Basic Support Life & First Aid

Article 1:

First Things First: Effectiveness
and Scalability of a Basic
Prehospital Trauma Care
Program for Lay First-
Responders in Kampala,
Uganda

First Things First: Effectiveness and Scalability of a Basic
Prehospital Trauma Care Program for Lay First-
Responders in Kampala, Uganda

Sudha Jayaraman1, Jacqueline R. Mabweijano2, Michael S. Lipnick3, Nolan Caldwell4, Justin Miyamoto4,
Robert Wangoda2, Cephas Mijumbi5, Renee Hsia6, Rochelle Dicker7, Doruk Ozgediz8*

1 Department of Surgery, University of California San Francisco, San Francisco, California, United States of America, 2 Department of Accident and Emergency, Mulago
Hospital and Makerere University, Kampala, Uganda, 3 Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts, United States of America,
4 School of Medicine, University of California San Francisco, Office of Medical Education, San Francisco, California, United States of America, 5 Department of Anaesthesia,
Mulago Hospital and Makerere University, Kampala, Uganda, 6 Department of Emergency Medicine, University of California San Francisco, San Francisco, California, United
States of America, 7 Department of Surgery, San Francisco General Hospital, San Francisco, California, United States of America, 8 Division of Pediatric Surgery, Hospital for
Sick Children, University of Toronto, Toronto, Ontario, Canada

Abstract

Background: We previously showed that in the absence of a formal emergency system, lay people face a heavy burden of
injuries in Kampala, Uganda, and we demonstrated the feasibility of a basic prehospital trauma course for lay people. This
study tests the effectiveness of this course and estimates the costs and cost-effectiveness of scaling up this training.

Methods and Findings: For six months, we prospectively followed 307 trainees (police, taxi drivers, and community leaders)
who completed a one-day basic prehospital trauma care program in 2008. Cross-sectional surveys and fund of knowledge
tests were used to measure their frequency of skill and supply use, reasons for not providing aid, perceived utility of the
course and kit, confidence in using skills, and knowledge of first-aid. We then estimated the cost-effectiveness of scaling up
the program. At six months, 188 (62%) of the trainees were followed up. Their knowledge retention remained high or
increased. The mean correct score on a basic fund of knowledge test was 92%, up from 86% after initial training (n = 146
pairs, p = 0.0016). 97% of participants had used at least one skill from the course: most commonly haemorrhage control,
recovery position and lifting/moving and 96% had used at least one first-aid item. Lack of knowledge was less of a barrier
and trainees were significantly more confident in providing first-aid. Based on cost estimates from the World Health
Organization, local injury data, and modelling from previous studies, the projected cost of scaling up this program was
$0.12 per capita or $25–75 per life year saved. Key limitations of the study include small sample size, possible reporter bias,
preliminary local validation of study instruments, and an indirect estimate of mortality reduction.

Conclusions: Lay first-responders effectively retained knowledge on prehospital trauma care and confidently used their
first-aid skills and supplies for at least six months. The costs of scaling up this intervention to cover Kampala are very
modest. This may be a cost-effective first step toward developing formal emergency services in Uganda other resource-
constrained settings. Further research is needed in this critical area of trauma care in low-income countries.

Citation: Jayaraman S, Mabweijano JR, Lipnick MS, Caldwell N, Miyamoto J, et al. (2009) First Things First: Effectiveness and Scalability of a Basic Prehospital
Trauma Care Program for Lay First-Responders in Kampala, Uganda. PLoS ONE 4(9): e6955. doi:10.1371/journal.pone.0006955

Editor: Peter McCulloch, University of Oxford, United Kingdom

Received June 28, 2009; Accepted August 2, 2009; Published September 11, 2009

Copyright: ß 2009 Jayaraman et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: Mulago Foundation; Laura Case Trust. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the
manuscript.

Competing Interests: The authors have declared that no competing interests exist.

* E-mail: [email protected]

Introduction has no formal emergency system, pre-hospital care in Kampala is
currently delivered by police, taxi drivers or community leaders
Low and middle income countries disproportionately bear the (Local Council officials) [5].
global burden of injury and annually, an estimated one to two
million injury deaths in these settings are potentially avertable To address this issue in Kampala, in 2008, we conducted a pilot
through improvements in trauma care [1,2,3]. However, many study on the burden of injury seen by lay people and designed and
low and middle income countries do not have formal emergency implemented a context-appropriate course on prehospital trauma
systems in place; thus an estimated 80% of injury deaths in these care for lay first-responders based on the World Health
countries occur in the pre-hospital setting [1]. Organization’s recommendations that lay first-responders should
be the first step towards developing formal emergency medical
Uganda is no different. The lack of access to emergency medical services in settings without a formal prehospital system [7,8,9].
care for seriously injured people in Uganda’s capital city,
Kampala, has been documented by others [4,5,6]. Since the city That pilot study demonstrated that lay people in Kampala
encountered a substantial number of emergencies and deaths and

PLoS ONE | www.plosone.org 1 September 2009 | Volume 4 | Issue 9 | e6955

Lay First-Responders

provided emergency care in many cases but were not well- were used to compare changes in fund of knowledge, barriers to
supported in this role. We also demonstrated the feasibility of providing aid, perceived utility of the kit and course and
designing and implementing a context-appropriate basic prehos- confidence in providing aid. Spearman’s correlation coefficient
pital trauma course that increased the fund of knowledge of these was used for comparing total test scores and confidence. Sub-
lay people in essential trauma care [10]. We now report the group analysis, to compare police trainees against all others, was
effectiveness of this pilot program in Kampala, Uganda and planned a priori, based on previous findings [9].
estimate the costs of scaling up the program to provide emergency
services across Kampala. Cost and Cost-Effectiveness Analysis
We measured the costs per capita, per death averted, and per
Since the initial training, we prospectively followed this cohort to
assess the effectiveness of our training program at six months. We life year saved of scaling up our training program based on
hypothesized that participants would retain knowledge of first-aid previous projections that 6000 to 12000 lay first-responders would
learned during the training course, use the skills and kit frequently be needed to provide basic prehospital trauma care for Kampala’s
and feel more empowered to provide emergency services. We also 1.2 million people. These projections, based on a conflict-affected
measured the costs and cost-effectiveness of scaling up this program setting, are outlined in the DCP2 [9]. We estimated projected cost-
to cover the city of Kampala based on the methods outlined in the effectiveness based on recent reports on the impact of using trained
Disease Control Priorities Project, 2nd edition (DCP2) [9]. lay first-responders on injury mortality [12]. Husum et al. reported
a 15% mortality reduction (95% CI 8–22%) when trained lay first-
Methods responders and paramedics provided prehospital care in Cambo-
dia and northern Iraq. Since similar estimates are not available for
First-Aid Program Kampala, we applied Husum’s figures to approximately 1600
Our training program, previously described in detail [10], was injury deaths recorded by the Mulago Hospital and Kampala City
Mortuaries every year. We used these figures as the base-case
conducted in Kampala (population: 1.2 million) [11] and consisted scenario in Kampala and conducted deterministic sensitivity
of a day-long modified basic first-aid course on trauma and a free analysis to account for variations in individual cost parameters
basic first-aid kit for each participant. Participants had been such as the impact of providing of first-aid supplies and a more
selected through convenience sampling from a 20 square mile conservative 7.5% mortality reduction on costs per death averted
catchment area around Mulago Hospital. The course curriculum and LY saved.
was adapted from prior models by local stakeholders. The core
curriculum included universal precautions, scene management, Results
external compression for haemorrhage control, basic airway
control, recovery position, safe lifting and transportation of injured Effectiveness of First-aid Program
victims, splinting fractures and triage. After the course, each We enrolled 309 trainees in the study and were able to evaluate
participant received a first-aid kit assembled using locally available
materials and instructions on restocking the kit. 127 participants (41%) at three months and 188 (62%) at six
months.
Study Design and Data Collection Methods
Cross-sectional surveys and fund of knowledge tests were Perception of training program and ability to provide
first-aid. Trainees rated the course and kit highly throughout
conducted immediately before and after training and at three and the six months. Overall, on a scale of one to five, with five being
six months. We collected data on the frequency and types of aid extremely useful, 98% of trainees rated the course at 4 or 5 out of
provided, supply use, barriers to providing aid, perceived utility of 5 at three months (n = 124) and 99% rated it the same at six
the course and kit and self-reported confidence in giving first-aid. months (n = 185). On the same scale, at three months, 97% rated
Fund of knowledge tests consisted of five questions covering each the kit 4 or 5 out of 5 (n = 125) and 94% did the same at 6 months
core skill area. Data collection instruments were designed in English (n = 185). When trainees’ confidence in providing first-aid for
and Luganda and pilot-tested twice in Kampala before use by study trauma was measured from zero to five, the mean before the
personnel. Structured one-on-one interviews were used at study training course was 3.1. At three months, this increased to 4.2
onset. Self-administered questionnaires were used at three months (n = 124, p,0.0001) and remained at that level at six months
and mobile phone surveys at six months for logistical reasons. A (n = 181, p,0.0001). There was no correlation between increasing
half-day refresher session was conducted at three months. levels of confidence and mean test scores, at six months.

All trainees were also asked to complete a pictographic Self-reported Use of First-Aid Skills and Supplies. At
encounter record after each emergency they encountered during three months, 84% of participants had used at least one skill taught
the study period. This record was designed and tested in both in the course. By six months, this frequency increased to 97%. At six
languages and was distributed at study onset (Figure 1). Completed months, the most commonly used skills were haemorrhage control,
records were collected at Mulago Hospital and at the five recovery position and lifting/moving (table 1). In paired analysis,
community clinics within the catchment area. participants reported using basic airway control and external
compression for haemorrhage control significantly more frequently
Institutional Review Board approvals were obtained from the at six months than they did at before the course but used safe lifting
University of California, San Francisco, Mulago Hospital and the and moving techniques significantly less frequently (table 2).
Uganda National Council of Science and Technology. Informed
consent to participate in the study was obtained from all trainees. At three months, 85% of participants had used at least one
Throughout this study, participants received no financial incentive component of the study kit and by six months, this frequency
to perform first aid in the field, or to complete encounter records. increased to 96%. At six months, the most commonly used items
Nominal stipends were given on training days to replace lost income. were gloves, cotton and gauze (table 1). Every item was considered
‘‘hard to stock’’ by trainees at 6 months despite the restocking help
Statistical Analysis provided by the study team.
All data were analyzed using Stata version 10 (College Station,
Barriers to Providing First-Aid. Inability to provide aid
TX, 2007). Paired t-tests, x2 tests and logistic regression analysis due to lack of knowledge decreased significantly at six months

PLoS ONE | www.plosone.org 2 September 2009 | Volume 4 | Issue 9 | e6955

Lay First-Responders

Figure 1. Prospective Data Collection Form for First Responders.
doi:10.1371/journal.pone.0006955.g001

PLoS ONE | www.plosone.org 3 September 2009 | Volume 4 | Issue 9 | e6955

Lay First-Responders

Table 1. Most commonly used skills and supplies from the Table 2. Use of specific skills before and six months after the
first-aid course. first-aid course.

Rank Survey data n = 179 Encounter records n = 153 Before Six Months After

Skills Hemorrhage control (74%) Lifting/moving (47%) Skill Training Training* P value{
1 Recovery position (65%) Hemorrhage control (42%)
2 Lifting/moving (65%) Recovery position (40%) Checking airway/breathing 50% 64% 0.0041
3 Control bleeding 61% 74% 0.0075
Supplies Gloves (92%) Gloves (83%) Splinting 31% 35% NS
1 Cotton (79%) Gauze (63%) Lifting/moving 78% 67% 0.0072
2 Gauze (74%) Bandage (43%) Recovery position{ 43% 57% 0.04
3
*n = 171 pairs, except for checking airway/breathing: n = 169 pairs.
doi:10.1371/journal.pone.0006955.t001 {by paired t test.
{was evaluated at three months and not at initial training.

doi:10.1371/journal.pone.0006955.t002

compared to pre-training although more trainees were unable to records, followed by Local Council officials (24%) and taxi drivers
help in an emergency due to work or travel-related reasons (16%). Injured victims were mostly young men (median age = 25,
(table 3). male = 88%) with injuries most commonly to the extremities (67%)
or head and neck (52%). Skill and supply use reported were similar
Knowledge Retention. At six months, trainees answered to findings from the cross-sectional surveys (table 2). Time from
92% of all fund of knowledge questions correctly (mean %) injury to response by study participants and subsequent arrival to
compared to 86% immediately after the initial training (n = 146 the hospital was 39 min (median). In comparison, the Mulago
pairs, p = 0.0016). Hospital Trauma Registry data notes a median of 120 minutes
from time of injury-to-hospital arrival.
Over 95% of trainees scored correctly on scene management,
basic airway control and recovery position and 80% did so on Cost Analysis
external compression for haemorrhage control and safe transpor- The WHO and World Bank suggest that 9000 trainees (range
tation (table 4). Attending the three month refresher training did
not affect the fund of knowledge test scores at 6 months. 6000–12000) would need to be trained to provide prehospital
services for Kampala’s 1.2 million people. For a base case
Sub-group analysis scenario, we calculated the annual cost of training 9000 trainees
over 3 years using our program. This resulted in a projected
Effect of being a Police trainee on outcomes at six annual cost of $47,854 and a per capita cost of $0.12 for the base
months. Police trainees were likely to have a higher test score case (table 5). When one-way sensitivity analysis was to evaluate
at 6 months compared to other trainees and this persisted after the addition of a $16 first-aid kit and an additional $16 restocking
controlling for refresher training attendance (OR = 2.3, p = 0.010). supplies, we found that overall costs ranged up $143,854 and per
They were also more likely to report using the recovery position capita costs up to $0.36 (table 6).
(OR = 1.9, p = 0.05), safe lifting or moving techniques (OR = 2.8,
p = 0.002) and splinting (OR = 1.9, p = 0.06) at six months, than Projected Cost-Effectiveness of Implementing Lay First-
the other trainees. These findings were adjusted for refresher Responder Training in Kampala
course attendance. When frequencies of skill and supply use were
measured, police trainees reported using both more frequently If our program was scaled up to cover Kampala, 240 deaths
than other trainees (skill use: OR = 1.9, p,0.025 and kit use: OR (15%) could be averted every year, which results in a projected
3.1, p,0.0001). cost per death averted of $598 using the base case or a cost per life
year (LY) saved of $25 based on an average age of death from
Correlation of survey data with encounter records.. We injury of 27 with a life expectancy of 24 additional years at that
collected 153 encounter records prospectively from trainees over
the six month study period. Police officers submitted 44% of

Table 3. Reported reasons for not providing assistance.

Reason* Before Training At six months P value{

I did not know what medical help was needed 8% 3% 0.018
I did not feel safe providing help 2.8% 3.3% NS
I did not have equipment to provide help 6% 9.5% NS
I could not help due to work-related reasons 2.8% 13% 0.0003
I could not help because I was travelling and could not stop 2% 0.12% 0.0006
I had other reason for not providing aid (n = 178) 4.5% 8.4% NS
I had more than one reason to not provide aid (n = 180) 6% 9% NS

*n = 179 unless otherwise noted
{by paired t test
doi:10.1371/journal.pone.0006955.t003

PLoS ONE | www.plosone.org 4 September 2009 | Volume 4 | Issue 9 | e6955

Lay First-Responders

Table 4. Participants’ fund of knowledge at six months compared to after initial training.

Scene management (n = 146) % answering correctly % answering correctly p value*
Checking airway/breathing(n = 146) immediately after training at six months
External compression for haemorrhage control (n = 147) NS
Use of recovery position (n = 145) 95% 98% NS
Speeding to hospital (n = 138) 92% 95% 0.04
76% 84% NS
*by paired t test. 94% 95% NS
doi:10.1371/journal.pone.0006955.t004 87% 80%

time. Results of the sensitivity analysis when a $16 kit and supplies critical to dissuade any incentive to charge informal fees for
were included showed that costs per death averted increased to a services. This study also showed shorter times from injury to
maximum of $1798 or $75 per life year saved. When the more hospital arrival, although this compares two different datasets; our
conservative (7.5%) mortality reduction was used, this upper limit prospectively collected encounter records and the hospital trauma
further increased to $3596 per death averted or $150 per life year registry. Nonetheless, the findings suggest that delivery of
saved (table 7). prehospital care did not delay access to care.

Discussion This study can be compared with others’ experiences in
developing prehospital systems in resource-constrained settings.
To address global disparities in trauma care, the World Health In urban Ghana, truck drivers who completed a similar context-
Organization (WHO) Essential Trauma Care Guidelines outline appropriate first-aid course showed a significant increase in skill
the ‘‘needs of the injured patient,’’ such as airway management use at 10 months based on response rate of 28% [18]. However,
and bleeding control, as essential health services for all people trainees assembled first-aid kits at their own expense and at 10
regardless of resources or context. The mandate of health systems months, only 27% of trainees carried materials useful for universal
to provide essential services has also been stressed through the precautions such as gloves or plastic bags. Our study conducted
rights-based approach to health [13]. The availability of adequate more in-depth and periodic assessments of participants’ fund of
emergency medical services is often considered a basic human knowledge and use of free first-aid kits.
right in high-income countries [14,15,16] and this right should not
be neglected in low and middle income countries which bear the Husum et al. have demonstrated a 15% reduction in mortality
disproportionate burden of injury. In addition, calls for injury of war and landmine victims in Iraq and Cambodia where
research in resource-constrained settings have particularly stressed paramedics and lay first-responders were trained to provide
the need for intervention-based research [17]. prehospital trauma care [12,19,20]. Despite extensive training
programs for both groups, most prehospital care provided was
Our study thus fills these critical needs. Despite a substantial solely basic life support. Others have argued for implementation of
burden of injury in Uganda, Kampala, its largest city, has no this model in non-conflict settings without formal emergency
formal prehospital emergency system. We found that lay people systems [9]. We designed our study based on a belief that a simple
can effectively retain knowledge of prehospital trauma care ‘‘scoop-and-run’’ approach with basic life-saving interventions will
learned through a context-appropriate first-aid course for at least be most practical and effective in the absence of second-tier
six months. Trainees found this basic intervention useful and after responders such as paramedics who are trained in advance airway
the training, were able to more confidently deploy these new skills. management, vascular access and fluid resuscitation. In addition,
Our findings suggest that police may be the ideal first-responders we could not provide extensive training for our participants due to
in Kampala given their higher knowledge retention and skills and opportunity costs of removing police officers from duty and
supply use compared with other trainees. Their established imposing a greater loss of income for the taxi drivers. Therefore,
communication and transportation networks are also an advan- while not directly comparable to previous work, our study suggests
tage. However, appropriate recognition for such services will be that high knowledge retention and greater use of first-aid skills may
be rough proxies for improvements in trauma care.
Table 5. Annual cost of training 3000 lay first-responders
(Base case). We also estimated that the costs of scaling up this intervention to
cover all of Kampala’s 1.2 million people was $25–150 per life year
Trainees for full-day training 3000 saved ($598–$3596 per death averted) which is comparable to the
Trainee salary at WHO level 1 $14,610 costs of traffic enforcement in Kampala at $27 per life year saved
Trainer days at 6 trainers & 50 trainees per day 360 [21] or the distribution of antiretroviral medications for HIV in
Trainer salary at WHO level 3 $3,244 Uganda at $600 per life year saved [22]. As such, this intervention
Room rental ($75/day) $27,000 could be considered very cost-effective. Additionally, Uganda’s
Materials ($1/trainee) $3,000 health policy is based on a minimum package of essential health
Total costs for base case $47,854 services, which was initially costed at $34 per capita by the WHO
[23,24]. Current public expenditure on health is USD $22 per
doi:10.1371/journal.pone.0006955.t005 capita in Uganda [25]. Thus, $0.12 per capita would be a modest
price to pay for prehospital care, especially since the current
package does not include emergency services for trauma.

However, our study has several limitations. First, we did not
directly measure changes in mortality. Our small sample size and

PLoS ONE | www.plosone.org 5 September 2009 | Volume 4 | Issue 9 | e6955

Lay First-Responders

Table 6. Costs per capita of training 9000 lay first-responders for Kampala.

Base case Annual Costs Costs to train all 9000 Costs per capita
Base case with $8 kit
Base case with $8 kit and $8 worth of extra supplies $47,854 $143,561 $0.12
Base case with $16 kit $71,854 $215,561 $0.18
Base case with $16 and $16 worth of extra supplies $95,854 $287,561 $0.24
$95,854 $287,561 $0.24
doi:10.1371/journal.pone.0006955.t006 $143,854 $431,561 $0.36

poor health system infrastructure limit the capacity to follow an Despite study limitations, many areas for further research exist.
injured victim across the continuum of care. Second, skill and Better methods are needed to measure the impact of training lay
equipment use may have been over-reported by participants. first-responders on trauma morbidity and mortality. While we
Although we attempted to address this using prospective encounter established the need for a simple first-aid kit for effective provision
records, hospital-level outcomes data would have increased the of prehospital care, mechanisms to ensure restocking of kits also
rigor of our study. Third, our study design may not be ideal. While need to be determined. Facility-based records could more
a randomized controlled trial is the gold standard to prove accurately measure economic burden and may help determine
intervention effectiveness, its use to test the benefits of essential the accuracy of our projections.
skills of basic trauma care in the prehospital setting is not ethical.
Furthermore, we did not use a control group in this study and Training lay first-responders may be a critical and cost-effective
relied on before-and-after comparisons to establish effectiveness first step to develop a formal emergency medical system. However,
which may be less than ideal. this intervention needs to be complemented by a second-tier of
responders, a call center, and appropriate transportation mecha-
Other contextual factors may have also affected our results. We nisms as well as facility-based training for clinicians to provide
could not control for changes in the frequency of skill use based on appropriate trauma care when victims reach health facilities. The
changes in job posting. Police trainees are rotated periodically, precedent for such comprehensive trauma systems development is
which removes them from active field duty, and thus potentially being established in other resource-constrained settings and could
limits their ability to be present at emergencies. During the study, provide practical lessons for Ugandan policy-makers
more trainees reported being unable to assist in emergencies due [26,27,28,29,30]. Since injury disproportionately affects the
to work or travel related reasons suggesting that trainee selection working poor who rely on high-risk modes of transport such as
was not ideal or that opportunity costs prohibited our trainees non-motorized vehicles, motorized two-wheeled vehicles and foot
from responding to emergencies. Finally, the implications of some traffic, by improving access to emergency services, a formal
of our results remain unclear. Some skills, such as safe lifting and prehospital emergency system in Uganda can become a ‘pro-poor’
moving, were reported less frequently at six months than after the strategy and increase equity [31,32]. Additionally, organized
initial training perhaps because their new knowledge or increased emergency services could improve health system efficiency by
credibility as emergency providers allowed them to harness help to building on existing informal mechanisms of care. Most impor-
transfer victims in emergencies or perhaps this skill simply may tantly, the global health community must urgently provide greater
need more emphasis in future training. priority and resources for injury intervention research in resource-
constrained settings that are particularly burdened with high
Our cost and cost-effectiveness analyses must be interpreted incidence of trauma.
with caution, even though sensitivity analyses were used. Estimates
of the number of lay first-responders needed for Kampala and In conclusion, a lay first-responder training program is a
their potential impact on mortality were determined based on practical and effective first-step towards developing a formal
work primarily in a conflict-affected setting. By comparison, a emergency system in Uganda. It is likely to be very cost-effective in
decreased incidence of trauma and associated frequency of skill use this setting. Establishing and scaling up this intervention with in
in urban Kampala may have led to an overestimate of the cost- Kampala should be a key priority for Ugandan policy-makers.
effectiveness. Incorporating emergency services into the essential package of

Table 7. Cost per death averted and per life year saved of training 9000 lay first responders.

