Sentinel surveillance for rotavirus in children

Abstract

Background:
Rotavirus is the leading cause of severe acute gastroenteritis (AGE) in children <5 years of age in Myanmar. The purpose of this analysis is to report from the sentinel surveillance system for rotavirus gastroenteritis (RVGE), which collects information on the epidemiology and circulating genotypes to assess the disease burden and support vaccine introduction in Myanmar.

Methods:
Prospective, active surveillance for RVGE-associated hospitalizations was conducted during 2009 -2014 at Yangon Children's Hospital. Stool samples collected from children <5 years of age admitted for AGE were screened for rotavirus antigen by ELISA (ProSpecT™ Rotavirus, OXOID-UK). G and P genotyping was performed by reverse transcription polymerase chain reaction.

Results:
Overall, 1860/3724 (49.9%) of stool samples tested positive for rotavirus, ranging from 42-56% of hospitalized AGE cases each year. RVGE was predominant in the 6-11 months age group 889/1860 (47.8%) as compared with 12-23 months 633/1860 (34.0%), 0-5 months 226/1860 (12.2 %) and 24-59 months 112/1860 (6.0%). RVGE occurred in a seasonal cycle with peak occurrence in the cold and dry months (November to February), accounting for 65.3% (1151/1763) among enrolled AGE cases. Vomiting (84.1% Vs 67.9%; P < .01), fever (84.5% Vs 75.6%; P < .01) and dehydration (78% Vs 69%; P < .01) were more frequently observed in RVGE than non-RVGE. Genotyping revealed that G1P[8] was predominant from January to June 2009, G12P[8] was predominant throughout 2009-2012 which was replaced in 2012-2013 by G2P[4] and changed again to G1P[8] in 2013-2014 and G9P[8] in late 2014.

Conclusions:
Rotavirus is accounting for approximately half of AGE-associated hospitalizations among children <5 years of age in Myanmar. There is immense diversity of rotavirus strains similar to that reported previously for other countries in the region. Information gained from this surveillance system highlights consideration of rotavirus vaccine introduction into this target population.