Estimated Deaths averted, n (%) Costs to 240 (15%) Cost per 120 (7.5%) Cost per
train all 9000 LY saved LY saved
Base case Cost per Cost per
Base case with $8 kit $143,561 death averted $25 death averted $50
Base case with $8 kit and $8 worth of extra supplies $215,561 $37 $75
Base case with $16 kit $287,561 $598 $50 $1,196 $100
Base case with $16 and $16 worth of extra supplies $287,561 $898 $50 $1,796 $100
doi:10.1371/journal.pone.0006955.t007 $431,561 $1,198 $75 $2,396 $150
$1,198 $2,396
PLoS ONE | www.plosone.org $1,798 $3,596

6 September 2009 | Volume 4 | Issue 9 | e6955

Lay First-Responders

health care would critically address the disproportionate global tor. Thanks also to UCSF Department of Surgery Scientific
injury burden shouldered by the poor. Results of this program Publications Manager, Pamela Derish for editorial support and
could be useful in other resource-constrained settings that lack Dr. Kayvan Roayaie for statistical assistance.
emergency medical systems.
Author Contributions
Acknowledgments
Conceived and designed the experiments: SJ JRM MSL RYH RD DO.
We would like to sincerely thank the following contributors who Performed the experiments: SJ JRM MSL NC JM RW CM RYH DO.
gave valuable assistance at during various stages of this project: Analyzed the data: SJ JRM MSL NC JM RW CM RYH RD DO.
Moses Dumba from the Ugandan Taxi Operators and Drivers Contributed reagents/materials/analysis tools: SJ JRM MSL NC JM RW
Association, Red Cross First-Aid Instructors: Robert Okuyat, CM RYH RD DO. Wrote the paper: SJ JRM RYH RD DO. Obtained
Juliet Kiyimba, Ivan Luwaga, Daniel Nsubuga, and Red Cross funding: DO.
Volunteer: Yusuf Kimbowa. Sarah Nakitto, Research Coordina-
16. Baker R (1999) American independence and the right to emergency care. JAMA
References 281: 859–860.

1. Mock CN, Jurkovich GJ, nii-Amon-Kotei D, Arreola-Risa C, Maier RV (1998) 17. Roberts I, Shakur H, Edwards P, Yates D, Sandercock P (2005) Trauma care
Trauma mortality patterns in three nations at different economic levels: research and the war on uncertainty. Bmj 331: 1094–1096.
implications for global trauma system development. J Trauma 44: 804–812;
discussion 812–804. 18. Mock CN, Tiska M, Adu-Ampofo M, Boakye G (2002) Improvements in
prehospital trauma care in an African country with no formal emergency
2. Mock C, Nguyen S, Quansah R, Arreola-Risa C, Viradia R, et al. (2006) medical services. J Trauma 53: 90–97.
Evaluation of Trauma Care capabilities in four countries using the WHO-
IATSIC Guidelines for Essential Trauma Care. World J Surg 30: 946–956. 19. Husum H, Gilbert M, Wisborg T, Van Heng Y, Murad M (2003) Rural
prehospital trauma systems improve trauma outcome in low-income countries: a
3. Mathers CD, Loncar D (2006) Projections of global mortality and burden of prospective study from North Iraq and Cambodia. J Trauma 54: 1188–1196.
disease from 2002 to 2030. PLoS Med 3: e442.
20. Husum H, Gilbert M, Wisborg G, Pilgram-Larsen J (2003) Landmine injuries: a
4. Kobusingye OC, Lett RR (2000) Hospital-based trauma registries in Uganda. study of 708 victims in North Iraq and Combodia. Military Medicine 168:
J Trauma 48: 498–502. 934–940.

5. Kobusingye OC, Guwatudde D, Owor G, Lett RR (2002) Citywide trauma 21. Bishai D, Asiimwe B, Abbas S, Hyder AA, Bazeyo W (2008) Cost-effectiveness of
experience in Kampala, Uganda: a call for intervention. Inj Prev 8: 133–136. traffic enforcement: case study from Uganda. Inj Prev 14: 223–227.

6. Mock CN, nii-Amon-Kotei D, Maier RV (1997) Low utilization of formal 22. Bishai D, Colchero A, Durack DT (2007) The cost effectiveness of antiretroviral
medical services by injured persons in a developing nation: health service data treatment strategies in resource-limited settings. Aids 21: 1333–1340.
underestimate the importance of trauma. J Trauma 42: 504–511; discussion
511–503. 23. Sachs J (2001) Investing in Health: The Report of the Commission on
Macroeconomics and Health. Geneva: World Health Organization.
7. WHO (2005) Prehospital Trauma Care Systems. Geneva: World Health
Organization. 24. (2000)Health Sector Strategic Plan I (2000/1–2004/5). Kampala, Uganda:
Ministry of Health.
8. Mock C, Lormand J, Goosen J, Joshipura M, Peden M (2004) Guidelines for
Essential Trauma Care. Geneva: World Health Organization. 25. WHO (2008) World Health Statistics.
26. Gautham S (2008) Strengthening Trauma Care in the United Arab Emirates.
9. Kobusingye OC, Hyder AA, Bishai D, Joshipura M, Hicks ER, et al. (2006)
Emergency Medical Services. In: Jamison DT, et al. (2006) Disease Control American College of Surgeons 94th Annual Clinical Congress. San Francisco,
Priorities in Developing Countries. 2nd ed. New York and Oxford: Oxford CA USA.
University Press. 27. Tiska MA, Adu-Ampofo M, Boakye G, Tuuli L, Mock CN (2004) A model of
prehospital trauma training for lay persons devised in Africa. Emerg Med J 21:
10. Jayaraman S, Mabweijano J, Lipnick M, Caldwell N, Miyamoto J, et al. (2009) 237–239.
Current patterns of prehospital trauma care in Kampala, Uganda and the 28. Aboutanos MB, Rodas EB, Aboutanos SZ, Mora FE, Wolfe LG, et al. (2007)
feasibility of a lay-first-responder training program. World Journal of Surgery Trauma education and care in the jungle of Ecuador, where there is no
(accepted for publication). advanced trauma life support. J Trauma 62: 714–719.
29. Ayyaz M (2008) Strenthening trauma care in Pakistan. American College of
11. Kampala City Council District Profile. Surgeons 94th Annual Clinical Congress. San Francisco, CA USA.
12. Husum H, Gilbert M, Wisborg T (2003) Training pre-hospital trauma care in 30. Poggetti R (2008) Strengthening trauma care in Brazil. American College of
Surgeons 94th Annual Clinical Congress. San Francisco, CA USA.
low-income countries: the ‘Village University’ experience. Med Teach 25: 31. Nantulya VM, Reich MR (2003) Equity dimensions of road traffic injuries in
142–148. low- and middle-income countries. Inj Control Saf Promot 10: 13–20.
13. Backman G, Hunt P, Khosla R, Jaramillo-Strouss C, Fikre BM, et al. (2008) 32. Peden M, Scurdfield R, Sleet D, Mohan D, Hyder AA, et al. (2004) World
Health systems and the right to health: an assessment of 194 countries. Lancet report on road traffic injury prevention. Geneva: World Health Organization.
372: 2047–2085.
14. Mock C, Quansah R, Krishnan R, Arreola-Risa C, Rivara F (2004)
Strengthening the prevention and care of injuries worldwide. Lancet 363:
2172–2179.
15. Razzak JA, Kellermann AL (2002) Emergency medical care in developing
countries: is it worthwhile? Bull World Health Organ 80: 900–905.

PLoS ONE | www.plosone.org 7 September 2009 | Volume 4 | Issue 9 | e6955

Article 2:

GNPs-CS/KGM as Hemostatic
First Aid Wound Dressing
with Antibiotic Effect: In Vitro
and In Vivo Study

GNPs-CS/KGM as Hemostatic First Aid Wound Dressing
with Antibiotic Effect: In Vitro and In Vivo Study

Li Fan1., Chong Cheng1., Youbei Qiao1, Fei Li1, Wei Li1, Hong Wu1*, Bo Ren2*

1 Department of Pharmaceutical Chemistry and Analysis, School of Pharmacy, Fourth Military Medical University, Xi’an, Shaanxi, China, 2 Department of Anesthesia, 302
Military Hospital of China, Beijing, China

Abstract

Ideal wound dressing materials should create a good healing environment, with immediate hemostatic effects and
antimicrobial activity. In this study, chitosan/konjac glucomannan (CS/KGM) films embedded with gentamicin-loaded
poly(dex-GMA/AAc) nanoparticles (giving GNP-CS/KGM films) were prepared as novel wound dressings. The results revealed
that the modified CS/KGM films could be used as effective wound dressings and had significant hemostatic effects. With
their microporous structure, the films could effectively absorb water from blood and trap blood cells. The gentamicinloaded
poly(dex-GMA/AAc) nanoparticles (GNPs) also further promoted blood clotting, with their favorable water uptake capacity.
Thus, the GNP-CS/KGM films had wound healing and synergistic effects that helped to stop bleeding from injuries, and also
showed good antibiotic abilities by addition of gentamicin to the NPs. These GNPCS/KGM films can be considered as
promising novel biodegradable and biocompatible wound dressings with hemostatic capabilities and antibiotic effects for
treatment of external bleeding injuries.

Citation: Fan L, Cheng C, Qiao Y, Li F, Li W, et al. (2013) GNPs-CS/KGM as Hemostatic First Aid Wound Dressing with Antibiotic Effect: In Vitro and In Vivo
Study. PLoS ONE 8(7): e66890. doi:10.1371/journal.pone.0066890

Editor: Vishal Shah, Dowling College, United States of America

Received January 31, 2013; Accepted May 10, 2013; Published July 16, 2013

Copyright: ß 2013 Fan et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This work was partially funded by grants from the National Natural Science Foundation of China No. 81201179 and No. 81271687. No additional
external funding received for this study. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the
manuscript.

Competing Interests: No conflict of interest exists in the submission of this manuscript.

* E-mail: [email protected] (HW); [email protected] (BR)

. These authors contributed equally to this work.

Introduction based nanoparticles may have great valuable for wound healing
and hemorrhage control.
Ideal wound dressing materials should have following charac-
teristics [1–3]: create a moist wound healing environment, absorb However, CS dressing has poor tensile strength and elasticity.
excess exudate, allow gaseous exchange, be removed easily without Hence development of high strength composites that are
trauma to the wound and better with effects of hemorrhage biocompatible and that can help in wound healing may be
control, antimicrobial activities, nontoxic and biocompatible. necessary for CS as wound dressing applications. Improvement of
water absorbing and water permeating properties is also necessary
Many efforts have been made to prepare such wound healing for the products derived from CS. Konjac glucomannan (KGM),
materials including proteins (e.g., fibrinogen [4], thrombin [5], could significantly improve mechanical properties of CS as a
collagen [6], gelatin [7], albumin [8]), and polysaccharides [9] wound dressing materials. The tensile strength and breaking
(chitosan, chitin, poly (Nacetyl glucosamine) and cellulose). elongation of blend films were enhanced about 40% and 30%,
Although most of these materials have been proved valuable for respectively [24]. Moreover, the addition of KGM improves the
wound healing [10–13] and hemorrhage control [14–16] in many biocompatibility of CS materials [25].
cases, their main limitations are lack of efficacy in bleeding, anti-
inflammatory and antibiosis. Though films based on KGM/CS has been found to be
nontoxic, biodegradable, biofunctional, biocompatible in addition
In recent years, scientists have attempted to apply dehydrated to having antimicrobial characteristics [26,27], the antibacterial
zeolite material to the bleeding site in order to induce hemostasis activity of chitosan was inferior to that of the organic antibacterial
through dehydration of the wounded area. However, zeolite compounds and could not provide efficient antimicrobial activity
particles possess poor adhesive properties, and they will release or a continuous and sustained release of the antibacterial agent on
heat when in contact with water [17]. Many biomaterials based the wound surface.
nanoparticles, such as modified polysaccharides with acrylic
polymers [18–20], with similar strong capabilities of water uptake In recent times, there has been considerable interest in
with zeolite particles, could absorb large amount water from blood preparations of antibiotics loaded nanoparticles and films in order
to form a clot quickly. And polysaccharides dressing itself, such as to enhance the antimicrobial activity of wound dressing [28]. It
chitosan (CS) dressing, possessing blood clot formation activity and was reported by R. Hamblin that a dressing combining CS acetate
cell adhesiveness [21] has already been proved to have desirable with silver nanoparticles leaded to improved antimicrobial efficacy
qualities for wound h ealing [22] and hemostasis [23]. Thus, we against fatal infections [28]. In our previous study, we have
hypothesized that if CS dressings composited with polysaccharides

PLOS ONE | www.plosone.org 1 July 2013 | Volume 8 | Issue 7 | e66890

Antibiotic Hemostatic First Aid Wound Dressing

Figure 1. Schematic of derivatization of dextran and the free-radical mediated polymerization of the crosslinked polymer
nanoparticles.
doi:10.1371/journal.pone.0066890.g001

developed nanoparticles based on derivative dextran that have obtained from Dept. of Laboratory in Xijing hospital. Yunnan
shown great capabilities in drug-controlled release [29,30]. baiyao as a positive control was also obtained from Xijing hospital.

In this study, poly (dex-GMA/AAc) nanoparticles were also Poly (DEX-GMA/AAc) blank nanoparticles and
used as antibiotics gentamicin delivery vehicles in order to keep Gentamicin loaded nanoparticles synthesis and
gentamicin sustainable release. We kept on adjusting ratio between characterization
KGM and CS in order to obtain more efficient wound dressing
film with better tensile strength and breaking elongation. It was DEX-GMA precursor and Poly (DEX-GMA/AAc) nanoparti-
revealed by research result that gentamicin got well sustainable cles were synthesized as has been previously reported [30] in our
drug release profile from poly (dex-GMA/AAc) nanoparticles. paper. Though 3 methods have been reported in our previous
And the antibacterial test result revealed that it possessed paper for synthesis of Poly (DEX-GMA/AAc) nanoparticles,
continuously bacteriostatic activity after adhere to skin surface. method of free radical polymerization was testified to be the
Also, it was confirmed by in vitro and vivo study that CS/KGM preferred one with best repeatability and size distribution (As
film was valuable for wound healing and hemorrhage control due shown in Figure 1). In brief, dextran (5.0 g) and DMAP (1.0 g) was
to its significant promoting wound healing effect and fast dissolved in 50 ml of DMSO at room temperature. After
hemostatic effect. dissolution of DMAP, GMA (0.8 g) was added. The mixture was
stirred for 30 h at room temperature under nitrogen. The obtain
Materials and Methods dextran polymer was then precipitated with ethanol and fluffy
product polymers were obtained. The polymers were further
Materials dissolved in deionized water and reprecipitated out with ethanol
Dextran (Mn,70,000 g/mol) was obtained from Leuconostoc three times. The product was dispersed into distilled water,
dialyzed for 1 week at 4uC. After lyophilizing, the white Dex-
spp., N, N-Dimethylpyridin-4-amine (DMAP, 99%), Glycidyl GMA was obtained. The purified Dex-GMA was characterized by
methacrylate (GMA, 97%), Chitosan (Mn,75,000 g/mol, 75– 1H-NMR spectroscopy. Poly (DEX-GMA/AAc) blank nanopar-
85% deacetylated), and Gentamicin were purchased from Sigma- ticles were synthesized in 30 ml pH 8.0 phosphate buffers by a free
Aldrich. Dimethylsulfoxide (DMSO), N, N’-Methylenebisacryla- radical emulsion polymerization. Gentamicin loaded nanoparticles
mide (MBA), ammonium persulfate (APS), acrylic acid (AA), were obtained as the same method with initially adding
acetylacetone, and other chemical agent were acquired from Gentamicin (50 mg). AAc (0.2 g) was dissolved in 5 mL PBS
Fluka. Konjac Glucomannan (KGM) from Chengdu new inter- and then neutralized by NaOH solution (0.25 mol/L). Dex-GMA
state development Co., LTD, Dulbecco’s modified Eagle media (0.6 g) and MBA (2 mg/mL, 15 mL) were added into AAc
(DMEM) from Gibco and fetal calf serum (FBS) were used without solution and obtained mixture 1. Tween-80 (0.1 mL) as emulsifier
further purification. Phosphate buffered saline (PBS) was prepared was added directly into PBS, and mixture 1 was dropwised added
by dissolving 8.00 g NaCl, 0.20 g KCl, 1.15 g Na2HPO4, and into the solution. The reaction mixture was purged with nitrogen
0.24 g KH2PO4 into ,900 mL of water. The pH was adjusted to for 20 min and the reaction temperature was increased up to
7.40 with 1 M NaOH or 1 M HCl, and the solution was mixed 70uC. APS (10% w/v, 1 mL) as initiator was added and reacted
with additional water to 1.00 L in a volumetric flask. Bacteria for 5 h under nitrogen. Poly (dex-GMA/AAc) nanoparticles were
strains staphylococcus aureus (ATCC 25923), escherichia coli collected by centrifugation at 12000 rpm for 30 min. Excess
(ATCC 25922) and Pseudomonas aeruginosa (ATCC 27853) were

PLOS ONE | www.plosone.org 2 July 2013 | Volume 8 | Issue 7 | e66890

Antibiotic Hemostatic First Aid Wound Dressing

Figure 2. 1H-NMR spectra of DEX-GMA. (a), Morphology of nanoparticles observed by TEM (b) (A: blank nanoparticles, B: drug loaded
nanoparticles), Particle size distribution from DLS analysis (c) (A: blank nanoparticles, B: drug loaded nanoparticles).
doi:10.1371/journal.pone.0066890.g002

surfactant and unencapsulated gentamicin were removed by EE(%)~ amountofdruginnanoparticles |100
dialysis (dialysis bag with 10000 MWCO) for 1 day and then initialamountof drug
nanoparticles solution was lyophilized. The blank and drug loaded
nanoparticles were characterized for their size and surface KGM/CS film preparation and characterization
morphology by dynamic laser scattering (DLS) (Malvern Zetasizer KGM/CS membrane was prepared following Zhang’s previous
Nano S90) and transmission electron microscopy (TEM) (Hitachi
HT-7700). paper [32] using casting and solvent evaporation technique
[33,34] with some modification. KGM was purified by extraction
Gentamicin encapsulation efficiency (EE) and loading efficiency of phenol and ethanol (4:1, v/v) for 5 times and extraction of
(LE) were determined by dissolving 100 mg of drug loaded chloroform and ethanol (5:1, v/v) for 3 times. Purified KGM was
nanoparticles in 50 ml PBS buffer with 5 ml 0.1 mol/l HCl for obtained after vacuum dried. Then purified totally soluble KGM
12 h under 90uC water bath. Then filter the solution using was dissolved in distilled water to a concentration of 1 wt%. CS
Millipore Ultrafiltration (UF) membranes with MWCO 1000 and was dissolved in a 1wt% aqueous acetic acid to prepare a
the filtrate was brought to volume of 100 mL. Gentamicin was concentration of 1 wt% solution. The solutions of KGM and CS
diluted with 5 ml of water by vortexing and assayed photomet- with different mixing ratios [25/75, 50/50, and 75/25 KGM/CS
rically (310 nm) after derivation with o-phthalaldehyde [31]. EE (w/w)] were cast onto polystyrene plates and lyophilized. A series
and LE were calculated by the formula below of blend membranes were there by obtained and coded as
K25C75, K50C50, and K75C25. Pure CS film and pure KGM
LE(%)~ amountofdruginnanoparticles |100 film were prepared following the same protocol. The films were
amountof drugloadednanopaticles seal up for safekeeping after Co-60 irradiation. Film samples of

Figure 3. In vitro cumulative release profile of Gentamicin. From Poly (DEX-GMA/AAc) nanoparticles (A) and GNPs-CS/KGM (B).
doi:10.1371/journal.pone.0066890.g003

PLOS ONE | www.plosone.org 3 July 2013 | Volume 8 | Issue 7 | e66890

Antibiotic Hemostatic First Aid Wound Dressing

Figure 4. The cross-section morphology of blending film with different KGM/CS ratio by SEM.
doi:10.1371/journal.pone.0066890.g004

about 100 mm thickness were coated with gold in 0.1Torr following equation:
(0.13 mBar ) vacuum degrees. The cross section morphologies
were observed on a Hitachi S-3400N SEM. The tensile strength Table 1. The water vapor transmission rate of different
(TS) and breaking elongation (E) of the films were measured on an composition of CS/KGM films.
Instron apparatus (Model 3342, Instron Corp., Canton, MA). The
film (0.260.3 cm) were used in the stretch test under the stress of 1 Composition (CS/KGM) (wt%) WVTR (g?m22?day21)
and 2 N. Draft distance of clip was set as 20 mm with stretching
speed of 50 mm?min21. TS and E were calculated following the C25K75 3170 6 36
formula below C50K50 2580 6 90
C75K25 2282 6 73
TS~F=S; E~(Lb{L0)=L0|100% CS 2950 6 44

F (N): maximum stress of stretching, S (mm2): initial sectional area of Mean 6 S.D. (n = 6).
sample; L0: initial sample length; Lb: sample length at breakage doi:10.1371/journal.pone.0066890.t001
(n = 6).
SDP%~ Ws-Wo |100(n~6)
The swelling behaviors of pure CS and CS/KGM blend films Wo
were studied according to the method described by Wang et al.
[35]. The samples of a cylinder shape (diameter 30 mm, thickness
4 mm) were weighted (Wo) before immersing in PBS at 37uC and
immersed in PBS for different time intervals and weighted (Ws).
The swelling degree in PBS (SDP) was calculated using the

PLOS ONE | www.plosone.org 4 July 2013 | Volume 8 | Issue 7 | e66890

Antibiotic Hemostatic First Aid Wound Dressing

Where Ws is the weight of the hydrogel swollen in PBS, Wo is the In vitro Gentamicin release from Poly (DEX-GMA/AAc)
initial weight of hydrogel samples. The results are presented as a nanoparticles
mean value with a standard deviation (n = 6).
Dynamic dialysis method was employed to determine in vitro
The water vapor transmission rate (WVTR) across the pure CS drug release profile of Gentamincin loaded Poly (DEX-GMA/
and CS/KGM blend films was determined to the method AAc) nanoparticles. The drug content in the drug-loaded
described by Tsao et al. [36]. The pure CS and CS/KGM blend nanoparticles was determined upon natural degradation of
films were cut into 20 mm620 mm with a thickness of 3 mm, and nanoparticles without enzyme and drug release caused by
mounted on the mouth of cylindrical aluminium cups (14 mm nanoparticles swelling due to AAc residue on the polymer
diameter) containing 10 ml distilled water, and then placed in an backbone. An amount exactly weight (0.5 g) of drug-loaded
incubator at 37uC. The WVTR was calculated using the following nanoparticles dispersed in PBS (3 mL) was introduced into dialysis
equation: bag and dialysis against PBS (50 mL). Dialysate was incubated at
37uC under magnetically stirring. At each time interval, Dialysate
WVTR~ Wo-Wt |106(g=m2=day) (1.5 mL) was taken out and carefully transferred to a test tube.
A After each measurement, 1.5 mL of fresh buffer was added. After
each measurement, 3 mL of fresh buffer was added. 5 mL
Where WVTR is expressed in g/m2/day, A (mm2) is the bottle derivatization regent was added into each sample and 310 nm
mouth area, Wo (g) and Wf (g) are the weights of the device before was set as detective wavelength.
and after being placed in an incubator for 24 h, respectively. The
results are presented as a mean value with a standard deviation In vivo wound healing experiments
(n = 6). The wounds below were created in animals to mimic the

The in vitro degradation experiments were carried out in PBS conditions encountered in the surgical patients. The animal
at 37uC. The dry films were weighed and immersed in PBS for experiments were approved by the Ethics Committee of the
different time intervals during 70 days. Then the samples were Fourth Military Medical University. Rats were randomly divided
freeze-dried for 24 h to remove excess water, and weighed again. into 5 groups and anesthetized with ethyl ether. A1.5 cm-wide
Finally, the remainder weight was weighed to evaluate the stability wound was cut with scissors on the back of each shaved rat down
of pure CS and CS/KGM blend films using the following to the fascia layer. Rats of 4 groups were dressed with pure CS and
equation. CS/KGM blend films and the lasted group was dressed with
gauze as control. Wound closure observation was assessed by
Degradationrateð%Þ~ Wd |100% digital camera in the day 3, 7 and 14. The wound closure rate was
Wo calculated using the following equation:

Where degradation rate is expressed in %, and Wo and Wd are the Woundclosurerate~ So-Sd |100%
weights of the films before and after degradation for a specific time So
interval, respectively. The results are presented as a mean value
with a standard deviation (n = 6). Where the wound closure rate is expressed as a percentage %, and
So and Sd are the area of the originally created wound, and the
Figure 5. Degree of water uptake of the pure CS and KGM/CS area of the wound post-operative for a specific time interval,
blending films (a), TS and breaking elongation of KGM/CS respectively. The results are presented as a mean value with a
blending films with different KGM/CS ratio (b), Degradation standard deviation (n = 6).
profile of KGM/CS blending films with different KGM/CS ratio
(c). In addition, the wounds and the surrounding skin of post-
doi:10.1371/journal.pone.0066890.g005 operative for day 7 and 14 were fixed with 4% formaldehyde
solution, paraffin embedded and stained with hematoxylin-eosin
(HE) reagent for histological examinations.

Preparation of GNPs-CS/KGM
GNPs-CS/KGM was prepared according to the procedure of

C75K25 film with some modification. After purification of KGM,
CS solution (60 mL) was mixed with KGM solution (20 mL) and
0.2 g Gentamicin loaded nanoparticles was added into the mixture
to form GNPs-CS/KGM.

In vitro Gentamicin release from GNPs-CS/KGM
Dynamic dialysis method as described above was also employed

to determine in vitro drug release profile of Gentamincin from
GNPs-CS/KGM. An amount exactly weight (1.0 g) of drug-
loaded GNPs-CS/KGM in PBS (3 mL) was introduced into
dialysis bag and dialysis against PBS (50 mL). Dialysate was
incubated at RT under magnetically stirring. At each time interval
in 72 h, Dialysate (3 mL) was taken out and carefully transferred
to a test tube. After each measurement, 3 mL of fresh buffer was
added. 5 mL derivatization regent was added into each sample
and 310 nm was set as detective wavelength.

PLOS ONE | www.plosone.org 5 July 2013 | Volume 8 | Issue 7 | e66890

Antibiotic Hemostatic First Aid Wound Dressing

Figure 6. Histological examination by HE staining. Effect of different KGM/CS formulae, CS film and control group on incision wound in rat skin
(A) after 7d, (B) after 14d, (C) GNPs-CS/KGM treated after 3d.
doi:10.1371/journal.pone.0066890.g006

Biocompatibility evaluation of GNPs-CS/KGM used to assay the cytotoxicity on the Poly (dex-GMA/AAc)
Mice skin fibroblast L929, obtained from stomatological nanoparticles, C75K25 film and blank NPs-CS/KGM. Poly (dex-
GMA/AAc) nanoparticles, C75K25 film and blank NPs-CS/
hospital affiliated with Fourth military medical University, was KGM were added into DMEM cell culture medium with 10%

PLOS ONE | www.plosone.org 6 July 2013 | Volume 8 | Issue 7 | e66890

Antibiotic Hemostatic First Aid Wound Dressing

Figure 7. Inhibitory effect against the bacteria by Disc agar diffusion. A, B, and C presented C75K25 film, Drug loaded Poly (dex-GMA/AAc)
nanoparticles and GNPs-CS/KGM treated with three kinds of bacteria, respectively. (a) Staphylococcus aureus ATCC25923, (b) Escherichia coli
ATCC25922, (c) Green copper pseudomonas ATCC27853.
doi:10.1371/journal.pone.0066890.g007

FBS to obtain mixture of 0.1 g/mL, and extractedfor 24 h in contact areas of edible films with agar surface were then measured
incubator at 37uC. After applying extracts, medium was suck out (n = 6).
and diluted using fresh DMEM medium with 10% FBS to make
different concentration gradient sample solution (0.1, 0.05 and In vivo wound healing experiments
0.01 g/mL). L929 cell suspension (1|105/mL, 100 mL/well) was This study was conducted in accordance with Guide for the
seeded in 96 well plates, and cultured in a humidified incubator at
37uC with 95% air and 5% CO2 for 24 h. Sample solution of Care and Use of Laboratory Animals. Male Sprague-Dawley rats
different concentration (100 mL/well) was added separately and (250–400 g) were used for the study. They were anesthetized with
cultured for another 1d, 3d, and 5d. The medium was suck out an intraperitoneal injection of pentobarbital (50 mg/kg, Jinan
and MTT (0.5 mg/ml, 20 mL/well) with medium (100 mL/well) Haohua Industry Co., Ltd). The rat’s femoral vein was exposed
was added in each well, then culturedfor 4 h. The medium was and cut with scissors. Poly (dex-GMA/AAc) nanoparticles,
then discarded. Dimethylsulfoxide (DMSO) (150 mL) was added C75K25 film, native CS film and GNPs-CS/KGM were applied
into each well and OD value of the solutions was measured at 490 to the area of hemorrhage, separately, and firm pressure was
nm using plate reader (xMark). Mean values were obtained from applied. In order to ensure that the dressings was tightly adhered
six wells per group. to wound area, analytical weights of 50 g was used on the dressing
to give a certain pressure. At the meanwhile, timers was started
Antibacterial activity evaluation and ended until bleeding was stopped. Spilled blood was suck up
Antimicrobial activity tests of Poly (dex-GMA/AAc) nanopar- using gauze after weighing, and blood loss was calculated by gauze
and sample mass difference before and after bleeding. Yunnan
ticles, C75K25 film and GNPs-CS/KGM were carried out using baiyao powder was chosen as positive control and gauze to cover
agar diffusion method. C75K25 film and GNPs-CS/KGM were hemostatic as negative control. Hemostatic performance was
cut into a disc form of 7 mm diameter using a puncher and evaluated by blood loss and hemostatic time. (n = 6).
nanoparticles were scattered on sheep blood agar plates, which
had been previously seeded with inoculum (0.1 ml, 0.5 McIntosh Rats were randomly divided into 2 groups and anesthetized
concentrations) of Staphylococcus aureus ATCC 25923, Esche- with ethyl ether. A1.5 cm-wide wound was cut with scissors on the
richia coli ATCC 25922, and Green copper pseudomonas ATCC back of each shaved rat down to the fascia layer. One group was
27853. The plates were then incubated at 37uC for 16 h. The dressed with GNPs-CS/KGM and the other group was dressed
diameters of inhibitory zones surrounding film discs as well as the with gauze as control. Wound closure observation was assessed by
digital camera in the day 3, 7 and 14. The wound closure rate was
Table 2. Effect of KGM/CS films, CS film and control on calculated using the same equation as mentioned above.
wound area contraction during 14 days (values are mean 6
S.D., n = 6 observations in each group). In addition, the wounds and the surrounding skin of post-
operative for day 7 and 14 were fixed with 4% formaldehyde
Group Percent wound contraction (%) Day 3
Table 3. Effect of GNPs-CS/KGM and control on wound area
contraction during 14 days (values are mean 6 S.D., n = 6
observations in each group).

Day 3 Day 7 Day 14

C25K75 10. 464.21 28.4 65.34 79.1 64.26* Group Percent wound contraction (%) Day 3
C50K50 15.2 63.26* 31.165.27* 80.7 65.69**
C75K25 18.9 64.18** 32.064.24** 81.1 63.76** Day 3 Day 7 Day 14
CS 16.8 65.21* 30.7 63.43* 80.4 64.48**
GNPs-CS/KGM 19.7 62.28** 33.463.87** 82.8 64.25**

Control (gauze) 7.9 62.65 22.7 63.12 70.8 62.85 Control (gauze) 7.662.92 24.263.46 72.9 63.45

Values are significant (*) at P,0.05 and (**) at P,0.01as compared to the control Values are significant (*) at P,0.05 and (**) at P,0.01as compared to the control
group. group.
doi:10.1371/journal.pone.0066890.t002 doi:10.1371/journal.pone.0066890.t003

PLOS ONE | www.plosone.org 7 July 2013 | Volume 8 | Issue 7 | e66890

Antibiotic Hemostatic First Aid Wound Dressing

Figure 8. Effect of 0.25 wt% samples on heparinized rats blood and rat’s femoral vein injury treated with GNPs-CS/KGM group.
doi:10.1371/journal.pone.0066890.g008

solution, paraffin embedded and stained with hematoxylin-eosin mincin standard curve (A = 0.0227C-0.0408, R2 = 0.9996, linear
(HE) reagent for histological examinations. range, 10–35mg/mL). The shape, morphology and size of
nanoparticles obtained from Poly (DEX-GMA/AAc) blank
Statistical analyses nanoparticles and Gentamicin loaded nanoparticles were analyzed
Statistical analyses were performed with software (SPSS 13.0). by TEM and DLS. The TEM image of freeze-dried nanoparticles
was presented in Fig. 2b. The size distribution of nanoparticles in
Primary statistical analysis of data (mean 6 standard deviations) water was shown in Fig. 2c. It could be seen from DLS data that
was performed with the analysis of variance (ANOVA) to the blank and drug loaded nanoparticles diameters had peak at
determine mean differences in every tested group, followed by 50–100 nm, which was nearly consistent with the TEM and
Student-Newman-Keuls –q (SNK-q) test for multiple comparisons. images.
P,0.05 was considered to indicate statistical significance.
In vitro cumulative release profiles of Gentamicin from
Ethical approval nanoparticles are shown in Fig. 3a. The release profiles appeared
The animal experiments were approved by the Ethics Com- to have two phases. The first phase was a rapid release in the prior
period and about 70% were released in this phase (within the first
mittee of the Fourth Military Medical University. 6 h). The rapid releasing process was mainly due to the
nanoparticles surface drugs could easily diffuse in the initial time
Results and Discussion and the swelling of nanoparticles promoted drug release. The
second phase was a relatively slow release ranging from 6 to 24 h,
Characterization of polymer and nanoparticles and about 80% drug were released in this phase due to the
1H-NMR spectra of DEX-GMA were recorded with a Bruker swelling equilibrium and degradation of nanoparticles. The
dissolved drugs were diffused into the release medium. The
AC at 500 MHz with D2O as solvent. As shown in spectra cumulative percentage release of drug from nanoparticles was
(Fig. 2a), the signal from the proton at the anomeric carbon of the about 83% for 72 h.
a-1, 6 linkages at 4.9 ppm) was well separated from the multiplet
peaks of from 3.4 to 4.0 ppm. The typical peaks from the As shown in Fig. 3b, Gentamicin release rate of GNPs-CS/
methacryloyl group were observed at 1.95 ppm (methyl protons) as KGM has sustainable increase in 36 h and then drug release
well as at 5.9 and 6.2 ppm (protons at the double bond). became slower. The cumulative drug release rate was about 75%
for 72 h.
Poly (DEX-GMA/AAc) blank nanoparticles and Gentamicin
loaded nanoparticles were all obtained by free radical emulsion
polymerization. Encapsulation and loading efficiency were
89.9261.47% and 7.1560.16% calculated according to Genta-

PLOS ONE | www.plosone.org 8 July 2013 | Volume 8 | Issue 7 | e66890

Antibiotic Hemostatic First Aid Wound Dressing

Table 4. The Bacteriostatic ring diameter of C75K25 film, Drug loaded Poly (dex-GMA/AAc) nanoparticles and GNPs-CS/KGM
against different bacterial strain (values are mean 6 S.D., n = 6 observations in each group).

Group Staphylococcus aureus Escherichia coli Green copper
pseudomonas
C75K25 film 8.660.73 N
GNPs-CS/KGM 21.360.79 21.560.70 N
Drug loaded Poly (dex-GMA/AAc) 25.160.82 22.360.74 22.860.51
nanoparticles 22.460.89

doi:10.1371/journal.pone.0066890.t004

Characterization of KGM/CS blend film films in this study was found to be between 2282 to
3170 g?m22?day21, for instance, WVTR of CS film was
Fig. 4 shows the cross-section morphology for KGM/CS 2950 g?m22?day21 and C75K25 was 2282 g?m22?day21. As
blending film. The three-dimensional network structure of film shown in Table 1, the WVTR of CS/KGM films showed better
was observed very clearly, which was similar to our previous work. water uptake ability than that of CS itself due to good
The porous structure was benefit for cell adhesion and prolifer- hydrophilcity of KGM and large pore size of CS itself. And the
ation as well as promotes the growth of tissue. The porous layer results of WVTR showed that blend films obtained in our study
was arranged very regularly, which was benefit to separate could prevent water evaporation effectively and provided a good
components through the membranes and improve of the flux. moist environment for wound.
Porous C25K75 films had pores with an average diameter of
95616 mm and C50K50 film of 1067 mm. The pore size of CS Mechanical strength is one of indicators to evaluate materials
and C75K25 film were between them. KGM has stronger swelling mechanical properties. As shown in Fig. 5b, TS of the films
capacity which resulted in increased pore size while amino of CS increased significantly from 2.6 MPa to 4 MPa with increased
could form hydrogen bonds with hydroxy of KGM. This duality KGM composition and breaking elongation was about 10%. The
influence lead to the result that the tendency of increase with the results showed that mechanical strength of film was significantly
pore size appeared smaller before alonging with increasing enhanced by mixing KGM with CS. It was probably because
percentage of KGM. So when KGM/CS proportion became hydrogen bond created by –NH2 protonation of CS and hydroxyl
1:1, the pore size reached smallest. of KGM enhanced intermolecular forces and then increased
tensile strength of film.
To study the effects of KGM on blend film swelling, the degree
of swelling was plotted with different KGM/CS ratio coded as Degradation profiles of CS, C25K75, C50K50, and C75K25
K25C75, K50C50, and K75C25 for 72 h, as shown in Fig. 5a. It films are shown in Fig. 5c. It was revealed by the results that the
was noticed that all films has strong capability of water uptake in ratio of KGM has significantly influence on degradation rate of
3 h due to strong interaction between water molecules and the film. Degradation rate of blend films increased with increased
membrane containing OH groups and NH2 groups [37]. The KGM ratio. After 70 days, C25K75 film was almost completely
degree of swelling of blend films increased with more KGM due to degraded while residual mass of C50K50 and C75K25 film was
its good hydrophilicity. SDP of C25K75 reached (890636) %. about 40% and CS film was 80%.
After 18 h, swelling rate of blend films decreased due to dissolution
of KGM. Biocompatibility evaluation of GNPs-CS/KGM
The MTT assay is an indirect method of assaying cell growth
Water vapor transmission rate (WVTR) was one of the most
important indicators to evaluate the WVTR moisturizing perfor- and proliferation since the A490 values can be correlated to the
mance of wound dressing and hemostasis material. It was cell number. As the basis of cell growth, proliferation and
generally acknowledged that wound dressings with WVTR differentiation, cell attachment is an important measure to
between 2000 to 5000 g?m22?day21 could prevent excess evaluate the biocompatibility of biomaterials. To assess cellular
dehydration of wound surface, maintain wet environment of adhesion, Poly (dex-GMA/AAc) nanoparticles, C75K25 film and
wound and avoid excess accumulation of exudate [38]. WVTR of blank NPs-CS/KGM were seeded with the same density of human

Table 5. Hemostatic time and bleeding volume of different treatment group (values are mean 6 S.D., n = 6 observations in each
group).

Group hemostatic time (s) bleeding volume (g/kg)
2.1060.32*
Poly (dex-GMA/AAc) nanoparticles 256618* 1.4060.10**
C75K25 film 201617** 0.9560.20**
GNPs-CS/KGM 176614** 2.3860.48*
native CS film 24969* 1.4560.22**
Yunnan baiyao (positive control) 207612** 2.4260.47
Control (gauze) 360626
July 2013 | Volume 8 | Issue 7 | e66890
Values are significant (*) at P,0.05 and (**) at P,0.01as compared to the control group.
doi:10.1371/journal.pone.0066890.t005

PLOS ONE | www.plosone.org 9

Antibiotic Hemostatic First Aid Wound Dressing

fibroblasts. The cell viabilities measured by MTT assay of indicating that GNPs-CS/KGM could effectively promote con-
fibroblasts cultured on different materials were shown in Fig. 6. tractility of wound and has great potential as wound dressing
The OD value of each tested group had no obvious difference with materials. It was also revealed by HE staining that inflammatory
control group. Cell relative growth rate of tested groups were infiltration of GNPs-CS/KGM treatment group were apparent
higher than 90%, especially in C75K25 group with low less than control group (Fig. 7).
concentration, relative growth rate appeared to be higher than
100%. The MTT results revealed that Poly (dex-GMA/AAc) Antibacterial activity evaluation
nanoparticles, C75K25 film and blank NPs-CS/KGM had good Drug loaded Poly (dex-GMA/AAc) nanoparticles, C75K25 film
biocompatibility and no cytotoxicity.
and GNPs-CS/KGM were prepared in vitro for their microbial
Wound healing effects of KGM/CS blend films and GNPs- activity against Staphylococcus aureus, Escherichia coli, and
CS/KGM Green copper pseudomonas using disc diffusion method. It was
revealed by Fig. 8 that GNPs-CS/KGM and Drug loaded Poly
Wound healing is an interaction of a complex cascade of (dex-GMA/AAc) nanoparticles had strong inhibitory effect against
biochemical and cellular events that generates resurfacing, the bacteria mentioned above while C75K25 film only had
reconstitution and restoration of the tensile strength of injured inhibitory effect against Staphylococcus aureus. The diameter of
skin [39]. For evaluation of the wound healing capability of the Bacteriostatic ring was shown in Table 4.
preparations, percent wound contraction on incision wounds and
histopathological studies were measured. First, we studied wound Hemostatic activities evaluation
healing effects of blend film. As shown in Table 2, it was observed First, we compared the effects of Poly (dex-GMA/AAc)
that postoperative wound area had slightly contraction after 3d.
The cut began to scab after 1 week and the scab become detached nanoparticles, C75K25 film, GNPs-CS/KGM, native CS film
after 2 weeks with significant wound shrinkage. and Yunnan baiyao powder as positive control on heparinized rat
whole blood. In each case, we adjusted the polymer concentration
Healing of closed incisional wounds was also determined by the in the overall mixture to 0.25 wt%. Upon addition 3 kinds of
histopathological studies. Fig. 7 shows the histological studies on GNPs-CS/KGM, liquid blood is instantly transformed into a self-
different KGM/CS formulae, CS film and gauze control group. supporting gel, as can be seen from the photograph in Fig. 8 where
Granulation tissue of wound became thickening gradually along the sample holds its weight upon tube inversion. In comparison,
with the increasing healing time. It was revealed by HE staining the rest samples and blood remains a freely flowing liquid, as seen
that inflammatory infiltration of different degree could be from its corresponding photograph.
observed after 3d on all experiment and control group. Epidermal
cell layer of treatment group proliferated actively after 7d and According to the result above, we studied hemostatic time and
capillaries began to form in dermis. Hierarchical structure was bleeding volume of different treatment group. As shown in Table 5,
visible between epidermis and dermis. Epithelial structure such as GNPs-CS/KGM and Yunnan baiyao powder group has effec-
stratum basale and acanthosis cell layer was observed in CS and tively shorten hemostatic time and significantly decreased bleeding
C75K25 treatment group obviously, especially in C75K25 group, volume compared with control group (P,0.05). The results
cuticular layer was also apparently visible. However, in gauge indicated that the hemostatic effect of GNPs-CS/KGM and
control group, no clear dermal tissue structures were formed and Yunnan baiyao powder was superior to gauze and GNPs-CS/
there was no hierarchical structure between epidermis and dermis KGM has stronger hemostatic effect.
after 7d. After 14d, the photomicrographs for the section of
incision wound treated with treatment group especially C75K25 Conclusion
group showed significant hierarchical structure of epithelial tissue In this study, GNPs-CS/KGM was prepared as novel wound
covering the wound area together with remodeling of well-
developed collagen fibers that almost resembled normal tissue dressing with hemostatic capabilities and antibiotic effect. The
while in control group, there was still actively proliferated result revealed that microporous structure of CS/KGM has the
fibroblast in dermis. Cuticle, stratum granulosum, spinous cells ability of effectively haemostasis. GNP could one step further
layer, and basalis stratified clearly. Sections obtained from incision promote blood clotting due to its physicochemical properties.
wounds treated with C75K25 revealed almost complete healing GNPs-CS/KGM has the ability to transform whole liquid blood
with nearly full resolution of the granulation tissue, normal tissue into a gel, and it quickly stops bleeding from injuries in small
architecture, and new capillary distribution. It was known that CS animals. An additional important aspect is GNPs-CS/KGM has
itself has promoting wound healing effect and mixed with KGM significant wound healing effect and hemostatic properties.
could improve its mechanical properties. However, if the amount Besides, GNPs-CS/KGM obtained good antibiotics ability by
of KGM increased in CS/KGM, the solubility of film will increase addition of gentamicin in NPs. Thus, GNPs-CS/KGM may be
and film-forming property will decrease. So when the proportion considered promising biodegradable and biocompatible novel
between CS and KGM reached 75 to 25, the CS/KGM film has wound dressing with hemostatic capabilities as well as antibiotic
the best mechanical properties and film-forming property. That’s effect for treatment of external bleeding injuries.
why the histology study appears to favor the C75K25 films over
the other formulations. Author Contributions

Then GNPs-CS/KGM was also studied for wound healing Conceived and designed the experiments: HW LF. Performed the
effects. As shown in Table 3, contraction rate of GNPs-CS/KGM experiments: LF CC FL WL. Analyzed the data: LF CC. Contributed
has significantly increased compared with gauze group (P,0.05) reagents/materials/analysis tools: BR YQ. Wrote the paper: LF.