Full text

UNCORRECTED PROOF
Vaccine xxx (2017) xxx-xxx
Contents lists available at ScienceDirect
Vaccine
journal homepage: www.elsevier.com
Sentinel surveillance for rotavirus in children <5 years of age admitted for Diarrheal
illness to Yangon Children’s Hospital, Myanmar, 2009–2014
Theingi Win Myata, ⁎, Hlaing Myat Thua, Ye Myint Kyawb, Khin Mar Ayea, Mo Mo Wina, Htin Lina,
Thin Thin Shwea, Win Mara, Khin Khin Ooa, Kyaw Zin Thanta
aDepartment of Medical Research, Ministry of Health and Sports, Myanmar
bYangon Children’s Hospital, Ministry of Health and Sports, Myanmar
ARTICLE INFO
Article history:
Available online xxx
Keywords:
Rotavirus
Acute gastroenteritis
Diarrhea
Hospitalization
Myanmar
ABSTRACT
Background
Rotavirus is the leading cause of severe acute gastroenteritis (AGE) in children <5 years of age in Myan-
mar. The purpose of this analysis is to report from the sentinel surveillance system for rotavirus gastroenteritis
(RVGE), which collects information on the epidemiology and circulating genotypes to assess the disease bur-
den and support vaccine introduction in Myanmar.
Methods
Prospective, active surveillance for RVGE-associated hospitalizations was conducted during 2009 –2014
at Yangon Children’s Hospital. Stool samples collected from children <5 years of age admitted for AGE were
screened for rotavirus antigen by ELISA (ProSpecT™ Rotavirus, OXOID-UK). G and P genotyping was per-
formed by reverse transcription polymerase chain reaction.
Results
Overall, 1860/3724 (49.9%) of stool samples tested positive for rotavirus, ranging from 42-56% of hos-
pitalized AGE cases each year. RVGE was predominant in the 6–11 months age group 889/1860 (47.8%) as
compared with 12–23 months 633/1860 (34.0%), 0–5 months 226/1860 (12.2 %) and 24–59months 112/1860
(6.0%). RVGE occurred in a seasonal cycle with peak occurrence in the cold and dry months (November
to February), accounting for 65.3% (1151/1763) among enrolled AGE cases. Vomiting (84.1% Vs 67.9%;
P <.01), fever (84.5% Vs 75.6%; P< .01) and dehydration (78% Vs 69%; P < .01) were more frequently ob-
served in RVGE than non-RVGE.
Genotyping revealed that G1P[8] was predominant from January to June 2009, G12P[8] was predomi-
nant throughout 2009–2012 which was replaced in 2012–2013 by G2P[4] and changed again to G1P[8] in
2013–2014 and G9P[8] in late 2014.
Conclusions
Rotavirus is accounting for approximately half of AGE-associated hospitalizations among children
<5 years of age in Myanmar. There is immense diversity of rotavirus strains similar to that reported previously
for other countries in the region. Information gained from this surveillance system highlights consideration of
rotavirus vaccine introduction into this target population.
© 2017.
1. Introduction
Acute gastroenteritis (AGE) is among the top causes of childhood
mortality worldwide, and rotavirus is the leading cause of severe AGE
in children <5 years of age [1]. The annual global burden of rotavirus
deaths during 2013 in this age group is estimated at 215,000 (range,
197,000–233,000) and approximately half (49%) of these deaths oc-
curred in India, Nigeria, Pakistan and Democratic Republic of Congo
[2].
Rotavirus gastroenteritis (RVGE) can be prevented through vacci-
nation, and WHO recommends that rotavirus vaccines (RV) be incor
⁎Corresponding author.
Email address: drtheingiwinmyat@gmail.com, royalwin@gmail.com (T.W.
Myat)
porated into the national immunization programs of all countries, par-
ticularly in those with high child mortality due to diarrhea, such as in
Southeast Asia and sub-Saharan Africa [3]. Currently, there are two
globally available vaccines with demonstrated efficacy against severe
rotavirus disease; Rotarix (GSK Bio), a monovalent vaccine contain-
ing a single G1P[8] strain) and RotaTeq (Merck), a pentavalent vac-
cine containing 4 common G types (G1–G4) and 1 common P type
P[8]) [4].
In Myanmar, diarrhea is among the priority childhood diseases,
with establishment of improved surveillance for rotavirus indicating
that the proportion of RVGE among hospitalized children with diar-
rhea in Yangon, reported as 22% in 1980, prior to surveillance es-
tablishment, is actually closer to 50% in the 2000 s [5,6]. Despite the
high prevalence of RVGE, rotavirus vaccine has not yet been intro-
duced. Data on the epidemiology of RVGE, disease burden and circu
https://doi.org/10.1016/j.vaccine.2017.11.002
0264-410/© 2017.