References monophosphate in an in situ-forming hydrogel wound dressing based on
oxidized alginate and gelatin. Biomaterials 27: 1355–1361.
1. Balakrishnan B, Mohanty M, Fernandez AC, Mohanan PV, Jayakrishnan A
(2006) Evaluation of the effect of incorporation of dibutyryl cyclic adenosine

PLOS ONE | www.plosone.org 10 July 2013 | Volume 8 | Issue 7 | e66890

Antibiotic Hemostatic First Aid Wound Dressing

2. Petrulyte S (2008) Advanced textile materials and biopolymers in wound 22. Dai T, Tanaka M, Huang YY, Hamblin MR (2011) Chitosan preparations for
management. Dan Med Bull 55: 72–77. wounds and burns: antimicrobial and wound-healing effects. Expert Rev Anti
Infect Ther 9: 857–879.
3. Corkhill PH, Hamilton CJ, Tighe BJ (1989) Synthetic hydrogels. VI. Hydrogel
composites as wound dressings and implant materials. Biomaterials 10: 3–10. 23. Zhang J, Xia W, Liu P, Cheng Q, Tahirou T, et al. (2010) Chitosan
modification and pharmaceutical/biomedical applications. Mar Drugs 8: 1962–
4. Perkins JG, Cap AP, Weiss BM, Reid TJ, Bolan CD (2008) Massive transfusion 1987.
and nonsurgical hemostatic agents. Crit Care Med 36: S325–339.
24. Li B, Li J, Xia J, Kennedy JF, Yie X, et al. (2011) Effect of gamma irradiation on
5. Losee JE, Fox I, Hua LB, Cladis FP, Serletti JM (2005) Transfusion-free the condensed state structure and mechanical properties of konjac glucoman-
pediatric burn surgery: techniques and strategies. Ann Plast Surg 54: 165–171. nan/chitosan blend films. Carbohydr Polym 83: 44–51.

6. Fernandes de Carvalho V, Paggiaro AO, Isaac C, Gringlas J, Ferreira MC 25. Huarong N, Xingxing S, Zhihua Z, Qingsong J, Yiwang C, et al. (2011)
(2011) Clinical trial comparing 3 different wound dressings for the management Electrospinning and characterization of konjac glucomannan/chitosan nanofi-
of partial-thickness skin graft donor sites. J Wound Ostomy Continence Nurs 38: brous scaffolds favoring the growth of bone mesenchymal stem cells. Carbohydr
643–647. Polym 85: 681–686.

7. Lo SF, Hsu MY, Hu WY (2007) Using wound bed preparation to heal a 26. Darmadji P, Izumimoto M (1994) Effect of chitosan in meat preservation. Meat
malignant fungating wound: a single case study. J Wound Care 16: 373–376. Sci 38: 243–254.

8. Pereira AL, Bachion MM (2005) Wound treatment: scientific production 27. Nakashima T, Matsuo M, Bin Y, Nakano Y, Kobayashi T, et al. (2006)
analysis published in the Revista Brasileira de Enfermagem from 1970-2003. Mechanical properties and antibacterial efficacy of chitosan films. Biocontrol Sci
Rev Bras Enferm 58: 208–213. 11: 27–36.

9. Gavriliuk BK, Gavriliuk VB (2011) Nanocomposite regeneration systems for 28. Huang L, Dai T, Xuan Y, Tegos GP, Hamblin MR (2011) Synergistic
wound healing. Biofizika 56: 1138–1141. combination of chitosan acetate with nanoparticle silver as a topical
antimicrobial: efficacy against bacterial burn infections. Antimicrob Agents
10. Sun G, Zhang X, Shen YI, Sebastian R, Dickinson LE, et al. (2011) Dextran Chemother 55: 3432–3438.
hydrogel scaffolds enhance angiogenic responses and promote complete skin
regeneration during burn wound healing. Proc Natl Acad Sci U S A 108: 29. Chen FM, Wu ZF, Sun HH, Wu H, Xin SN, et al. (2006) Release of bioactive
20976–20981. BMP from dextran-derived microspheres: a novel delivery concept. Int J Pharm
307: 23–32.
11. Laurens N, Koolwijk P, de Maat MP (2006) Fibrin structure and wound healing.
J Thromb Haemost 4: 932–939. 30. Hui Z, Hong W, Li F, Fei L, Chun-hu G, et al. (2008) Preparation and
characteristics of pH-sensitive derivated dextran hydrogel nanoparticles. Polym
12. Scherer SS, Pietramaggiori G, Matthews JC, Gennaoui A, Demcheva M, et al. composite 30: 1243–1250.
(2011) Poly-N-acetyl glucosamine fibers induce angiogenesis in ADP inhibitor-
treated diabetic mice. J Trauma 71: S183–186. 31. Virto MR, Elorza B, Torrado S, Elorza Mde L, Frutos G (2007) Improvement of
gentamicin poly(D,L-lactic-co-glycolic acid) microspheres for treatment of
13. Lenselink E, Andriessen A (2011) A cohort study on the efficacy of a osteomyelitis induced by orthopedic procedures. Biomaterials 28: 877–885.
polyhexanide-containing biocellulose dressing in the treatment of biofilms in
wounds. J Wound Care 20: 534, 536–539. 32. Hui Z, Chunhu G, Hong W, Li F, Fei L, et al. (2008) Immobilization of
derivatized dextran nanoparticles on konjac glucomannan/chitosan film as a
14. Dai C, Liu C, Wei J, Hong H, Zhao Q (2010) Molecular imprinted macroporous novel wound dressing. Biofactors 30: 227–240.
chitosan coated mesoporous silica xerogels for hemorrhage control. Biomaterials
31: 7620–7630. 33. Wittaya-areekul S, Prahsarn C (2006) Development and in vitro evaluation of
chitosan-polysaccharides composite wound dressings. Int J Pharm 313: 123–128.
15. Sugawara J, Suenaga K, Hoshiai T, Sato T, Nishigori H, et al. (2012) Efficacy of
Recombinant Human Soluble Thrombomodulin in Severe Postpartum 34. Ye X, Kennedy JF, Li B, Xie BJ (2006) Condensed state structure and
Hemorrhage With Disseminated Intravascular Coagulation. Clin Appl Thromb biocompatibility of the konjac glucomannan/chitosan blend films. Carbohydr
Hemost doi: 10.1177/1076029612443305. Polym 64: 532–538.

16. Connolly RJ (2004) Application of the poly-N-acetyl glucosamine-derived rapid 35. Tao Wang X-KZ, Xu-Ting Xue, Da-Yang Wu (2012) Hydrogel sheets of
deployment hemostat trauma dressing in severe/lethal Swine hemorrhage chitosan, honey and gelatin as burn wound dressings. Carbohydrate Polymers
trauma models. J Trauma 57: S26–28. 88: 75–83.

17. Zhang Y, Xu C, He Y, Wang X, Xing F, et al. (2011) Zeolite/polymer 36. Tsao CT, Chang CH, Lin YY, Wu MF, Wang JL, et al. (2011) Evaluation of
composite hollow microspheres containing antibiotics and the in vitro drug chitosan/-poly(glutamic acid) polyelectrolyte complex for wound dressing
release. J Biomater Sci Polym Ed 22: 809–822. materials. Carbohydr Polym 84: 812–819.

18. Laurienzo P, Malinconico M, Mattia G, Russo R, La Rotonda MI, et al. (2006) 37. Dhanuja G, Smitha B, Sridhar S (2005) Pervaporation of isopropanol-water
Novel alginate-acrylic polymers as a platform for drug delivery. J Biomed Mater mixtures through polyion complex membranes. Separation and Purification
Res A 78: 523–531. Technology 44: 130–138.

19. Daniela Pasqui MDCaRB (2012) Polysaccharide-Based Hydrogels: The Key 38. Kim IY, Yoo MK, Seo JH, Park SS, Na HS, et al. (2007) Evaluation of semi-
Role of Water in Affecting Mechanical Properties. Polymers 4: 1517–1534. interpenetrating polymer networks composed of chitosan and poloxamer for
wound dressing application. Int J Pharm 341: 35–43.
20. Persin KS-KZ, Foster TJ, van Dam JEG, Boeriu CG, Navard P (2011)
Challenges and opportunities in polysaccharides research and technology: The 39. Oberyszyn TM (2007) Inflammation and wound healing. Front Biosci 12: 2993–
EPNOE views for the next decade in the areas of materials, food and health 2999.
care. Carbohydr Polym 84: 22–32.

21. Rao SB, Sharma CP (1997) Use of chitosan as a biomaterial: studies on its safety
and hemostatic potential. J Biomed Mater Res 34: 21–28.

PLOS ONE | www.plosone.org 11 July 2013 | Volume 8 | Issue 7 | e66890

Article 3:

Non-Pneumatic Anti-Shock
Garment (NASG), a First-Aid
Device to Decrease Maternal
Mortality from Obstetric
Hemorrhage: A Cluster
Randomized Trial

Non-Pneumatic Anti-Shock Garment (NASG), a First-Aid
Device to Decrease Maternal Mortality from Obstetric
Hemorrhage: A Cluster Randomized Trial

Suellen Miller1*, Eduardo F. Bergel2, Alison M. El Ayadi1, Luz Gibbons2, Elizabeth A. Butrick1,
Thulani Magwali3, Gricelia Mkumba4, Christine Kaseba4, N. T. My Huong5, Jillian D. Geissler1,
Mario Merialdi5

1 Department of Obstetrics, Gynecology, and Reproductive Sciences, University of California San Francisco, San Francisco, California, United States of America, 2 Instituto
de Efectividad Cl´ınica y Sanitaria, Buenos Aires, Argentina, 3 Department of Obstetrics and Gynecology, University of Zimbabwe, Harare, Zimbabwe, 4 Department
of Obstetrics and Gynecology, University Teaching Hospital, Lusaka, Zambia, 5 The Department of Reproductive Health and Research of the United Nations
Development Programme/United Nations Population Fund/United Nations Children’s Fund/World Health Organization/World Bank Special Programme of Research,
Development and Research Training in Human Reproduction, World Health Organization, Geneva, Switzerland

Abstract

Background: Obstetric hemorrhage is the leading cause of maternal mortality. Using a cluster randomized design, we
investigated whether application of the Non-pneumatic Anti-Shock Garment (NASG) before transport to referral hospitals
(RHs) from primary health care centers (PHCs) decreased adverse outcomes among women with hypovolemic shock. We
hypothesized the NASG group would have a 50% reduction in adverse outcomes.

Methods and Findings: We randomly assigned 38 PHCs in Zambia and Zimbabwe to standard obstetric hemorrhage/shock
protocols or the same protocols plus NASG prior to transport. All women received the NASG at the RH. The primary
outcomes were maternal mortality; severe, end-organ failure maternal morbidity; and a composite mortality/morbidity
outcome, which we labeled extreme adverse outcome (EAO). We also examined whether the NASG contributed to negative
side effects and secondary outcomes. The sample size for statistical power was not reached; of a planned 2400 women, 880
were enrolled, 405 in the intervention group. The intervention was associated with a non-significant 46% reduced odds of
mortality (OR 0.54, 95% CI 0.14–2.05, p = 0.37) and 54% reduction in composite EAO (OR 0.46, 95% CI 0.13–1.62, p = 0.22).
Women with NASGs recovered from shock significantly faster (HR 1.25, 95% CI 1.02–1.52, p = 0.03). No differences were
observed in secondary outcomes or negative effects. The main limitation was small sample size.

Conclusions: Despite a lack of statistical significance, the 54% reduced odds of EAO and the significantly faster shock
recovery suggest there might be treatment benefits from earlier application of the NASG for women experiencing delays
obtaining definitive treatment for hypovolemic shock. As there are no other tools for shock management outside of referral
facilities, and no safety issues found, consideration of NASGs as a temporizing measure during delays may be warranted. A
pragmatic study with rigorous evaluation is suggested for further research.

Trial Registration: ClinicalTrials.gov NCT00488462

Citation: Miller S, Bergel EF, El Ayadi AM, Gibbons L, Butrick EA, et al. (2013) Non-Pneumatic Anti-Shock Garment (NASG), a First-Aid Device to Decrease Maternal
Mortality from Obstetric Hemorrhage: A Cluster Randomized Trial. PLoS ONE 8(10): e76477. doi:10.1371/journal.pone.0076477

Editor: Hany Abdel-Aleem, Assiut University Hospital, Egypt

Received May 15, 2013; Accepted August 21, 2013; Published October 23, 2013

Copyright: ß 2013 Miller et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: The project described was supported by Grant Number R01HD053129 from the National Institutes Of Child Health And Human Development.
Additional funding was provided by the Bill and Melinda Gates Foundation, Grant #48541. Technical and financial support was provided by the United Nations
Development Programme/United Nations Population Fund/United Nations Children’s Fund/World Health Organization (WHO)/World Bank Special programme of
research, development and research training in human reproduction (HRP/WHO). The content is solely the responsibility of the authors and does not necessarily
represent the official views of the National Institutes Of Child Health And Human Development or the National Institutes of Health, the Bill and Melinda Gates
Foundation, or the World Health Organization. Other than EFB, NTMH and MM, who are authors, HRP/WHO staff had no role in study design, data collection and
analysis, decision to publish, or preparation of the manuscript. The remaining funders had no role in study design, data collection and analysis, decision to
publish, or preparation of the manuscript. The corresponding author had full access to all the data in the study and had final responsibility for the decision to
submit for publication.

Competing Interests: The authors have declared that no competing interests exist.

* E-mail: [email protected]

Introduction shock, multiple organ dysfunction syndrome, and mortality.
Current obstetric and midwifery guidelines stress the use of
Obstetric hemorrhage (OH) is the leading cause of maternal uterotonics (oxytocin, misoprostol) to prevent postpartum hemor-
mortality, responsible for 25–50% of maternal deaths [1]. rhage (PPH) due to uterine atony [2,3]; however, even under
Uncontrolled hemorrhage can lead to irreversible hypovolemic randomized trial conditions, uterotonics can only reduce PPH by

PLOS ONE | www.plosone.org 1 October 2013 | Volume 8 | Issue 10 | e76477

NASG: A Cluster Randomized Trial

24% to 60% [4,5]. Uterotonics are also recommended for on randomized trials, although the quality of those trials was
treatment of atonic PPH, but they do not always stop hemorrhage. criticized [18].
Additional means of stopping atonic hemorrhage, such as balloon
tamponade, are currently being recommended [2], but not all OH Two quasi-experimental studies with pre-intervention phases
is due to atonic etiologies. Neither administration of uterotonics followed by NASG intervention phases demonstrated significantly
nor balloon tamponade will treat non-atonic OH (e.g. ruptured reduced measured blood loss [19–22], faster shock recovery [23],
uterus, ruptured ectopic pregnancy, vaginal/perineal lacerations, and decreased mortality (RR 0.56, 95% CI 0.35–0.89) [19] or
etc). Finally, if a woman bleeding from any OH etiology has lost so extreme adverse outcomes (severe morbidity and mortality
much blood that she has gone into shock, even if bleeding can be combined) (RR 0.32, 95% CI 0.16–0.63) [20] with the use of
controlled, the woman may still need blood transfusions. NASGs for women with severe OH (.1000 mL) and shock at
tertiary facilities [19–23]. A recent cost effectiveness analysis of
Blood transfusions and surgery are sometimes the only definitive that data at the tertiary level shows NASGs to be very cost effective
treatment for severe OH and hypovolemic shock, but they are for severe shock [24]. A systematic review of these quasi-
frequently only available at the highest level of the health system. experiments resulted in a World Health Organization (WHO)
For most women, accessing such care therefore relies on recommendation that the NASG be used as a temporizing
overcoming a series of delays that are associated with high rates measure while awaiting transfer [2], but noted that research
of maternal death in limited-resource settings: recognizing evaluating potential benefits, potential harm and use at the
complications, deciding to seek care, finding transport to care, primary level was needed [25].
and receiving quality comprehensive emergency obstetric care at
referral facilities [6,7]. Until recently, the only tools available at This study was conducted to evaluate whether NASG applica-
lower levels might be elevating the woman’s lower extremities and tion before transport from midwife-staffed primary health care
referral. Transport during referral can take hours, sometimes days. centers (PHCs) would result in reduced maternal mortality for
Upon arrival at a tertiary center (or even for women who begin women with hypovolemic shock secondary to OH, reduced time
hemorrhage in a tertiary facility), there may be long delays before to shock recovery, and if NASG use increased negative side effects.
blood transfusions can be arranged and completed. Zambia, with a maternal mortality ratio of 591 per 100,000 live
births [26], and Zimbabwe, with a maternal mortality ratio of 960
The Non-pneumatic Anti-Shock Garment (NASG) is a first-aid per 100,000 live births, were the settings for the study [27].
device that may assist women to survive delays in transport and
therefore receive definitive treatment (Figure 1). The NASG, made Methods
of neoprene and VelcroTM, compresses the lower body with nine
articulated segments closed tightly around the legs, pelvis, and Ethics Statement
abdomen. A foam ball in the abdominal segment increases Institutional review boards affiliated with the following institu-
compression (Figure 2). Circumferential compression reduces
vascular volume under the compressed areas, while expanding tions reviewed and approved study protocols and the informed
the central circulation by increasing preload, peripheral resistance, consent document and process: University of California, San
and cardiac output. Tamponade of abdominal, pelvic, and uterine Francisco (UCSF); University of Zambia, Lusaka; University of
vessels reduces blood loss [8–10]. Zimbabwe-UCSF Collaborative Programme on Health Research;
and the Department of Reproductive Health and Research of the
The NASG is ideal for OH for a number of reasons. The World Health Organization. Written informed consent (signature
abdominal panel stretches so that external uterine massage or from literate participants and thumb print from those who could
compression can be accomplished. The design permits perineal not sign) was obtained from all study participants who were
access for performing vaginal procedures (suturing lacerations, conscious and able to give consent. All ethics committees approved
manual exploration of the uterus and/or bimanual compression) a waiver of consent for unconscious women; written consent for an
or for inserting urinary catheters. Surgery can be performed by unconscious woman was either obtained from a relative and/or
simply opening the abdominal segment immediately prior to the patient after she regained consciousness. No data were entered
beginning surgery, and then replacing this segment when surgery for data collection forms that were not accompanied by a signed
is completed; removal of the device for surgery is not necessary. consent form. Figure 1, a photograph of a model in an NASG, was
The majority of the pressure exerted by the device is in the not taken during the study, but is a similar image used for
abdomen, retroperitoneum, and pelvis, reducing blood flow in OH illustrative purposes only. The subject of the photograph has given
immediately upon application [8,10–13]. The NASG is very written informed consent, as outlined in the PLOS consent form,
simple to apply and training in application is rapid. to publication of her photograph.

The NASG was developed in the 1970s by the United States Study Design
(US) National Aeronautics and Space Administration (NASA)/ To test the efficacy and safety of the NASG, a cluster
Ames; the original patent has expired. Although it has been used
in the US for pre-hospital lower body trauma, the device was not randomized control trial (CRCT) with PHCs as the cluster units
used for obstetric hemorrhage in limited-resource settings until was conducted in 38 PHCs referring to one of five Referral
2002 [14]. Unlike the predecessor device, the pneumatic anti- Hospitals (RHs) in Harare, Zimbabwe, and Lusaka and the
shock garment (PASG; also known as medical/military anti-shock Copperbelt Province, Zambia. (The protocol for this trial and
trousers, MAST), the NASG has no air bladders, tubing, or supporting CONSORT checklist are available as supporting
gauges, so it is much simpler to use and cannot be over-inflated information; see Checklist S1 and Protocol S1.) We chose a cluster
[15]. The PASG/MAST had a history of some adverse effects that randomized design because it is impossible to blind providers to
might be related to over-inflation [16,17]. The lower pressure the intervention or to develop a placebo garment. Furthermore, it
exerted by the NASG may not increase risks of negative effects is difficult to require providers to randomize individual patients
(dyspnea, decreased urine output, compartment syndrome, or once they have had the opportunity to use the NASG and see the
ischemia to compressed areas). Furthermore, the PASG has been apparent results (decreased bleeding and improved vital signs).
cited in a Cochrane Review as being either ineffective or perhaps
harmful for the pre-hospital treatment of trauma patients, based The University of California, San Francisco (UCSF), University
of Zambia, Lusaka, and University of Zimbabwe-UCSF Collab-

PLOS ONE | www.plosone.org 2 October 2013 | Volume 8 | Issue 10 | e76477

NASG: A Cluster Randomized Trial

Figure 1. Non-pneumatic Anti-Shock Garment (NASG) Photo. A model in an NASG.
doi:10.1371/journal.pone.0076477.g001

orative Programme on Health Research, Harare, conducted the communities. Because the NASG is visible, blinding of participants
study. The Department of Reproductive Health and Research of and clinicians/data collectors was impossible; the UCSF research
the UNDP/UNFPA/UNICEF/WHO/World Bank Special Pro- team was blind to outcomes.
gramme of Research, Development and Research Training in
Human Reproduction (HRP) served as the Data Coordinating Participants
Center (DCC). The Institute for Clinical Effectiveness and Health Clusters. Eligible PHCs were peri-urban, had at least 500
Policy (IECS), Buenos Aires, Argentina, conducted data monitor-
ing and analysis. All institutional review boards/ethical commit- annual deliveries conducted by midwives, and referred OH cases
tees approved the protocols. ($500 mL blood loss by visual estimation) to one of five study
RHs. Lusaka and Harare began in 2007; the Copperbelt Province
Because the sites were research na¨ıve, lacked baseline data, and was added in 2008. PHCs each covered the public assistance
clinicians were unfamiliar with NASGs, the study was implement- population in a given geographic catchment area. All PHCs had a
ed in three phases. Although the timelines were slightly different, maternity department, where midwives attended deliveries; any
all sites conducted each phase. The first phase, 2007–2008, was an woman .24 weeks gestation with bleeding would be seen in the
11-month preparatory phase, to familiarize the clinicians/data maternity department. Midwives were trained to provide prophy-
collectors with accurate form completion and adherence to lactic uterotonics, treat PPH with uterotonics and IV fluids, repair
standardized protocols. In the second phase, 10 months in first- and second-degree perineal lacerations, and refer any patient
2008–2009, we implemented the NASG at the RHs, to give with estimated blood loss $500 mL to the RH. Women with early
clinicians experience with it and to collect baseline outcome data. pregnancy bleeding of ,24 weeks gestation (ectopic pregnancy,
In the final phase, 2009–2012, the PHCs were randomized and complications of abortion, trophoblastic/molar pregnancy) were
NASGs implemented in the intervention group. seen in the outpatient department by either midwives or clinical
officers; these women in early pregnancy were also referred to the
Based on baseline data, a covariate-constrained randomization RH for bleeding $500 mL. No PHC was equipped to provide
procedure was used to ensure that intervention and control PHCs blood transfusion, surgery, or manual vacuum aspiration (MVA).
were balanced on number of deliveries, number of deliveries per All PHCs had access to a shared ambulance dispatch system to
midwife, distance to the RH, and proportion of OH cases request ambulance transfer for their patients to the designated
expected [28]. The DCC allocated the PHCs to intervention or RH.
control group; allocation assignment was known by PHC staff and
health authorities, but not by women in the PHC catchment

PLOS ONE | www.plosone.org 3 October 2013 | Volume 8 | Issue 10 | e76477

NASG: A Cluster Randomized Trial

Figure 2. Schematic of the NASG. This figure shows an opened NASG. The articulated leg segments (1, 2, 3) are attached to the pelvic segment
(4), and the abdominal segments (5, 6) contain a foam ball for extra pressure.
doi:10.1371/journal.pone.0076477.g002

Individual participants. Participants were included in a the study the NASG at the RH was a) by the start of the RCT in
cluster if they sought maternity care in a study PHC in their 2009, the RHs were using the NASG on OH patients in the
neighborhood. Participants were women with OH from any facility, b) the providers would find it difficult to refrain from
etiology and hypovolemic shock, with at least two of the following applying the NASG after seeing the benefits at the RH level, and
eligibility criteria: visually estimated blood loss $500 mL, pulse c) the goal of the study was to determine the efficacy when applied
$100 BPM, systolic blood pressure #100 mm Hg. Women with earlier in the OH/shock trajectory.
antepartum hemorrhage were excluded if the fetus was viable.
Participants were consented when they became eligible, if they All women received standard shock/hemorrhage protocol:
were conscious and able to give consent. All ethics committees oxygen, IV fluids, uterotonics/uterine massage (for uterine atony),
approved waiver of consent for unconscious women; consent for suturing of lacerations, manual removal of placenta or retained
an unconscious women was either obtained from a relative and/or tissues, MVA, surgery, and blood transfusion, as necessary. The
the patient after she regained consciousness. only differences in treatment received depended on hemorrhage
etiologies, e.g. ruptured ectopic pregnancies required surgery.
Interventions Upon RH arrival, all women had absorbent pads removed and
PHC level interventions. The main intervention was weighed to determine blood loss during transport and had a
calibrated blood measurement drape (Brass V Drape, Excellent
applying the NASG first-aid compression device (Zoex, Colma, Fixable Drapes, Madurai, India) placed. Staff at the RH removed
CA, USA). The NASG was rapidly applied, sequentially starting at the NASG when both criteria were met: the patient’s bleeding
the ankles, as the first step in shock resuscitation, and an absorbent decreased to ,50 mL per hour and the pulse was ,100 BPM for
perineal pad (Stay Dry Briefs, McKesson, San Francisco, CA, two hours. Removal was incremental, beginning at the NASG
USA) was applied for blood loss measurement. Control group ankle segments; vital signs were monitored for fifteen minutes to
patients also had absorbent perineal pads applied at study entry. ensure hemodynamic stability before proceeding to subsequent
Both groups received the same hemorrhage/hypovolemic shock segments.
protocol: intravenous (IV) fluids, uterotonics and uterine massage
(for uterine atony), and suturing of first- and second-degree Data Collected
lacerations. Data collected included reason for patient admission, age,

Management at the RH. Women in the control arm also gravidity, parity, weeks gestation, delivery information, prophy-
received the NASG upon RH arrival; those in the intervention lactic uterotonics, time hemorrhage started, treatment uterotonics,
arm remained in the NASG. The rationale for giving all women in IV fluids, blood transfusions, and hemostatic procedures and/or

PLOS ONE | www.plosone.org 4 October 2013 | Volume 8 | Issue 10 | e76477

NASG: A Cluster Randomized Trial

surgeries. Blood loss in the drape and urine output in urine Board (DSMB) reviewed the enrollment and outcome rates from
collection bags were recorded hourly, and vital signs were the baseline data collection phase and they changed the sample
recorded every 15 minutes from study entry to NASG removal. size target to 1944 women.
Mean Arterial Pressure (MAP) ,60 mmHg, ([(2 6 diastolic)+sys-
tolic]/3), defined more serious shock [29]. Mortality, morbidities, The DSMB performed power calculations in April 2011 and
diagnosis, and negative side effects that could potentially be February 2012. In both instances the DSMB noted that accrual
attributed to the NASG (respiratory distress, reduced urine output, was low, but recommended continuing the study and either
nausea, vomiting, abdominal pain) were recorded. In-facility increasing clusters or extending enrollment. However we were
clinicians/data collectors or study-funded midwives (when avail- unable to secure additional funding; enrollment ended May 2012,
able) recorded data. approximately 30 months post-randomization.

Data Management We undertook intention-to-treat analyses to compare treatment
Data forms were reviewed by study coordinators and checked groups with a pretest-posttest design, where the outcome rates
were measured at the cluster level prior to random assignment and
against medical records to resolve inconsistencies. Annually, a again after randomization and implementation of intervention. As
random sample of 10% of data forms was checked against medical pre-specified in the protocol, two sets of analyses were conducted,
records to confirm accuracy. Data were double-entered in both accounting for the cluster randomized study design: a) post-
OpenClinica (Akaza Research, Waltham, MA, USA), queried test observations only and b) post-test observations adjusted for
and cleaned, and analyzed using SAS (SAS Institute, Cary, NC, baseline measurements. We estimated random effects logistic
USA) and STATA (STATA Corp, College Station, TX, USA). regression models for binary outcomes. To adjust for the outcome
measurements at baseline, we included the logit of the cluster
Outcomes specific outcome rate as a covariate [34]. Diagnosis was entered in
The primary outcomes of the trial were the frequency of the regression model to adjust for an imbalance among
participants recruited after randomization. The effect size was
maternal mortality; survival with severe maternal morbidity; and reported as OR with 95% CIs. For continuous outcomes, a
extreme adverse outcome (EAO), a composite outcome of the two. random effects linear regression model was estimated. Measured
Mortality was defined as dying before hospital discharge, as blood loss values were transformed into the log metric for
women were not followed up after discharge. Severe morbidity normality, and the ratio of the geometric mean and its 95% CI
was defined as end-organ failure (cardiac, pulmonary, renal, were reported. To compare SI recovery trajectories, Cox
cerebral) persisting 24 hours or more beyond shock resuscitation regression models were estimated to evaluate group differences
[30]. Patients who were lost to follow-up between the PHC and accounting for study design effects by including a working
the RH were tracked to determine if there were outcome data on correlation matrix to adjust the standard errors. Statistical tests
mortality. (See Text S1 for a detailed description of the lost to were 2-sided and performed at the 5% significance level.
follow-up protocol).
This study is registered with ClinicalTrials.gov number,
Secondary outcomes included median blood loss measured by NCT00488462.
weighing the absorbent pad(s) upon RH admission; blood loss
measured in the drape at the RH; blood loss during surgery; Results
frequency of emergency hysterectomy for intractable uterine
atony; and time to recovery from shock, defined as return to Shock Figure 3 shows the trial profile. Fifty-five clinics were assessed
Index (SI) ,0.98 (SI = Heart Rate/Systolic Blood Pressure) [31]. for eligibility; thirty-eight met criteria and were included in the
Negative effects that might be attributable to the NASG included baseline period and subsequently randomized. During the baseline
decreased urine output, respiratory difficulties, nausea, vomiting, period, 114 women and 99 women were enrolled in what would
and abdominal pain. become, after the clinics were randomized, the intervention and
control groups respectively. After randomization, 548 women at
Co-interventions included: IV fluids, blood transfusions, receipt the control PHCs were assessed for eligibility and 482 were
of the NASG at the RH, and duration of NASG use. allocated to the control group; 445 women at the intervention
PHCs were assessed for eligibility and 405 were allocated to the
Statistical Methods intervention group. No clusters were lost to follow-up and all
Sample size was estimated in Acluster (Metaxis, Inc., Vista, CA) clusters enrolled participants. Seven women were lost to follow-up
in the control group; none in the intervention group. (See Text S1
using the incidence rate of the primary composite EAO, based on for a detailed description of the lost to follow-up protocol.) Among
an NASG pilot study conducted in Nigeria in 2005 (9% incidence, women in the intervention group, 366 (90%) received the NASG
50% effect size reduction) [32]. A sample size of 2400 women was at the PHC; none received it at the PHC in the control group.
calculated based on a reduction in incidence of EAO from 9.0% to
4.5% in EAOs at 20 clinics, of varying sizes, 80% power, two-sided PHC and Participant Characteristics
type 1 error rate of 5%, and an intra-cluster correlation coefficient As shown in Table 1, PHC characteristics measured at baseline
of 0.01 [33]. Achieving this target required enrolling approxi-
mately 3.3 women per clinic per month over 3 years, which we felt were comparable between groups. Among women recruited
was possible based on an assessment of delivery rates and reported during the baseline period, mortality was higher in the clinics
OH rates in Harare and Lusaka conducted while writing the later randomized to the intervention group. Other baseline
proposal. variables were similar, including diagnoses. Women in both
groups enrolled during the intervention period (after randomiza-
Initially, 12 clusters in Lusaka and 12 clusters in Harare were tion) were similar, with the exception of diagnosis: abortion and
included with a planned enrollment period of three years. During placental abruption were more common in the control group;
the baseline period, an additional 14 clusters in the Copperbelt uterine atony, retained placenta, placenta accreta, and lacerations
Province, Zambia were added in 2008 because the accrual rate were more common in the intervention group (Table 2). Women
was lower than expected, mainly due to lower than expected in the intervention group spent a median of 100 minutes (IQR 70–
incidence of OH/shock and lower incidence of EAOs (5%). In
2008, prior to randomization, the Data Safety and Monitoring

PLOS ONE | www.plosone.org 5 October 2013 | Volume 8 | Issue 10 | e76477

NASG: A Cluster Randomized Trial

Figure 3. Cluster Randomized Trial Design.
doi:10.1371/journal.pone.0076477.g003

135) in the NASG between study enrollment and arrival at the the RH was 375 for women in the intervention group compared to
RH, women in the control group spent a median of 110 minutes 420 for women in the control group (OR 0.91, 95% CI 0.77–1.08,
(IQR 78–155) between study enrollment and being placed in the p = 0.28).
NASG at the RH (not shown). Median transfer time was 40
minutes for the intervention group (IQR 30–60) and 49 minutes Outcomes
for the control group (IQR 33–67) (not shown). Outcomes for the intention-to-treat analysis are in Table 4. The

Co-Interventions intervention was associated with a 46% reduction in the odds of
Results on co-interventions: IV fluids, blood transfusions, and mortality (OR 0.54, 95% CI 0.14–2.05, p = 0.37) and 54%
reduction in the odds of the composite EAO (OR 0.46, 95% CI
receipt of NASG at RH (Table 3), show a significant difference 0.13–1.62, p = 0.22); these differences were not statistically
only in rapidity of receipt of blood transfusions, with more women significant. The results were similar after adjusting for the outcome
in the NASG receiving blood within one hour of arrival at RH rates measured during the baseline period (mortality: AOR 0.55,
(OR 3.21, 95% CI 1.23–8.35, p = 0.02). However, the overall rate 95% CI 0.14–2.18, p = 0.40; EAO: AOR 0.46 95% CI 0.13–1.67,
of transfusions between groups was similar; 42% intervention vs. p = 0.24), and after adjusting for both the rate of outcomes at
38.6% control (OR 1.34, 95% CI 0.94–1.92, p = 0.11). Of the baseline and definitive diagnosis imbalance, mortality: AOR 0.47
women in the control group, 13.2% did not receive the NASG at (95% CI 0.12–1.87, p = 0.28) and EAO: AOR 0.39 (95% CI 0.11–
the RH and 18.8% of women in the intervention group who had 1.44, p = 0.16) (not shown).
not received the NASG at the PHC also did not receive it at the
RH (not shown). The median number of minutes in the NASG at Only one secondary outcome was statistically significant. The
median time in minutes to return to SI ,0.98 (HR 99/SBP101)

PLOS ONE | www.plosone.org 6 October 2013 | Volume 8 | Issue 10 | e76477

NASG: A Cluster Randomized Trial

Table 1. Clinic and Women’s Characteristics and Outcome Assessment during Baseline Period.

Intervention Group Control Group

(Number of (Number of
clinics = 19) clinics = 19)

n (%) n (%)

Clinic characteristics at Volume of births Low (,1235 births/yr) 5 (26%) 5 (26%)
baselinea
9 (48%) 10 (53%)
Medium (1235–2751 births/yr) 5 (26%) 4 (21%)
2 (11%) 4 (21%)
High (.2751 births/yr) 13 (68%) 11 (58%)
4 (21%) 4 (21%)
Nu of midwives Low (,8) 12.19 (5.22) 11.84 (6.25)
114 99
Medium (8–15)
26.8 (5.4) 26.6 (5.7)
High (.15) 2 (1–3) 2 (1–4)
36.8 (3.0) 37.1 (2.4)
Characteristics of women at Distance to RH (km)b 15/113 (13.3%) 18/99 (18.2%)
baseline Nu of women 25/113 (22.1%) 31/99 (31.3%)
33/113 (29.2%) 28/99 (28.3%)
Ageb,c 22/113 (19.5%) 16/99 (16.2%)
Parityd,e 10/113 (8.9%) 4/99 (4.0%)
Gestational ageb,f 4/113 (3.5%) 0/99 (0.0%)
3/113 (2.7%) 0/99 (0.0%)
Diagnosis Complications of Abortion 0/113 (0.0%) 1/99 (1.0%)
1/113 (0.9%) 0/99 (0.0%)
Postpartum Uterine Atony 0/113 (0.0%) 0/99 (0.0%)
0/113 (0.0%) 1/99 (1.0%)
Retained Placenta 600 (450–800) 500 (350–600)
20/114 (17.5%) 25/96 (26.0%)
Lacerations/Genital Trauma 1/114 (0.9%) 0/99 (0.0%)
0/114 (0.0%) 1/99 (1.0%)
Placental Abruption
6/114 (5.3%) 3/99 (3.0%)
Placenta Previa 6/114 (5.3%) 4/99 (4.0%)

Ruptured Uterus

Ectopic Pregnancy

Placenta Accreta

Molar Pregnancy

Other
Estimated REVEALED blood loss at study entry (mL)d
MAP,60 at study entryg

Outcome assessment at Unconscious at study entry
baseline Survived with severe morbidityh

Mortality

Extreme adverse outcome

aVariables used for randomization, using Phase 1 clinic statistics.
bMean (Standard Deviation).
cNu of women in the Intervention Group: 113; Nu of women in the Control Group: 99.
dMedian (Interquartile Range) is reported.
eNu of women in the Intervention Group: 104; Nu of women in the Control Group: 90.
fFor those $24 weeks; does not include cases with molar or ectopic pregnancies or abortion. Nu of women in the Intervention Group: 92; Nu of women in the Control

Group: 69.
gMAP was measured as ([(2 * diastolic BP)+systolic BP]/3); includes women with non-palpable BP.
hIncludes acute renal failure, acute respiratory distress syndrome, heart failure, cerebral impairment (seizures, unconsciousness, motor/cognitive loss) among women

who survived.

doi:10.1371/journal.pone.0076477.t001

was 170 minutes (IQR 96–299) for the intervention group vs. 209 Discussion
minutes (IQR 114–386) for the control group (HR 1.25 (95% CI
1.02–1.52, p = 0.03), indicating a 25% faster rate of recovery for Women with hypovolemic shock secondary to OH at the PHC
women in the intervention group. transported to an RH in the NASG had a non-significant
unadjusted 46% lower mortality, and 54% lower composite
There were no significant differences by group for women who EAO, than women in the control group, with an adjusted 61%
experienced negative side effects at the PHC. At the RH, women decrease in EAO. Women in the intervention group also had a
in the intervention group were more likely to have reported 25% statistically significant reduction in time to lower Shock
abdominal pain (OR 1.96, 95% CI 1.12–3.45, p = 0.02) (Table 5). Index. These results suggest a treatment benefit from earlier
No women required removal of the NASG because of pain.

PLOS ONE | www.plosone.org 7 October 2013 | Volume 8 | Issue 10 | e76477

NASG: A Cluster Randomized Trial

Table 2. Women’s Characteristics Enrolled during Intervention Period.

Characteristics of women enrolled during Nu of women Intervention Control Group
intervention period Group n (%)
482
n (%)

405

Agea,b 26.9 (5.9) 27.3 (6.3)
2 (1–3) 2 (1–3)
Parityc,d 37.7 (2.6) 37.4 (2.9)
73/405 (18.0%) 177/478 (37.0%)
Gestational agea,e 163/405 (40.3%) 146/478 (30.5%)
97/405 (24.0%) 85/478 (17.8%)
Diagnosisf Complications of Abortionk 51/405 (12.6%) 34/478 (7.1%)
7/405 (1.7%) 19/478 (4.0%)
Postpartum Uterine Atonyj 3/405 (0.7%) 5/478 (1.1%)
3/405 (0.7%) 5/478 (1.1%)
Retained Placentai 2/405 (0.5%) 5/478 (1.1%)
5/405 (1.2%) 0/478 (0.0%)
Lacerations/Genital Traumaj 1/405 (0.3%) 2/478 (0.4%)
500 (480–700) 500 (500–800)
Placental Abruptioni 129/399 (32.3%) 149/475 (31.4%)
11/403 (2.7%) 13/477 (2.7%)
Placenta Previa

Ectopic Pregnancy

Ruptured Uterus
Placenta Accretai

Molar Pregnancy
Estimated REVEALED blood loss at study entry (mL)c,g
MAP,60 at study entryh

Unconscious at study entry

Note: Wilcoxon Rank Sum test utilized to test all continuous variables due to non-normality. Chi-square test used for categorical values except where noted.
aMean (Standard Deviation).
bNu of women in the Intervention Group: 404; Nu of women in the Control Group: 476.
cMedian (Interquartile Range) is reported.
dNu of women in the Intervention Group: 404; Nu of women in the Control Group: 472.
eFor those $24 weeks; does not include cases with molar or ectopic pregnancies or abortion. Nu of women in the Intervention Group: 291; Nu of women in the Control

Group: 250.
fFisher’s exact test used for categorical values.
gNu of women in the Intervention Group: 391; Nu of women in the Control Group: 447.
hMAP was measured as ([(2 * diastolic BP)+systolic BP]/3); includes women with non-palpable BP.
iSignificant difference between intervention groups where p,0.05.
jSignificant difference between intervention groups where p,0.01.
kSignificant difference between intervention groups where p,0.001.

doi:10.1371/journal.pone.0076477.t002

application of the NASG. As expected, other secondary outcomes recovery in both the NASG group and the control group (170
were not different, as all women received the NASG at the RH. minutes and 209 minutes, respectively) occurred before receipt of
There were no significant negative effects from NASG use at the blood transfusions (229 minutes and 260 minutes, respectively).
PHCs, but an increase in abdominal pain was reported at the RH.
Strengths of this study include that it was conducted in settings
The difference in median length of time in the NASG between where the NASG was more likely to have a large impact. In
groups was only 55 minutes. The median time in NASG for limited-resource settings with high maternal mortality, PHCs and
women in the intervention group was 100 minutes before RH RHs are busy, understaffed, and often characterized by delays in
arrival and 375 minutes after arrival (475 total), while the women transport and time to receipt of definitive therapies. Conducting a
in the control group were in the NASG at the RH for 420 minutes. methodologically rigorous randomized trial under these conditions
These results suggest that it is not the length of time, but the earlier was a challenge, but doing so provides results that can be applied
application that affected the outcomes. We also interpret the to similar settings. A further strength of this study was having
significantly more rapid recovery of the SI to ,0.98 to be a result mortality as an outcome. The reduction in maternal mortality
of earlier application of the NASG. An alternative explanation observed in this trial is uniquely high for a single intervention.
could be the more rapid administration of blood transfusions, as a Measuring maternal mortality is difficult and it is rarely used as an
higher proportion of the intervention group received a blood outcome [35,36]. A review of maternal health intervention studies
transfusion in the first two hours after RH admission. However, showed only four with a mortality outcome [37–40]. Only one was
there was no difference between the two groups in the median a trial of a single medical intervention, the MAGPIE trial, a
time from study entry or RH entry to time of blood transfusion CRCT of magnesium sulphate for pre-eclampsia, which showed a
(Table 3). It is more likely that the observed effect was due to the non-significant 45% mortality reduction (RR 0.55, CI 0.26–1.14)
earlier application of the NASG than to the blood transfusions. [38]. Another strength of the trial was a high rate of follow-up on
Further supporting evidence of this explanation is that median our primary outcome; ,1% of individuals were missing data on
recovery time to SI in both groups was 50–60 minutes before the mortality. However, missingness was higher for some secondary
median time to blood transfusion. The median time to shock outcomes.

PLOS ONE | www.plosone.org 8 October 2013 | Volume 8 | Issue 10 | e76477

NASG: A Cluster Randomized Trial

Table 3. Co-interventions at the Referral Hospital.

Intervention Control Group Odds Ratio P-value
Group
0.28
Minutes in NASG at RHa,b (n = 405) (n = 475) (95% CI) 0.41
n/N (%) n/N (%) 0.49
Women receiving .1500 mL of IV fluids within 1 hour of study 375 (240–588) 420 (280–683) 0.91 (0.77–1.08)c 0.02
admissiond 59/402 (14.6%) 57/456 (12.5%) 1.20 (0.78–1.87) 0.04
0.11
Women with uterine atony who received uterotonics within 84/161 (52.2%) 68/138 (49.3%) 1.32 (0.60–2.86) 0.23
1 hour of study admission 0.65
32/398 (8.0%) 19/435 (4.4%) 3.21 (1.23–8.35)
Women receiving blood transfusion within
1 hour of hospital admission 88/398 (22.1%) 73/435 (16.8%) 1.98 (1.02–3.86)

Women receiving blood transfusion within 2 hours of hospital 167/398 (42.0%) 168/435 (38.6%) 1.34 (0.94–1.92)
admission 117.5 (75–265) 135 (90–270) 0.78 (0.52–1.17)c
229 (165–380) 260 (195–420) 0.93 (0.67–1.29)c
Women receiving blood transfusion ever
Time to blood transfusion from RH arrivala,e
Time to blood transfusion from study entrya,f

aFor each group the median and the interquartile range is reported.
bNu women in the Intervention Group: 381; Nu women in the Control Group: 268.
cFor the estimation of the effect the variable was transformed into the log metric for normality and the ratio of the mean is reported.
dThe protocol asked for 1500 mL to be administered in the first hour of resuscitation.
eNu women in the Intervention Group: 166; Nu women in the Control Group: 163.
fNu women in the Intervention Group: 166; Nu women in the Control Group: 162.

doi:10.1371/journal.pone.0076477.t003

A major limitation to the study was low accrual. The lack of trials in which these rates were .3% [19,21], mainly renal failure.
statistical significance on the primary outcome may be due to a A lack of nephrology units and dialysis may have contributed to a
smaller than expected number of women with OH at PHCs and a lack of survival with this morbidity. An alternative explanation
lower than expected event rate. (See Text S2 for more information could be that the training efforts made to assure protocol
on low accrual.) The low incidence of severe end-organ failure adherence improved patient outcomes, not an uncommon finding
maternal morbidities (,0.2%) was not consistent with previous

Table 4. Primary and Secondary Outcomes.