UNCORRECTED PROOF
2 Vaccine xxx (2017) xxx-xxx
lating genotypes is crucial to provide data supporting the introduction
of RV into Myanmar’s national immunization program. These data
may also be used for post-vaccination monitoring of vaccine impact
and effectiveness. The objective of this analysis is to report data from
rotavirus sentinel surveillance conducted in Yangon Children’s Hos-
pital (YCH) during 2009–2014.
2. Materials and methods
2.1. Study Design and setting
Prospective, active, hospital based sentinel surveillance for AGE
and rotavirus-associated admission among children < 5 years of age
was conducted at Yangon Children’s Hospital (YCH) from January
2009 through December 2014 according to the World Health Organi-
zation (WHO) generic protocol [7]. YCH is the largest and main ter-
tiary care referral hospital in Yangon with a pediatric inpatient unit
consisting of 3 wards and 1300 beds. Approximately 10,000–13,000
children under 5 years of age are admitted year round to these wards,
with admissions to each ward occurring on a rotating basis.
2.2. Selection criteria for participants
Cases eligible for enrollment were children < 5 years of age pre-
senting with AGE (≥3 looser-than-normal stools in a 24-h period dur-
ing the illness, with onset of diarrhea ≤ 14 days at presentation) and
treated at one of the three pediatric-medical wards of YCH. We ex-
cluded those with diarrhea duration longer than 14 days (chronic diar-
rhea) on admission and presence of either blood or mucus in the stool.
2.3. Surveillance methodology
Active case finding was conducted 5 days per week (Monday-Fri-
day) on the post-admission days of each of the pediatric-medical
wards. As such, enrolment did not occur for children who were ad-
mitted Friday or Saturday, and children admitted Sunday were ap-
proached for enrolment on Monday. All children <5 years of age hos-
pitalized due to AGE, who met the selection criteria and whose par-
ents or legal guardian consented were enrolled in the study. On enroll-
ment, a case report form containing information on demography, clin-
ical history, physical examination, treatment and outcome was com-
pleted. A stool sample containing not less than 3 ml was collected
within 48 h of admission, using wide-mouth screw capped bottles.
The bottles were labeled and transported on the same day in cold
chain to the Virology Research Division laboratory at the Department
of Medical Research, and subsequently stored at −20 °C until test-
ing was performed. Upon discharge, the date and outcome were also
recorded on the case report form.
2.4. Laboratory analysis
All samples were tested for rotavirus antigen by ProSpecT™ Ro-
tavirus ELISA kit, Oxoid, UK according to the manufacturer's in-
structions. A subset of rotavirus-positive stool samples (approximately
30% of ELISA positive samples from each month) was randomly cho-
sen for G (VP7) and P (VP4) genotyping. RNA was extracted by us-
ing QIAamp Viral RNA Mini Kit (QIAGEN GmbH, Germany) and
the extracted dsRNA was amplified by RT-PCR using specific oligin-
ucleotide primers provided by the WHO Rotavirus Reference Labora-
tory, CMC, Vellore, India [8].
2.5. Statistical analysis
We report trends in rotavirus testing during the surveillance pe-
riod. Additionally, we categorized the patients into rotavirus-positive
and rotavirus-negative and compared these groups with respect to de-
mographics, clinical, treatment and outcome parameters. Chi-square
test was used to determine statistically significant differences between
these groups and a p-value <0.05 was considered significant.
2.6. Ethical consideration
This study was approved by the Ethics Review Committee, Depart-
ment of Medical Research, Myanmar. Informed consent was obtained
from the parents or guardians of children prior to enrolment.
3. Results
During 2009 through 2014, 4359 eligible children were identified
through the surveillance system, and of these, 3724 (85.4%) were en-
rolled. Six hundred and thirty-five children were not enrolled. Rea-
sons for not enrolling children include: lack of caregiver availability,
child being discharged before surveillance staff approached them, and
caregivers not consenting to participate. Of the 3724 AGE cases en-
rolled, 1860 (49.9%) were rotavirus positive. The mean number of
rotavirus cases enrolled per year was 620 (range, 282–875) and the
mean proportion of rotavirus positive specimens ranged from the low-
est of 42% in 2009 to the highest of 56% in 2011 (Fig. 1). AGE
Fig. 1. Seasonality of RVGE among hospitalized <5 years old children.