Intervention Control Group ICC Odds Ratio P-value
Group (95% CI)
(n = 475) –
Primary Outcomes Survived with severe morbiditya (n = 405) 0.022 – 0.37
n/N(%) 0.019 0.54 (0.14–2.05) 0.22
Mortality n/N(%) 0.46 (0.13–1.62) 0.75
1/465 (0.2%) 1.04 (0.80–1.36)d 0.30
Extreme adverse outcome 0/403 (0.0%) 11/475 (2.3%) 1.31 (0.79–2.16)d 0.73
Secondary Outcomes Blood loss in transit (mL)b,c 4/405 (1.0%) 12/465 (2.6%) 1.06 (0.76–1.49)d –
4/403 (1.0%) 218 (95–461) – 0.03
Blood loss after arrivalc,e 205 (105–405) 50 (30–150) 1.25 (1.02–1.52)i
Total blood lossc,f 60 (30–280) 336 (145–599)
Emergency hysterectomyg 355 (160–655) 0/296 (0.0%)
Minutes to normal Shock Indexh 1/240 (0.4%) 209 (114–386)
170 (96–299)

aIncludes acute renal failure, acute respiratory distress syndrome, heart failure, cerebral impairment (seizures, unconsciousness, motor/cognitive loss) among women

who survived.
bThe blood loss in transit was measured in 155 women in the Intervention Group and 175 women in the Control Group.
cFor each group the median and the interquartile range is reported.
dFor estimation of the effect the variable was transformed into the log metric for normality and the ratio of the mean is reported.
eBlood loss after arrival was measured in 267 women in the Intervention Group and 269 women in the Control Group at post-randomization.
fTotal blood loss was measured in 125 women in the Intervention Group and 123 women in the Control Group at post-randomization.
gHysterectomy among women with diagnosis of uterine atony and complications of abortion.
hShock Index was calculated by (heart rate/systolic blood pressure). For each group the median and interquartile range and hazard ratio is reported. Shock Index was

measured in 326 women in the Intervention Group and 358 women in the Control Group at post-randomization.
iHazard ratio.

doi:10.1371/journal.pone.0076477.t004

PLOS ONE | www.plosone.org 9 October 2013 | Volume 8 | Issue 10 | e76477

NASG: A Cluster Randomized Trial

Table 5. Side Effects.

At the clinic Respiratory symptoms/dyspnea Intervention Group Control Group Odds Ratio P-value
At the RH Abdominal pain
Nausea (n = 405) (n = 475) (95% CI) 0.12
Vomiting 0.24
Respiratory symptoms/dyspnea 33/382 (8.7%) 23/420 (5.5%) 1.68 (0.88–3.22) 0.85
Reduced urine output 190/383 (49.6%) 243/426 (57.0%) 0.77 (0.50–1.19) 0.27
Abdominal pain 26/383 (6.8%) 35/421 (8.3%) 0.93 (0.45–1.94) 0.83
Nausea 15/382 (3.9%) 27/422 (6.4%) 0.66 (0.31–1.38) 0.74
Vomiting 12/394 (3.1%) 19/437 (4.4%) 0.90 (0.34–2.36) 0.02
2/394 (0.5%) 3/438 (0.7%) 0.74 (0.12–4.46) 0.49
213/395 (53.9%) 179/437 (41.0%) 1.96 (1.12–3.45) 0.15
9/393 (2.3%) 7/437 (1.6%) 1.45 (0.50–4.26)
11/393 (2.8%) 6/437 (1.4%) 2.13 (0.76–5.97)

doi:10.1371/journal.pone.0076477.t005

in trials conducted over time [41]. (See Table S1 for an or a difference in surgeries and/or pain medications administered
explanation of study-sponsored trainings). for post-surgical patients. Since the women in the intervention
group actually had the NASGs on for less time at the RH (375
Although the sample size was not reached, our results are minutes vs. 420 minutes), we are unsure of the cause of the
consistent with the hypothesized .50% reduction in EAOs and increased pain or what to do to alleviate it. This should be followed
significantly more rapid decrease in SI among those treated with up in any future studies.
the NASG, with no increase in adverse health effects. These
findings are similar to prior findings of the NASG at the tertiary Another limitation was the imbalance in hemorrhage etiologies,
level [19,21,23]. which is difficult to explain given the balanced randomization and
equal distribution of etiologies at baseline. The most significant
A finding that was not expected was a lack of a statistically differences appear to be the increased enrollment of postpartum
significant difference in measured blood loss in transit (205 mL hemorrhage (PPH) etiologies in the intervention group (uterine
NASG vs. 218 mL control); measured blood loss has consistently atony, retained placenta, genital lacerations, and placenta accreta).
been statistically significant in the quasi-experimental trials at the As noted in the methods and results sections, we conducted an
tertiary level [19–22]. However, 62% of data was missing on blood adjusted analysis for the imbalance and found an even greater
loss in transit, and 72% of women had no total blood loss protective effect for the NASG. Inclusion criteria were the same at
recorded. all PHCs and in all wards at the PHCs. However, the staff at
intervention clinics had to do a new procedure, application of the
The two primary limitations to this study, lack of reaching NASG; this may have been more strictly adhered to in the
sample size for power and high rate of missing values for blood loss maternity wards than in the outpatient departments where
in transit data, should be addressed. As we note in (Text S2), the abortion patients were seen. It is also possible that the acuity of
rate of severe hemorrhage with hypovolemic shock was lower than the PPH patients was perceived as more severe.
we had anticipated. Given the increasing global practice of
prophylactic uterotonics and improved management of early PPH, Despite these differences in etiologies, however, the most
we feel that obtaining a sample size adequate to demonstrate a important prognostic factor, condition on study entry, was similar
statistically significant decrease in maternal mortality, even with between study groups, with approximately 32% of women with
the large effect size of the NASG, in a cluster randomized trial, is MAP,60 mmHg [19,44]. The overall effect of the imbalance in
unattainable. We do not feel that this type of randomized efficacy etiologies was recruitment of a population at higher risk in the
trial with detailed data collection will be able to be repeated in a intervention group, which might have negatively biased the
period of time that would be supportable by the majority of comparison. However, adjustment for these etiologies post-
funders. We therefore recommend more pragmatic trials with a randomization did not diminish the effect size. In fact the AOR
rigorous evaluation component. The other primary limitation, the for EAO was 0.39 (0.11–1.44, p = 0.16), actually strengthening our
high rate of missing values for blood loss in transit, could be results, since the PPH etiologies are more likely to be associated
overcome in future studies by greater attention to weighing pads with extreme adverse outcomes in this sample.
on arrival at the RH, and more careful calibration and use of
electronic scales. While the amount of blood loss is a secondary The slightly more rapid time between study enrollment and
outcome, it is also a proxy for the more crucial mortality outcome, arrival at RH (100 minutes vs. 110 minutes) for the intervention
and therefore understanding the effect of the NASG on blood loss group and the more rapid receipt of blood transfusions may reflect
during transit would still be of interest. the NASG as a visual cue indicating severity and need for action.
The low level of all women in the sample receiving blood
We did not expect to observe a statistically significant increase transfusions (,40%) does not necessarily reflect their condition;
in abdominal pain at the RH. This was inconsistent with a prior blood transfusions may be difficult to obtain in these settings, and
quasi-experimental study that found no difference between study the amount of blood transfused (or even ever receiving a
groups in experience of abdominal pain [42,43]. While it can be transfusion) may be more a reflection of blood availability than
unpleasant, abdominal pain is not a safety concern. The reason for patient need.
the increase in reports of abdominal pain at the RH level in the
intervention group is unclear. Perhaps the increase in this study The generalizability of the findings is limited by the specific
might be related to different distribution of hemorrhage diagnosis, settings in Zambia and Zimbabwe and the nature of a research

PLOS ONE | www.plosone.org 10 October 2013 | Volume 8 | Issue 10 | e76477

NASG: A Cluster Randomized Trial

project compared to real-world clinical care. While the individual students and interns who supported the project and did special
women are most likely similar to other women using public sub-projects. We thank our Data and Safety Monitoring Board
facilities in sub-Saharan Africa, the clusters may be different; most including Jose Belizan, Fernando Althabe, Shrikant Bangdiwala,
of the clinics were peri-urban, referred to teaching facilities, and Fridah Kazembe, and Anna Colletor-Penduka for their hard work
staffed by midwives, nurses, or clinical officers. While the NASG is and support. Also thanks to Daniel Giordano of the Centro
simple to apply, the training for the project also included frequent Rosarino de Estudios Perinatales for initial database set up and to
reviews of evidence-based protocols on prevention, identification, Sheri Lippman of UCSF for analysis and data cleaning oversight.
and management of OH/shock. Furthermore, there was supervi-
sion and protocol reinforcement associated with research. Supporting Information

Despite the lack of statistical significance, these findings on the Table S1 Study Sponsored Training Content and Sched-
primary outcome and the significantly faster shock recovery ule. Study-sponsored trainings were conducted at the beginning
suggest that as a first-aid device, there might be a treatment benefit of each of the three study phases. In addition, after randomization
for NASG use at the PHC level, with no risk of safety issues. annual update/refresher trainings were held in each city (Harare,
Concerns expressed by clinicians, based on previous experience Lusaka, Kitwe or Ndola [Copperbelt cities] ) for all RH and PHC
with PASGs about safety (that the NASG might exert too much staff, and, as necessary, PHC-specific or unit-specific refresher
pressure, or increase the risk of anuria/oliguria or dyspnea), do not trainings were held in places where enrollment was low or multiple
appear to be an issue. Currently there are no other tools available errors in filling out data collection forms were reported. This table
to stabilize women with hypovolemic shock and severe OH until contains an outline of the various training program topics, their
definitive therapies can be reached and administered. Therefore, scheduling during the study phases, and their intended audience(s).
these results could be helpful in informing policy and clinical (PDF)
decisions to incorporate the NASG into health systems seeking to
overcome delays contributing to maternal mortality. Text S1 Lost to Follow-Up, Tracking Participants
between PHCs and RHs. Cases were considered lost to
Given the potential clinical benefit to application of NASG to follow-up if there was no data on the primary outcome, mortality.
women suffering hypovolemic shock, policy makers may be Difficulties in communication between the PHCs and the RHs
interested in investing in NASGs for their maternal health systems. resulted in 7 of 887 total women (0.8%) having no primary
Our understanding of the current treatment and prevention outcome data available, and being considered lost to follow-up for
options for PPH indicate that a holistic approach to maternal the analysis, after the tracking methods described in this
mortality reduction would include: investments in misoprostol for supplement were utilized.
prevention or treatment where safe oxytocin injection is not (PDF)
possible; referral and transport strengthening to enable community
level, PHC, or BEmOC facilities to transport women in shock to Text S2 Low Accrual. Our estimated sample size was based on
CEmOC facilities; and the application of the NASG prior to a pilot study conducted in Nigeria in 2004–2006 which found an
transport. NASGs could also be placed on ambulances when adverse outcome rate of 9.0%. Based on a reduction from 9.0% to
hemorrhaging women are picked up for transport. If the NASG 4.5% in incidence of adverse outcomes, 80% power, two-sided
has not been applied prior to transport, then it could be applied at type I error of 5%, and an intra-cluster correlation coefficient of
those CEmOC facilities that have delays. The NASG plays a 0.01, we calculated a sample size of 2400 women. By the time data
unique role in hemorrhage and hypovolemic shock management. collection began, mass forced migration of the peri-urban areas
If a woman is given prophylaxis but still hemorrhages, or she has served by our study sites and uptake of pre-delivery uterotonic
an etiology that is not responsive to uterotonics, the NASG might prophylaxis resulted in reductions in both the pool of delivering
reverse shock more quickly and contribute to her surviving during women and the actual incidence of PPH. Attempts to increase
transport or delays at referral centers. Therefore, systems should enrollment by adding 16 study sites in the Copperbelt region in
not consider investing in either uterotonics or the NASG, but in 2008 did not result in increased clinic mean enrollment greater
both: uterotonics to prevent and treat atonic PPH; and the NASG than 0.8 cases per clinic per month.
for shock reversal, more rapid recovery from shock, and for non- (PDF)
atonic hemorrhage etiologies.
Checklist S1 CONSORT Checklist. The CONSORT 2010
The difficulty and expense of conducting another NASG RCT checklist of information to include when reporting a cluster
at the PHC level may preclude future efficacy trials. However, we randomized trial.
believe that more research is needed; the next step could be a (PDF)
pragmatic multi-country study, set in high-volume facilities and
referring communities with high maternal mortality, with a NIH Protocol S1 The NASG RCT protocol.
rigorous evaluation framework. (PDF)

Acknowledgments Related Article S1 A related article published in PLOS
ONE.
We thank the many doctors, nurses, midwives, and health (PDF)
system staff in Lusaka District, Kalulushi District, Kitwe District,
Ndola District, and University Teaching Hospital, Kitwe Central Author Contributions
Hospital, and Ndola Central Hospital in Zambia; and the City
Council Clinics and Pariyenatwa Group of Hospitals and Harare Conceived and designed the experiments: SM. Analyzed the data: EFB AE
Central Hospitals in Harare, Zimbabwe, who collected and LG. Wrote the paper: SM EFB LG EAB AE. Supervised the study,
reviewed data and cared for women in this study. Field provided administrative & technical support and revised drafts: TM GM
coordinators include Rhoda Amamfumba, Violet Mambo, Jessica CK. Principal investigator for the study and oversaw implementation of the
DeMulder, Kathleen McDonald, Kelly Winter, Althea Anderson, study and interpreted results: SM. Oversaw database creation manage-
Stephanie Boarden, and Ashley Leech. Many thanks as well to the ment: EFB. Supervised and study and cleaned the data: EAB. Involved in
literature searches, figure development and data cleaning: JG. Helped

PLOS ONE | www.plosone.org 11 October 2013 | Volume 8 | Issue 10 | e76477

NASG: A Cluster Randomized Trial

acquire funding: MM. Created figures & tables, data cleaning and data 24. Sutherland T, Downing J, Miller S, Bishai D, Butrick E, et al. (2013) Use of the
interpretation: AE. Database management: NTMH. non-pneumatic anti-shock garment (NASG) for life-threatening obstetric
hemorrhage: A cost-effectiveness analysis in Egypt and Nigeria. PLoS ONE.
References
25. WHO (2012) WHO recommendations for the prevention and treatment of
1. Lozano R, Wang H, Foreman KJ, Rajaratnam JK, Naghavi M, et al. (2011) postpartum haemorrhage: Evidence base. Geneva: World Health Organization.
Progress towards Millennium Development Goals 4 and 5 on maternal and child 278 p.
mortality: an updated systematic analysis. Lancet 378: 1139–1165.
26. (2007) Zambia Demographic and Health Survey. Calverton, MD: Central
2. WHO (2012) WHO recommendations for the prevention and treatment of Statistical Office, Lusaka, Zambia; Ministry of Health, Lusaka, Zambia;
postpartum haemorrhage. Geneva: World Health Organization. 42 p. Tropical Diseases Research Centre, Ndola, Zambia; University of Zambia,
Lusaka, Zambia.
3. Lalonde A (2012) Prevention and treatment of postpartum hemorrhage in low-
resource settings. Int J Gynaecol Obstet 117: 108–118. 27. (2010–2011) Zimbabwe Demographic and Health Survey. Calverton, MD:
Zimbabwe National Statistics Agency.
4. Mobeen N, Durocher J, Zuberi N, Jahan N, Blum J, et al. (2010) Use of
misoprostol by trained traditional birth attendants to prevent postpartum 28. Ivers NM, Halperin IJ, Barnsley J, Grimshaw JM, Shah BR, et al. (2012)
haemorrhage during home deliveries in Pakistan: a randomised placebo- Allocation techniques for balance at baseline in cluster randomized trials: a
controlled trial. British Journal of Obstetrics and Gynaecology 118: 353–361. methodological review. Trials 13: 120.

5. Derman RJ, Kodkany BS, Goudar SS, Geller SE, Naik VA, et al. (2006) Oral 29. The Clinician’s Ultimate Reference (2010) Mean Arterial Pressure.
misoprostol in preventing postpartum haemorrhage in resource-poor commu- 30. Mantel GD, Buchmann E, Rees H, Pattinson RC (1998) Severe acute maternal
nities: a randomised controlled trial. Lancet 368: 1248–1253.
morbidity: a pilot study of a definition for a near-miss. BJOG, 105 985–990.
6. McCarthy J, Maine D (1992) A framework for analyzing the determinants of 31. Mitra B, Fitzgerald M, Chan J (2013) The utility of a shock index./ = 1 as an
maternal mortality. Stud Fam Plann 23: 23–33.
indication for pre-hospital oxygen carrier administration in major trauma.
7. Ransom EI, Yinger NV (2002) Making Motherhood Safer: Overcoming Injury.
Obstacles on the Pathway to Care.: Population Reference Bureau. 32. Hensleigh P, Ojengbede O, Miller S, Morhason-Bello IO, Fabamwo OA (2005)
Non-pneumatic Anti Shock Garment Resuscitation in Patients with Post Partum
8. Lester F, Stenson A, Meyer C, Morris J, Vargas J, et al. (2011) Impact of the Hemorrhage, Abstract # 029 Presented at the Society of Gynaecology and
Non-pneumatic Anti-shock Garment on pelvic blood flow in healthy postpartum Obstetrics of Nigeria (SOGON) Annual Conference. Tropical Journal of
women. Am J Obstet Gynecol 204: 409e401–405. Obstetrics and Gynaecology 22: S14.
33. Donner A, Klar N (2000) Design and Analysis of Cluster Randomisation Trials:
9. McSwain NE (1988) Pneumatic anti-shock garment: state of the art 1988. Ann In Health Research: Oxford University Press. 178 pages p.
Emerg Med 17: 506–525. 34. Nixon RM, Thompson SG (2003) Baseline adjustments for binary data in
repeated cross-sectional cluster randomized trials. Stat Med 22: 2673–2692.
10. Stenson A, Miller S, Lester F (2012) The Mechanisms of Action of the Non- 35. Miller S, Sloan NL, Winikoff B, Langer A, Fikree FF (2003) Where is the ‘‘E’’ in
pneumatic Anti-Shock Garment (NASG). In: Arulkumaran S, Karoshi M, L.G MCH? The need for an evidence-based approach in safe motherhood.
K, Lalonde A, B-Lynch C, editors. A Comprehensive Textbook of Postpartum J Midwifery Womens Health 48: 10–18.
Hemorrhage: An Essential Clinical Reference for Effective Management. 2nd 36. Campbell OM, Graham WJ (2006) Strategies for reducing maternal mortality:
ed. UK: Sapiens Publications. getting on with what works. Lancet 368: 1284–1299.
37. West KPJ, Katz J, Khatry SK, LeClerq SC, Pradhan EK, et al. (1999) Double
11. McSwain MJ, McSwain NE (2000) Pneumatic antishock garment: state of the blind, cluster randomised trial of low dose supplementation with vitamin A or
art at the turn of the century. Trauma 2: 63–75. beta carotene on mortality related to pregnancy in Nepal. The NNIPS-2 Study
Group. BMJ 318: 570–575.
12. Hauswald M, Greene ER (1986) Aortic blood flow during sequential MAST 38. Altman D, Carroli G, Duley L, Farrell B, Moodley J, et al. (2002) Do women
inflation. Ann Emerg Med 15: 1297–1299. with pre-eclampsia, and their babies, benefit from magnesium sulphate? The
Magpie Trial: a randomised placebo-controlled trial. Lancet 359: 1877–1890.
13. Hauswald M, Greene ER (2003) Regional blood flow after pneumatic anti-shock 39. Manandhar DS, Osrin D, Shrestha BP, Mesko N, Morrison J, et al. (2004) Effect
garment inflation. Prehosp Emerg Care 7: 225–228. of a participatory intervention with women’s groups on birth outcomes in Nepal:
cluster-randomised controlled trial. Lancet 364: 970–979.
14. Hensleigh PA (2002) Anti-shock garment provides resuscitation and haemostasis 40. Jokhio AH, Winter HR, Cheng KK (2005) An intervention involving traditional
for obstetric haemorrhage. BJOG 109: 1377–1384. birth attendants and perinatal and maternal mortality in Pakistan. N Engl J Med
352: 2091–2099.
15. Pelligra R (1994) Non-pneumatic antishock garment use. Emergency: 53–56. 41. Goudar SS, Chakraborty H, Edlavitch SA, Naik VA, Bellad MB, et al. (2008)
16. O’Connor RE, Domeier RM (1997) An evaluation of the pneumatic anti-shock Variation in the postpartum hemorrhage rate in a clinical trial of oral
misoprostol. J Matern Fetal Neonatal Med 21: 559–564.
garment (PASG) in various clinical settings. Prehosp Emerg Care 1: 36–44. 42. Fathalla MF, Mourad-Youssif MM, Meyer C, Camlin C, Turan J, et al. (2011)
17. Christensen KS (1986) Pneumatic antishock garments (PASG): do they Non-Atonic Obstetric Haemorrhage: Effectiveness of the Non-pneumatic Anti-
Shock Garment in Egypt. ISRN Obstet Gynecol 2011.
precipitate lower-extremity compartment syndromes? J Trauma 26: 1102–1105. 43. Morris J, Meyer C, Fathalla MF, Youssif MM, Al-Hussaini TK, et al. (2011)
18. Dickinson K, Roberts I (2000) Medical anti-shock trousers (pneumatic anti- Treating Uterine Atony with the Non-pneumatic Anti-Shock Garment in Egypt.
Afr J Midwifery Womens Health 5: 37–42.
shock garments) for circulatory support in patients with trauma. Cochrane 44. Turan J, Ojengbede O, Fathalla M, Mourad-Youssif M, Morhason-Bello IO, et
Database Syst Rev: CD001856. al. (2011) Positive effects of the non-pneumatic anti-shock garment on delays in
19. Miller S, Fathalla MM, Ojengbede OA, Camlin C, Mourad-Youssif M, et al. accessing care for postpartum and postabortion hemorrhage in Egypt and
(2010) Obstetric hemorrhage and shock management: using the low technology Nigeria. J Womens Health (Larchmt) 20: 91–98.
Non-pneumatic Anti-Shock Garment in Nigerian and Egyptian tertiary care
facilities. BMC Pregnancy Childbirth 10: 64.
20. Miller S, Fathalla MM, Youssif MM, Turan J, Camlin C, et al. (2010) A
comparative study of the non-pneumatic anti-shock garment for the treatment of
obstetric hemorrhage in Egypt. Int J Gynaecol Obstet 109: 20–24.
21. Miller S, Hamza S, Bray E, Gipson R, Nada K, et al. (2006) First aid for
obstetrical haemorrhage: the pilot study of the non-pneumatic anti-shock
garment (NASG) in Egypt. BJOG 113: 424–429.
22. Miller S, Ojengbede O, Turan JM, Morhason-Bello IO, Martin HB, et al. (2009)
A comparative study of the non-pneumatic anti-shock garment for the treatment
of obstetric hemorrhage in Nigeria. Int J Gynaecol Obstet 107: 121–125.
23. Miller S, Turan JM, Dau K, Fathalla M, Mourad M, et al. (2007) Use of the
non-pneumatic anti-shock garment (NASG) to reduce blood loss and time to
recovery from shock for women with obstetric haemorrhage in Egypt. Glob
Public Health 2: 110–124.

PLOS ONE | www.plosone.org 12 October 2013 | Volume 8 | Issue 10 | e76477

Article 4:

Timing of Antiretroviral
Therapy Initiation after a First
AIDS-Defining Event: Temporal
Changes in Clinical
Attitudes in the ICONA Cohort

Timing of Antiretroviral Therapy Initiation after a First
AIDS-Defining Event: Temporal Changes in Clinical
Attitudes in the ICONA Cohort

Antonella Cingolani1*, Alessandro Cozzi-Lepri2, Adriana Ammassari3, Cristina Mussini4,
Maria Alessandra Ursitti5, Pietro Caramello6, Gioacchino Angarano7, Paolo Bonfanti8, Andrea De Luca9,
Maria Stella Mura10, Enrico Girardi11, Andrea Antinori3, Antonela D’Arminio Monforte12,
for Icona Foundation Study group"

1 Department of Public Health, Institute of Infectious Diseases, Catholic University, Roma, Italy, 2 Research Department of Infection & Population Health, University College
London, London, United Kingdom, 3 Clinical Department, National Institute for Infectious Diseases ‘‘L. Spallanzani,’’ Roma, Italy, 4 Institute of Infectious Diseases,
University of Modena and Reggio Emilia, Modena, Italy, 5 Department of Infectious Diseases, S. Maria Nuova IRCCS Hospital, Reggio Emilia, Italy, 6 Infectious and Tropical
Diseases Unit I, Department of Infectious Diseases, Amedeo di Savoia Hospital, Torino, Italy, 7 Department of Infectious Diseases, University of Bari, Bari, Italy, 8 Unit of
Infectious Diseases, A. Manzoni Hospital, Lecco, Italy, 9 Department of Internal and Specialty Medicine, University Infectious Diseases Unit, Azienda Ospedaliera
Universitaria Senese, Siena, Italy, 10 Department of Infectious Diseases, University of Sassari, Sassari, Italy, 11 Department of Epidemiology, National Institute for Infectious
Diseases ‘‘L. Spallanzani,’’ Roma, Italy, 12 Department of Medicine, Surgery and Dentistry University of Milan Clinic of Infectious Diseases, ‘‘San Paolo’’ Hospital, Milan, Italy

Abstract

Background: Time of starting antiretroviral therapy (ART) after diagnosis of specific AIDS-defining event (ADE) is a crucial
aspect. Objectives of this study were to evaluate if in patients diagnosed with ADE the time to ART initiation may vary
according to year of diagnosis and type of ADE.

Methods: All HIV+ persons diagnosed with an ADE over the 6 months prior to or after enrolment in the Icona Foundation
study cohort and while ART-naive were grouped according to type of diagnosis: Those with ADE requiring medications
interacting with ART [group A], those with ADE treatable only with ART [B] and other ADE [C]. Survival analysis by Kaplan-
Meier was used to estimate the percentage of people starting ART, overall and after stratification for calendar period and
ADE group. Multivariable Cox regression model was used to investigate association between calendar year of specific ADE
and time to ART initiation.

Results: 720 persons with first ADE were observed over 1996–2013 (group A, n = 171; B, n = 115; C, n = 434). By 30 days from
diagnosis, 27% (95% CI: 22–32) of those diagnosed in 1996–2000 had started ART vs. 32% (95% CI: 24–40) in 2001–2008 and
43% (95% CI: 33–47) after 2008 (log-rank p = 0.001). The proportion of patients starting ART by 30 days was 13% (95% CI 7–
19), 40% (95% CI: 30–50) and 38% (95% CI 33–43) in ADE groups A, B and C (log-rank p = 0.0001). After adjustment for
potential confounders, people diagnosed after 2008 remained at increased probability of starting ART more promptly than
those diagnosed in 1996–1999 (AHR 1.72 (95% CI 1.16–2.56).

Conclusions: In our ‘‘real-life’’ setting, the time from ADE to ART initiation was significantly shorter in people diagnosed in
more recent years, although perhaps less prompt than expected.

Citation: Cingolani A, Cozzi-Lepri A, Ammassari A, Mussini C, Ursitti MA, et al. (2014) Timing of Antiretroviral Therapy Initiation after a First AIDS-Defining Event:
Temporal Changes in Clinical Attitudes in the ICONA Cohort. PLoS ONE 9(2): e89861. doi:10.1371/journal.pone.0089861

Editor: Yazdan Yazdanpanah, Hopital Bichat Claude Bernard, France

Received September 17, 2013; Accepted January 25, 2014; Published February 27, 2014

Copyright: ß 2014 Cingolani et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: The Icona Foundation Study is supported by unrestricted educational grants of Abbott, Bristol-Myers Squibb, Gilead, JANNSEN, ViiV. The funders had
no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing Interests: The authors received funding from commercial sources, as has been declared by each authors in their ‘‘Declaration of Interests’’ form;
nevertheless the authors confirm that this does not alter their adherence to all the PLOS ONE policies on sharing data and materials.

* E-mail: [email protected]

" Membership of the Icona Foundation Study group is provided in the Acknowledgments.

Introduction of starting ART for people presenting with ADE has been debated
for a long time. The results of the AIDS Clinical Trials Group
A large proportion of HIV-infected people still present for care (ACTG) protocol 5164 showed that starting antiretroviral therapy
with low CD4 cell count or with an AIDS-defining event (ADE) at within the first 30 days after a diagnosis of opportunistic infections
first diagnosis of HIV infection [1]. This represents a population (OI) other than tuberculosis reduces AIDS progression and death
with higher probability of clinical progression and death [2] and by 50% [4] compared to delayed initiation. Moreover, a
lower chances of immunological recovery [3]. The optimal timing subsequent analysis of the same trial demonstrated that early

PLOS ONE | www.plosone.org 1 February 2014 | Volume 9 | Issue 2 | e89861

Timing of ART after First AIDS-Defining Event

Figure 1. Flow diagram showing the original population of people diagnosed with AIDS while ART-naive, patients definitively
considered for the analysis, and reasons for exclusion.
doi:10.1371/journal.pone.0089861.g001

initiation is also a cost-effective approach [5]. Following the prospective follow-up in the cohort and while still ART-na¨ıve were
presentation of ACTG findings, a number of national and included in this analysis (Flow Chart, Figure 1).
international guidelines have been modified accordingly. In
particular, currently the Italian guidelines state that people with Ethics statements
different ADE (except for meningeal tuberculosis and criptococ- All individuals signed an informed consent prior to enrollment
cosis) should start ART within 30 days from ADE diagnosis [6–8].
Nevertheless, the extent to which clinicians in ‘real life’ strictly and the study was approved by the Ethics Committee of each
follow these guidelines remains unexplored. Preliminary results on participating institution that are listed in the Aknowledgments.
people diagnosed with Pneumocystis jiroveci pneumonia (Pcp)
demonstrated that it may be feasible to treat these patients very Statistical analysis
early [9]. Nevertheless, the risk of overlapping toxicities, as well as Patients included were classified in 3 groups according to the
pharmacokinetics/pharmacodinamics interactions between anti-
retrovirals and specific treatment of OI, and the high pill burden type of AIDS diagnosis observed using a ‘in house’ algorithm for
with subsequent risk of poor adherence may all represent factors classification. Specifically, the groups were defined ‘a priori’ on the
limiting the strict implementation of these new recommendations basis of the expected probability of prompt initiation of ART:
[10–13]. In order to evaluate the possible impact of changes in Group A) OIs treated with drugs that may have interaction with
Italian guidelines following the dissemination of the results of trials ART such as for example rifamicins and antineoplastic drugs with
such as ACTG 5164, we analysed temporal changes of the time both NNRTI and PIs: tuberculosis (TB), atypical mycobacteriosis,
from a first diagnosis of ADE to the time of starting antiretroviral non-Hodgkin lymphoma (NHL); Group B) ADE treatable mainly
treatement (ART) in patients of the Icona Foundation Study by ART: progressive multifocal leucoencephalopathy (PML),
cohort who were diagnosed with AIDS when ART-naive. Kaposi’s sarcoma (KS), wasting syndrome, cryptosporidiasis,;
and Group C) all other ADEs (i.e. Pcp, Cytomegalovirus retinitis,
Methods toxoplasmic encephalitis, cervical cancer, criptococcosis, esopha-
geal candidiasis). The expected timing of starting ART was the
Study population slowest for group A diseases, prompt for group B and intermediate
All HIV-1 infected patients of the ICONA Foundation Study for group C.

who were diagnosed with AIDS while ART-na¨ıve regardless of Notwithstanding specific treatment for some conditions includ-
time of enrolment in the cohort and of their CD4 cell count were ed in group C may potentially interact with ART but these
considered for this analysis (eligible patients). ICONA Foundation interactions have not been demonstrated at a pharmacokinetic
Study is an observational cohort of HIV-infected individuals who level.
are antiretroviral na¨ıve at the time of enrolment [14]. This cohort
was set up in January, 1997 and to date consists of more than The inclusion of Kaposi sarcoma in group B might be arbitrary,
10,000 patients from 50 infectious disease units in Italy. Initiation nevertheless, even though international treatment guidelines
and discontinuation dates of each antiretroviral drug, HIV-viral recommend the use of polichemotherapy (PCT) together with
load and CD4 cell count at each clinical visit (every 4–6 months on ART for people with moderate to severe disease ($T1), in clinical
an average) were recorded for each enrolled patient. AIDS- practice, starting ART represents the most rapid curative
defining diseases are recorded in the database at the date that this intervention for any patient presenting with KS.
diagnosis is confirmed or presumptive according to Centers for
Disease Control and Prevention (CDC) criteria [15]. Of the Further, different indications for timing of starting ART for
eligible patients, only individuals who were diagnosed over the 6 NHL and TB (both included in group A) are reported in updated
months prior to enrolment in the cohort or those diagnosed under guidelines and could have influenced the decision of clinicians;
however, considering that the calendar periods of observation are
very different in the studied population, we decided to use the
potential risk of drug-drug interaction as a more homogeneous ‘‘a

PLOS ONE | www.plosone.org 2 February 2014 | Volume 9 | Issue 2 | e89861

Timing of ART after First AIDS-Defining Event

Table 1. Characteristics of patients according to time period of AIDS diagnosis.

Characteristics at AIDS diagnosis 1996–2000 2001–2008 2009+ p-value* Total
N = 332 N = 172 N = 216 N = 720
Age, years
Median (IQR) 37 (32, 43) 40 (34, 47) 39 (34, 47) ,.001 38 (33, 45)

Gender, n(%) 86 (25.9%) 58 (33.7%) 58 (26.9%) 0.161 202 (28.1%)
Female
296 (89.2%) 138 (80.2%) 142 (65.7%) ,.001 576 (80.0%)
Nation of birth, n(%) ,.001
Italian 123 (37.0%) 26 (15.1%) 18 (8.3%) 167 (23.2%)
72 (21.7%) 38 (22.1%) 52 (24.1%) 162 (22.5%)
Mode of HIV Transmission, n(%) 112 (33.7%) 89 (51.7%) 123 (56.9%) 324 (45.0%)
IDU 25 (7.5%) 19 (11.0%) 23 (10.6%) 67 (9.3%)
Homosexual contacts
Heterosexual contacts 1998 (1996, 2000) 2004 (2001, 2008) 2011 (2009, 2013) ,.001 2002 (1996, 2013)
Other/Unknown ,.001
283 (85.2%) 147 (85.5%) 154 (71.3%) 584 (81.1%)
Calendar year of AIDS diagnosis 6 (1.8%) 4 (2.3%) 2 (0.9%) ,.001 12 (1.7%)
Median (range) 43 (13.0%) 21 (12.2%) 60 (27.8%) 124 (17.2%)

HBsAg, n(%) 164 (49.4%) 112 (65.1%) 136 (63.0%) 412 (57.2%)
Negative 136 (41.0%) 37 (21.5%) 21 (9.7%) 194 (26.9%)
Positive 32 (9.6%) 23 (13.4%) 59 (27.3%) 114 (15.8%)
Not tested
60 (24, 150) 57 (24, 221) 51 (17, 165) 0.689 56 (22, 158)
HCVAb, n(%)
Negative 5.31 (4.80, 5.77) 5.14 (4.25, 5.56) 5.30 (4.67, 5.71) 0.065 5.27 (4.66, 5.70)
Positive
Not tested 494 (285, 791) 512 (311, 917) 580 (272, 1052) 0.603 526 (292, 889)

CD4 count, cells/mmc 0.12 (0.06, 0.22) 0.15 (0.07, 0.34) 0.14 (0.05, 0.30) 0.533 0.13 (0.06, 0.27)
Median (IQR) 0.676
86 (25.9%) 35 (20.3%) 50 (23.1%) 171 (23.8%)
Viral load, log10 copies/mL 49 (14.8%) 29 (16.9%) 37 (17.1%) 115 (16.0%)
Median (IQR) 197 (59.3%) 108 (62.8%) 129 (59.7%) 434 (60.3%)

CD8 count, cells/mmc 11 (10, 13) 11 (10, 13) 12 (10, 13) 0.682 11 (10, 13)
Median (IQR)
4000 (2980, 6200) 4700 (3500, 6200) 4570 (3155, 6665) 0.161 4390 (3100, 6230)
CD4/CD8 ratio, cells/mmc
Median (IQR) 31 (22, 51) 36 (21, 52) 34 (21, 58) 0.472 32 (21, 53)

ADE group, n(%) 35 (23, 56) 33 (24, 50) 36 (25, 57) 0.845 35 (24, 56)
Group A
Group B 88.94 (70.60, 92.93) 107.7 (87.27, 123.5) 105.3 (91.11, 128.7) 0.037 104.4 (87.27, 124.5)
Group C ,.001
49 (6.9%)
Haemglobin, g/dL 23 (6.9%) 21 (12.2%) 5 (2.4%) 82 (11.5%)
Median (IQR) 65 (19.6%) 7 (4.1%) 10 (4.9%) 7 (1.0%)
4 (1.2%) 1 (0.6%) 2 (1.0%)
White Blood Cell, cells/mmc
Median (IQR)

ALT, IU/L
Median (IQR)

AST, IU/L
Median (IQR)

eGFR, 100 mls/min/1.73 m2
Median (IQR)

Reason for enrolment, n(%)
No indications for ART
Patients decision
Contraindications

PLOS ONE | www.plosone.org 3 February 2014 | Volume 9 | Issue 2 | e89861

Timing of ART after First AIDS-Defining Event

Table 1. Cont.

Characteristics at AIDS diagnosis 1996–2000 2001–2008 2009+ p-value* Total
N = 216
Recent HIV diagnosis N = 332 N = 172 N = 720
Recent access to care 163 (79.1%)
Physician decision 167 (50.3%) 123 (71.5%) 22 (10.7%) 453 (63.8%)
Unknown 60 (18.1%) 16 (9.3%) 0 (0.0%) 98 (13.8%)
Time from enrolment to ADE, days 13 (3.9%) 3 (1.7%) 4 (1.9%) 16 (2.3%)
Median (range) 0 (0.0%) 1 (0.6%) 5 (0.7%)
Calendar year of HIV diagnosis
Median (IQR) 0 (0, 1243) 0 (0, 3685) 0 (0, 4358) ,.001 0 (0, 4358)

*Chi-square or Wilcoxon test as appropriate. 1997 (1991, 1998) 2003 (2000, 2005) 2011 (2010, 2012) ,.001 2000 (1997, 2009)
doi:10.1371/journal.pone.0089861.t001

priori’’ criterion used by clinicians to decide either to start ART or diagnosis are shown in Table 1. People who were diagnosed with
not. AIDS in 2009 or after were less likely to be Italian born (p,0.001),
intravenous drug users (p,0.001) and co-infected with HCV
In case of multiple diagnoses, clinicians are asked to rank the (p,0.001) than those diagnosed in previous years (Table 1).
illness in order of disease severity, and the most severe condition
was used for classification in the analysis (data not shown). Of the 720 individuals studied, 171 (24%) had a main diagnosis
of group A, 115 (16%) of group B and the remaining 434 (60%) of
ART was defined as a regimen of $1 drug belonging to one of group C. 120 patients (17%) were diagnosed with multiple ADE.
the 3 major historical drug classes (NRTI, NNRTI and PI). Patients in group B were older (p,0.001), more likely to be of
Italian origin (p,0.001), whereas patients in Group A had lower
Calendar periods of AIDS diagnosis were grouped as 1996– CD4 cell count at the time of AIDS diagnosis (both p,0.001) than
2000, 2001–2008, and 2009+, the earliest period reflecting the patients with OIs classified in other groups (Table S1).
years of old generation ART regimens typically with a heavy pill
burden and lower tolerability, the latter period reflecting the years Overall, considering only the major diagnosis for each
after the introduction of modern regimens and following the first individual in those with multiple ADE according to the criterion
presentation of the results of the ACTG 5164 trial at an of severity described in the Methods, the prevalence of specific
International Conference [16]. opportunistic conditions were 32% for Pneumocystis jrovecii
pneumonia (Pcp, n = 231), 16% for tuberculosis (TB, n = 115),
Standard survival analysis by Kaplan-Meier was used to 12% for esophageal candidiasis (n = 84), 5.4% for CMV disease
estimate the cumulative percentage of people starting ART from (n = 39), 7% for Kaposi sarcoma (n = 50), 6.3% for toxoplasma
the date of AIDS diagnosis (overall and after stratification for both encephalitis (n = 45), 7% for wasting syndrome (n = 23); 5.4% for
calendar periods and type of ADEs). non Hodgkin Lymphoma (NHL, n = 39), all other conditions were
observed at a frequency less than 3%. We then assessed trends
All KM plots have been truncated at 30 days. The threshold of over time for specific diagnosis with at least 30 patients being
30 days were based on current treatment guidelines as well as diagnosed. A trend for decreasing prevalence was observed for
assumptions. Thirty days is the maximum length of time whitin esophageal candidiasis (46%, 19%, 34% respectively, in the 3
which is recommended to start ART after a diagnosis of all periods), Pcp (45%, 25%, 29%) and TB (54%, 18%, 26%). In
opportunistic conditions, except for tuberculosis and criptococco- contrast, a stable prevalence was seen for CMV (23%, 36% and
sis. 41%) and KS (32%, 24% and 44%) and for NHL (31%, 31%,
38%).
Multivariable Cox regression model was used to investigate the
association between calendar periods of diagnosis and type of ADE Analysis of time to starting cART according to different
with time to ART initiation after controlling for age, gender, periods and groups of ADEs
nation of birth, HIV transmission route, hepatitis co-infection
status, reason for enrolment in the cohort, CD4 count at diagnosis Overall, 518 people (72%) started ART over follow-up after the
and number of concomitant ADE (single vs. multiple). All AIDS diagnosis; by 30 days from the AIDS diagnosis 33% started
demographics as well as laboratory markers that were associated ART (95% CI: 29–37).
in univariable analysis with a p-value = 0.15 were included in the
multivariable model with the exception of CD8 and white blood Out of 518 persons starting ART, 94 persons (19%) started a
cell counts which were collinear with CD4 counts. regimen containing a 2NRTI+NNRTI, 167 (32%) a 2NRTI+PI
without ritonavir boosting, 170 (33%) a 2NRTI+PI/r and 87
Results (17%) other types of regimen (these include 61 people starting less
than 3 drugs-53 with 2NRTI alone).
Baseline patients characteristics
A total of 720 individuals diagnosed with AIDS over the 6 Figure 2 shows the Kaplan Meier estimates of the cumulative
probability of starting ART, according to the calendar period of
months prior to enrolment in the cohort or under prospective starting ART (Figure 2a) and to group of ADEs (Figure 2b). The
follow-up while still ART-na¨ıve were included in this analysis. In Kaplan Meier estimate of the median time to ART initiation was
included patients, the mean time between HIV and AIDS 58 days [95% CI: 48–70] for patients diagnosed in the period
diagnosis was 32 (SD = 56) months and between enrolment and 1996–1999, 45 days [95%CI: 34–51] for those diagnosed during
AIDS diagnosis was 3 (SD = 15) months. Selected characteristics of
the study population overall and according to periods of AIDS

PLOS ONE | www.plosone.org 4 February 2014 | Volume 9 | Issue 2 | e89861

Timing of ART after First AIDS-Defining Event

PLOS ONE | www.plosone.org 5 February 2014 | Volume 9 | Issue 2 | e89861

Timing of ART after First AIDS-Defining Event

Figure 2. Kaplan-Meier estimates of the cumulative proportion of patients starting ART by 30 days from ADE diagnosis, according
to period of starting ART (a) and group of ADEs (b).
doi:10.1371/journal.pone.0089861.g002

2000–2008, and 38 days [95%CI: 33–48] for those with a promptly than those who were diagnosed in 1996–1999: adjusted
diagnosis of AIDS in/after 2009. The proportions of people HR (AHR) 1.72 (95% CI 1.16–2.56; p = 0.007). The strongest
starting ART by 30 days in these same groups were 27% (95% calendar year effect seemed to have occurred for diagnoses
CI:22–32) of those diagnosed in 1996–2000, 32% (95% CI:24–40) belonging to group A (AHR 3.61; 95%CI 1.48–8.77; p = 0.005)
of those in 2001–2008 and 43% (95% CI:33–47) of those in/after (Table 2a).
2009 (log-rank p = 0.001, Figure 2a).
Moreover, after controlling for potential confounders, time to
Patients in group A started ART after a median time of 116 ART initiation was shorter in people belonging to group B (ARH
days [95%CI 82–148], those in group B after a median of 35 days 1.63, 95% CI:1.05–2.53; p = 0.02) compared to patients in group
[95%CI 30–49] and those in group C after 38 days [95%CI 35– A. The difference in time to start of ART between groups B and A
41, log-rank p = 0.0002, Figure 2b). The proportion of patients was consistent across calendar periods. (p-value for interaction
starting ART by 30 days was 13% (95% CI 7–19), 40% (95% CI: p = 0.19, Table 3).
30–50) and 38% (95% CI 33–43) in ADE groups A, B and C,
respectively. After concomitantly stratifying for both these Analysis of other predictors of starting ART
variables, people belonging to group B diagnosed in or after We identified other factors that were associated with the time of
2009 had the highest overall probability of starting ART by 30
days (51%; 95%CI: 33–69) (Figure S1, data not shown). Of note, ART initiation independent of calendar year and type of AIDS
the Kaplan-Meier estimate of the median time to starting ART in diagnosis from fitting a Cox regression analysis (Table 4). People
the whole Icona cohort (11,303 individuals enrolled) is much born in Italy (vs. non Italian-born ARH = 1.83, 95% CI:1.21–
longer, at 157 months (95% CI:139–173). 2.78) those with lower CD4 cell count at the time of their ADE
diagnosis (per 100 cells/mm3 lower AHR = 1.22 [95%CI 1.11–
After adjusting for a number of potential confounding factors 1.34] and people who were enrolled in the cohort because recently
listed in the Methods, people diagnosed in 2009 or after, remained discovered to be HIV-infected (vs. those with no indications for
at significantly increased probability of starting ART more starting ART at enrolment, ARH = 1.86, 95% CI:1.17–2.93) were

Table 2. Relative Hazards (RH) of starting ART from fitting a Cox regression analysis, according to calendar periods and stratified
by type of ADE.

Crude and adjusted relative hazards of ART initiation

Crude RH (95% CI) p-value Adjusted* RH (95% CI) p-value

All patients 1.00 0.104 1.00 0.610
Calendar Period 1.19 (0.96, 1.48) ,.001 0.92 (0.67, 1.26) 0.007
1996–1999 1.51 (1.22, 1.86) 1.72 (1.16, 2.56)
2000–2008 0.578
2009+ 1.00 0.726 1.00 0.005
Group Aa 1.09 (0.66, 1.81) ,.001 1.24 (0.59, 2.61)
Calendar Period 2.38 (1.49, 3.81) 3.61 (1.48, 8.77) 0.090
1996–1999 0.736
2000–2008 1.00 0.851 1.00
2009+ 0.95 (0.55, 1.65) 0.058 0.39 (0.13, 1.16) 0.619
Group Bb 1.63 (0.98, 2.69) 1.23 (0.37, 4.02) 0.342
Calendar Period
1996–1999 1.00 0.156 1.00
2000–2008 1.21 (0.93, 1.57) 0.095 0.89 (0.58, 1.39)
2009+ 1.26 (0.96, 1.66) 1.32 (0.74, 2.37)
Group Cc
Calendar Period
1996–1999
2000–2008
2009+

*adjusted for age, gender, nation of birth, mode of HIV transmission,hepatitis co-infection status, type of AIDS diagnosis-all patients model only-,reason for enrolment in

Icona, CD4 count and viral load at diagnosis, number of concomitant AIDS diagnoses
amycobacteriosis, tubercolosis, Non-Hodgkin lymphomas
bisosporidiosis, criptosporidiosis, PML, Kaposi sarcoma andAIDS dementia complex
call other ADE

doi:10.1371/journal.pone.0089861.t002

PLOS ONE | www.plosone.org 6 February 2014 | Volume 9 | Issue 2 | e89861

Timing of ART after First AIDS-Defining Event

Table 3. Relative hazards (RH) of starting ART according with excluded (due to missing values of these variables) were statistically
groups of ADEs from fitting a Cox regression analysis and different for some demographic characteristics but large differ-
stratified by calendar period. ences were not observed (data not shown) with the exception of
year of diagnosis (more recent year of AIDS diagnosis in excluded
Crude and adjusted relative hazards of ART patients [2007 on average], compared to patients included in
initiation multivariable model [2000]). People who were diagnosed with
HIV in more recent years were more likely to have missing data
Crude RH p- Adjusted* RH p-value for CD4 count or viral load at time of their AIDS diagnosis. This
(95% CI) value (95% CI) could be due to the fact that data collection was modified
approximately in 2005 to collect only one historical value of these
All patients 1.00 ,.001 1.00 0.028 markers at entry in the cohort (two values were instead recorded in
Type of ADE 2.18 (1.63, 2.90) ,.001 1.63 (1.05, 2.53) 0.962 previous years).
Group Aa 2.04 (1.64, 2.55) 1.01 (0.71, 1.42)
Group Bb Discussion
Group Cc 1.00 ,.001 1.00 0.037
1996–1999 2.30 (1.52, 3.48) ,.001 1.91 (1.04, 3.52) 0.392 The results of our analysis show that in recent years Italian
Type of ADE 2.18 (1.60, 2.96) 1.22 (0.78, 1.91) clinicians tend to initiate ART more promptly than in the past
Group Aa after a diagnosis of AIDS. Nevertheless, even in recent years the
Group Bb 1.00 0.029 1.00 0.408 overall probability of starting ART seems to remain low (43% by
Group Cc 2.03 (1.08, 3.84) ,.001 1.55 (0.55, 4.40) 0.998 30 days from the AIDS diagnosis) with a median time to ART
2000–2008 2.42 (1.48, 3.97) 1.00 (0.40, 2.51) initiation of 21 days. Indeed, these estimates are considerably
Type of ADE lower than what has been suggested as optimal time of starting
Group Aa 1.00 0.010 1.00 ART on the basis of the results of the ACTG 5164 trial [4]. Even
Group Bb 1.99 (1.18, 3.36) 0.047 5.37 (1.08, 26.81) 0.040 in the subset of people who were diagnosed under prospective
Group Cc 1.55 (1.01, 2.38) 1.95 (0.52, 7.39) 0.325 follow-up in the cohort, the estimate remains low (38% instead of
2009+ 33% by 30 days). Of note, however, our estimates refer to any
Type of ADE diagnosis of OI while in the trial 75% of patients had been
Group Aa diagnosed with Pcp and cases of TB were excluded.
Group Bb
Group Cc To our knowledge ours is the first analysis evaluating whether
there has been a change in the timing of ART initiation in clinical
*adjusted for age, gender, nation of birth, mode of HIV transmission,hepatitis practice over time in cART-naive patients presenting for care with
different AIDS-defining conditions.
co-infection status, calendar period-all patients model only-,reason for
Recently, Geng et al reported a 3-fold increase in the rate of
enrolment in Icona, CD4 count and viral load at diagnosis, number of ART initiation and an increase in the probability of early ART
from 7% to 50% after the diffusion of ACTG 5164 findings in a
concomitantAIDS diagnoses single clinical center in the USA uniquely in patients diagnosed
amycobacteriosis, tubercolosis, Non-Hodgkin lymphomas with Pcp [9].
bisosporidiosis, criptosporidiosis, PML, Kaposi sarcoma andAIDS dementia
In a retrospective European and Canadian multi-cohort study,
complex aimed to analyse the clinical progression of AIDS-presenter
call other ADE patients, the proportion of patients starting ART by 30 days from
diagnosis was 50% when considered the overall study population,
doi:10.1371/journal.pone.0089861.t003 ranging from 70% in people diagnosed with KS to 30% in those
with NHL, with wide variation in the time to initiation also
significantly more likely to start ART. Interestingly, viral load at according to country of origin [2]. Of note, not all individuals
the time of AIDS diagnosis was not independently associated with studied by us were AIDS presenters, 462/720 (64%) developed
the probability of starting ART. AIDS after enrolment in the cohort.

Description of people not starting ART. Two hundred There are a number of potential concerns and obstacles to the
and two patients were still off-ART. Of these 8 had died and the early initiation of ART in patients presenting with opportunistic
remaining 194 had not commenced ART at the time that they conditions: the risk of immune reconstitution inflammatory
were last seen alive (Figure 1). The reasons for not starting ART or syndrome, the concern for poor adherence to multi-drug
being censored are not collected in our database. However, for the treatment, for drug-drug interactions potentially leading to adverse
633 patients who were diagnosed before or at the time of events or poor treatment responses, and the difficulties in taking or
enrolment we have the reason for being ART-na¨ıve. The main tolerating oral medications for critically ill patients. All these
reasons were: recent seroconversion (68%), recent access to care represent potential reasons for delaying antiretroviral treatment
(14%) and patients’ reasons (10%). Overall, the distribution of and may partially explain our findings [17]. Indeed, people with
reasons for being ART-naive at enrolment was similar in people AIDS defining conditions for which ART represents the only
who subsequently started ART or those who remained off-ART treatment option were treated earlier, whereas those whose
(data not shown). conditions were treatable with medications having potential
drug-drug interaction with antiretrovirals were treated later.
Assessment of missing data. Some of the variables in the
model were fitted as categorical and individuals with missing data Although both current Italian and International guidelines
for these factors have been grouped separately as ‘value unknown’. report a general agreement on the net benefit of a prompt start of
People with missing data for variables included as continuous (e.g. ART in people with AIDS, data evaluating the effect of early ART
CD4 count at the date of AIDS diagnosis) were, however, on specific opportunistic conditions other than Pcp, TB and
excluded from the multivariable calculations. People included and cryptococcosis are very limited, and this could lead the clinicians

PLOS ONE | www.plosone.org 7 February 2014 | Volume 9 | Issue 2 | e89861

Timing of ART after First AIDS-Defining Event

Table 4. Other factors associated with the chance of starting ART - Relative hazards (RH) from fitting a Cox regression.

Crude and adjusted relative hazards of ART initiation

Crude RH (95% CI) p-value Adjusted* RH (95% CI) p-value
0.652
Age 1.20 (1.09, 1.31) ,.001 0.97 (0.84, 1.12) 0.422
per 10 year older 1.00 0.004
Gender, n(%) 1.00 0.122 1.14 (0.83, 1.55) 0.664
Male 0.86 (0.71, 1.04) 1.00 0.797
Female ,.001 1.83 (1.21, 2.78) 0.956
Nation of birth, n(%) 1.00 1.00 0.610
Non Italian 1.50 (1.19, 1.90) 0.018 1.11 (0.69, 1.79) 0.007
Italian ,.001 1.07 (0.65, 1.76) 0.948
Mode of HIV Transmission, n(%) 1.00 0.029 1.02 (0.55, 1.89) 0.001
IDU 1.36 (1.05, 1.75) 1.00 ,.001
Homosexual contacts 1.49 (1.19, 1.87) 0.104 0.92 (0.67, 1.26) 0.501
Heterosexual contacts 1.47 (1.04, 2.07) ,.001 1.72 (1.16, 2.56)
Other/Unknown 1.00 0.028
Calendar Period 1.00 0.008 1.01 (0.67, 1.53) 0.962
1996–1999 1.19 (0.96, 1.48) 0.013 0.44 (0.27, 0.72)
2000–2008 1.51 (1.22, 1.86) ,.001 1.22 (1.11, 1.34) 0.097
2009+ ,.001 1.06 (0.90, 1.25) 0.107
Hepatitis B/C coinfection, n(%) 1.00 ,.001 0.008
Negative 0.77 (0.63, 0.93) 1.00 0.361
Positive 0.73 (0.57, 0.94) ,.001 1.63 (1.05, 2.53)
Not tested ,.001 1.01 (0.71, 1.42)
CD4 count 1.22 (1.15, 1.29) 0.089
per 100 cells/mmc lower ,.001 1.00
Viral load 1.37 (1.20, 1.57) 0.524 0.63 (0.37, 1.09)
per log10 copies/mL higher 0.627 0.19 (0.03, 1.43)
CD8 count 0.96 (0.94, 0.98) 0.147 1.86 (1.17, 2.93)
per 100 cells/mmc higher 1.26 (0.77, 2.08)
ADE group, n(%) 1.00 0.564
Group A 2.18 (1.63, 2.90) 0.396
Group B 2.04 (1.64, 2.55) ,.001
Group C 0.045
Haemglobin 0.63 (0.37, 1.07)
per 100 g/dL higher
White Blood Cell 0.91 (0.86, 0.95)
per 1000 cells/mmc higher
ALT 1.07 (0.86, 1.34)
per 100 IU/L higher
AST 1.06 (0.84, 1.34)
per 100 IU/L higher
eGFR 1.29 (0.91, 1.82)
per 50 100 mls/min/1.73 m2 higher
Reason for enrolment, n(%) 1.00
No indications for ART 1.12 (0.77, 1.62)
Patients decision 0.64 (0.23, 1.78)
Contraindications 2.04 (1.51, 2.76)
Recent HIV diagnosis 1.45 (1.01, 2.08)
Recent access to care

PLOS ONE | www.plosone.org 8 February 2014 | Volume 9 | Issue 2 | e89861

Timing of ART after First AIDS-Defining Event

Table 4. Cont.

Crude and adjusted relative hazards of ART initiation

Crude RH (95% CI) p-value Adjusted* RH (95% CI) p-value
0.723
No. of diagnoses, n(%) 1.21 (1.02, 1.44) 0.033 1.05 (0.79, 1.40)
Per 1 additional

doi:10.1371/journal.pone.0089861.t004

to hesitate in ART initiation in people diagnosed with less well- criterion according to CDC guidelines. As a consequence people
studied conditions such as CMV disease or esophageal candidiasis may be started on ART after a presumptive and before a
or toxoplasmosis. confirmative diagnosis and excluded from the analysis as a result.
This may have led to selection bias as well as over-estimation of
In our study population, TB accounted for 14% of people with the estimated time to start ART. However, only 88 of the 720
AIDS. When we looked at people diagnosed with TB alone the diagnoses (12%) analyzed here were made using a presumptive
median time to starting ART was 142 days (95% CI: 97–190). Of diagnostic criterion.
note, we found that the effect of calendar period was stronger in
group A (including TB) than in group C (including Pcp); this was Furthermore, we did not perform a survival analysis with
confirmed even after restricting the analysis to people with TB or endpoint death so we do not know what the impact of such a
Pcp only (data not shown). This may represent an unexpected conservative approach may be on mortality in a real practice
result, because recommendation on earlier start of ART based on setting. We are indeed planning to investigate the association
the data of ACTG 5164 concerned more specifically Pcp rather between the strategies of starting ART within 15 days vs. starting
than TB. It can be speculated that other factors may have affected within 30 days vs delaying for longer than 30 days or not start at
a prompt initiation of ART in people with TB such as a more all and the risk of death using g-computation in a separate report
extensive use of NNRTI in the recent years (with less concerns in [23]. Patients with missing data have been excluded from the
terms of drug-drug interactions), or a progressive increase of multivariable Cox regression analysis. However, included and
awareness of findings of randomized clinical trials on TB [18–21]. excluded patients were similar for the other measured character-
In addition, immune reconstitution inflammatory syndrome (IRIS) istics and therefore selection bias is unlikely. There was a large
is often observed after starting treatment for specific disorders (i.e., difference in calendar year of diagnosis likely to be due to a change
TB, CMV, criptococcosis) and this might be another reason why in modality of data collection over time which is unrelated to time
clinicians are less likely to promptly start ART in people with TB. to ART initiation.

The opposite trend was seen for people diagnosed with In conclusion, in a real life setting of patients seen for care in
cryptococcal meningitis (less chance to initiate ART in recent Italy, we show that the time from AIDS diagnosis to ART
periods) this might be due to increasing evidence for a worst initiation was significantly shorter in more recent years, even
survival in patients with this pathology starting ART earlier [21– though still considerably longer than what has been reported as
22]. beneficial on survival in AIDS patients. Reasons for that could be
a delay in the implementation of guidelines in clinical practice, but
Investigating the predictors of time to ART initiation, we found could also be due to real difficulties in treating these patients
that patients whose reason for enrolment in Icona was a recent concomitantly for opportunistic conditions and for HIV itself.
diagnosis of HIV infection, were more likely to be promptly started Similar prospective studies with data collection enrichment are
on ART compared to patients who had no indication for initiating warranted to explore in details the specific barriers to ART
ART. These are likely to be people who were diagnosed with HIV initiation in patients presenting with ADE.
when they presented with AIDS and it is conceivable that their
time to ART initiation is the shortest. Of note indication for Supporting Information
starting ART have also been modified in guidelines over time but
to a lower extent for people with already a diagnosis of AIDS. Figure S1 Kaplan-Meier estimates of the cumulative
proportion of patients starting ART by 30 days from
Moreover, patients of non-Italian nationality tended to start ADE diagnosis, according to different combination of
ART later than Italian patients. Some concerns regarding the period of starting ART and group of ADEs.
retention in care of migrant patients might explain the conserva- (TIFF)
tive approach of clinicians in starting ART in this subset of
patients. Table S1 Characteristics of patients according to
different groups of ADEs.
Our analysis has several limitations. (DOCX)
First, although a severity criterion could be establish for the
classification of ADE, for a given pathology, we do not collect data Author Contributions
on the degree of severity or progression of the disease. For
example, it could be hypothesized that in case of greater severity, Analyzed the data: ACL. Wrote the paper: AC ACL A. Ammassari CM
requiring also more intensive care, ART was not started ADL EG A. Antinori ADM PB GA MSM PC MAU.
immediately because patients were not able to take it. Moreover,
AIDS diagnoses are typically made using a presumptive diagnostic

References 2. Miro JM, Manzardo C, Mussini C, Johnson M, d’Arminio Monforte A, et al;
Late Presenters Investigators. (2011) Survival outcomes and effect of early vs.
1. Antinori A, Coenen T, Costagiola D, Dedes N, Ellefson M, et al for European
Late Presenter Consensus Working Group (2011). Late presentation of HIV
infection: a consensus definition. HIV Med. 12:61–4.

PLOS ONE | www.plosone.org 9 February 2014 | Volume 9 | Issue 2 | e89861

Timing of ART after First AIDS-Defining Event

deferred cART among HIV-infected patients diagnosed at the time of an AIDS- 13. Cingolani A, Torti L, Pinnetti C, de Gaetano Donati K, Murri R, et al. (2010)
defining event: a cohort analysis. PLoS One.; 6: e26009. Detrimental clinical interaction between ritonavir-boosted protease inhibitors
3. Mocroft A, Phillips AN, Gatell J, Ledergerber B, Fisher M, et al. (2007) and vinblastine in HIV-infected patients with Hodgkin’s lymphoma. AIDS.
Normalization of CD4 counts in patients with HIV-1 infection and maximum 24(15):2408–12.
virological suppression who are taking combination antiretroviral therapy: an
observational cohort study. Lancet 370:407–413. 14. d’Arminio Monforte A, Cozzi-Lepri A, Girardi E, Castagna A, Mussini C, et al
4. Zolopa AR, Andersen J, Komarow L, Sanchez A, Sanne I, et al. (2009) Early for Icona Foundation Study Group. (2011) Late presenters in new HIV
antiretroviral therapy reduces AIDS progression/death in individuals with acute diagnoses from an Italian cohort of HIV-infected patients: prevalence and
opportunistic infections: a multicenter randomized strategy trial. PlosOne clinical outcome. Antivir Ther. 16(7):1103–12.
4:5575–84.
5. Sax PE Sloan CE, Schackman BR, Grant PM, Rong J, et al. (2010) Early 15. Revised Classification system for HIV infection and expanded surveillance case
antiretroviral therapy for patients with acute AIDS related opportunistic definition for AIDS among adolescents and adults. MMWR Recomm Rep 1992;
infections: a cost-effectiveness analysis of ACTG 5164. HIV Clin Trials 11; 41 (RR-17):1–19.
248–2.
6. Guidelines for Prevention and Treatment of Opportunistic Infections in HIV- 16. Zolopa A, Andersen J, Komarow L, et al. Immediate vs deferred ART in the
Infected Adults and Adolescents Centers for Disease Control and Prevention the setting of acute AIDS-related OI: final results of a randomized strategy trial,
National Institutes of Health, and the HIV Medicine Association of the ACTG A5164. 15th CROI, 2008; Boston, MA, Abstract 142.
Infectious Diseases Society of America 2013.Downloaded from http://aidsinfo.
nih.gov/guidelines on 7/24/2013. 17. Muller M, Wandel S, Colebunders R, Attia S, Furrer H, et al. (2010) Immune
7. EACS Clinical management and treatment of HIV infected adults in Europe. reconstitution inflammatory syndrome in patients starting antiretroviral therapy
Version 5, November 2009. for HIV infection: a systematic review and meta-analysis. Lancet Infect Dis,
8. Antinori A, Marcotullio S, Ammassari A, Andreoni M, Angarano G, et al for 10:251–261.
Italian HIV Guidelines Working Group. (2012) Italian guidelines for the use of
antiretroviral agents and the diagnostic-clinical management of HIV-1 infected 18. Havlir DV, Kendall MA, Ive P, Kumwenda J, Swindells S, et al. (2011) Timing
persons. Update 2011. New Microbiol. 35(2):113–59. of antiretroviral therapy for HIV-1 infection and tuberculosis. N Engl J Med.
9. Geng EH, Kahn JS, Chang OC, Hare CB, Christopoulos KA, et al. (2011) The 365(16):1482–91.
effect of ACTG Clinical Trials Group Protocol 5164 on the time from
Pneumocystis jirovecii pneumonia diagnosis to antiretroviral initiation in routine 19. Blanc FX, Sok T, Laureillard D Borand L, Rekacewicz C, et al. (2011) Earlier
clinical practice: a case study of diffusion, dissemination, and implementation. versus later start of antiretroviral therapy in HIV-infected adults with
Clin Infect Dis 53:1008–101. tuberculosis. N Engl J Med. 365(16):1471–81.
10. Boulanger C, Hollender E,Farrell K, Stambaugh JJ, Maasen D, et al. (2009)
Pharmacokinetic evaluation of rifabutin in combination with lopinavir-ritonavir 20. Abdool Karim SS, Naidoo K, Grobler A, Padayatchi N, Baxter C, et al. (2011)
in patients with HIV infection and active tuberculosis. Clin Infect Dis 49: 1305– Integration of antiretroviral therapy with tuberculosis treatment. New Engl J Med
1311. 365:1492–1501.
11. Khachi H, O’Connell R, Ladenheim D, Orkin C (2009) Pharmacokinetic
interaction between lopinavir/ritonavir an rifabutin in HIV-infected patients 21. Bisson GP, Molefi M, Bellamy S, Thakur R, Steenhoff A, et al. (2013) Early
with mycobacterial co-infection. J Antimicrob Chemother 64: 871–873. versus delayed antiretroviral therapy and cerebrospinal fluid fungal clearance in
12. Centers for Disease Control and Prevention. Managing drug interactions in the adults with HIV and cryptococcal meningitis. Clin Infect Dis 56; 1165–1173.
treatment of HIV-related tuberculosis http://www.cdc.gov/tb/publications/
guidelines/TB_HIV_Drugs/default.htm. 22. Boulware D, et al. ART initiation within the first 2 weeks of cryptococcal
meningitis is associated with higher mortality: a multisite randomized trial. 20th
Conference on Retroviruses and Opportunistic Infections, 3–6 March 2013,
Atlanta, Abstract 144.

23. Daniel RM, Cousens SN, De Stavola BL, Kenward MG, Sterne JA (2013)
Methods for dealing with time-dependent confounding. Stat Med. 32(9):1584–
618.

PLOS ONE | www.plosone.org 10 February 2014 | Volume 9 | Issue 2 | e89861