UNCORRECTED PROOF
Vaccine xxx (2017) xxx-xxx 3
cases were admitted to YCH year round and rotavirus positivity was
also detected year round. Among them, AGE due to rotavirus was
found to be high from November through February, the peak being in
January and February, comprising 63–83% positivity each year (Fig.
1).
Among both rotavirus positive and negative cases, male predom-
inance was observed. The proportion of male children accounted for
61.7% in RVGE and 59.2% in non-RVGE. The male to female ratio
in RVGE was 1.6:1 and in non-RVGE was 1.5:1 though this was not
statistically significant.
By age group, the 6–11 months age group (47.8%) was found
to be significantly affected by rotavirus (P < 0.01), followed by the
12–23 months age group (34.0%) and the 0–5 months age group
(12.2%) and finally, the 24–59 months age group (6.0%). Both overall
and annually, around 60% of the RVGE cases were found among chil-
dren aged <12 months, and over 90% among children aged <24months
(Table 1).
Hospitalized AGE cases commonly presented with vomiting, fever
and dehydration. Among them, vomiting (84.1% Vs 67.9%; P < .01),
fever (84.5% Vs 75.6%; P < .01) and dehydration (78% Vs 69%;
P < .01) were more frequently observed in RVGE compared with
non-RVGE patients. Furthermore, oral rehydration (66.1% Vs
Table 1
Characteristics and clinical presentations of hospitalized children with rotavirus diar-
rhea and non-rotavirus diarrhea—Myanmar, 2009–2014 (N = 3724).
Characteristics RVGE (%) Non RVGE (%) P value
N =1860 (49.9%) N =1864 (50.1%)
Sex
Male 1147 (61.7%) 1103 (59.2%) 0.12
Female 713 (38.3%) 761 (40.8%)
Age in months
0–5 months 226 (12.2%) 370 (19.8%) <0.01
6–11 months 889 (47.8%) 666 (35.7%)
12–23 months 633 (34.0%) 524 (28.1%)
24–59 months 112 (6.0%) 304 (16.3%)
Clinical Symptoms (%) Yes
Vomiting 1564 (84.1%) 1266 (67.9%) <0.01
Fever 1515 (84.5%) 1409 (75.6%) <0.01
Dehydration 1447 (77.8%) 1291 (69.4%) <0.01
Treatment
ORS 1230 (66.1%) 1137 (60.9%) <0.01
IV fluids 649 (34.9%) 589 (31.6%) <0.05
Outcome
Died 1 (0.05%) 1(0.05%) 1.0
60.9%; P < .01) and intravenous rehydration (34.9% Vs 31.6%;
P < .05) were more fre quently required in RVGE. (Table 1).
Distribution of rotavirus genotypes by seasonal years (from July
to June of the following year) is shown in Table 2. In 2009 from
January to June, G1P[8] was predominant (15/43, 35%) followed by
G12P[8] (6/43, 14%). The G12P[8] genotype which started to emerge
in early 2009 became predominant, detected in 61.7% (50/81), 75.2%
(103/137) and 26% (45/173) of all strains detected in 2009–2010,
2010–2011 and 2011–2012 respectively. In 2012–2013, G2P[4] be-
came the most common genotype, accounting for 73.3% (22/30) of all
detected strains. The most prevalent genotype detected in 2013–2014
was G1P[8] accounting for 41.9% (31/74) followed by G9P[8] ac-
counting 27% (20/74) which increased up to 51.3% (20/39) of all
strains detected in the second half of 2014.
4. Discussion
We found that rotavirus was responsible for a considerable propor-
tion of AGE admissions among children under 5 years of age in Myan-
mar during 2009–2014. Every year, >80% of eligible AGE cases were
enrolled and stool samples were collected and tested from all enrolled
cases. The percentage of rotavirus positivity ranged from 42% to 56%,
highlighting that about half of the hospitalized AGE cases were due to
rotavirus. This proportion is similar to that of the neighboring coun-
tries; 53.4% in Kolkata in 2011–2013 [9], 44% in 39 inpatient stud-
ies conducted between 1998 and 2013 in China [10]. According to the
WHO’s rotavirus surveillance network data, the median rotavirus de-
tection in the Southeast Asian Region was 35% in 2012–2013 with
Myanmar having the highest proportion at 47% [11]. We addition-
ally demonstrated that rotavirus infection has a strong seasonal peak
in colder, drier months as seen in other Asian countries [12].
RVGE in this study showed male predominance which is in accor-
dance with the findings of other studies, such as in Lahore, where 60%
of enrolled children were male [13], and Uganda, where 61% of chil-
dren were male [14]. The factors underlying this difference are poorly
understood and further study is warranted. Regarding the age distrib-
ution of RVGE patients, the highest proportion rotavirus positive was
among children 6–11 months of age; over 60% were infants and more
than 90% of the cases were <2 years of age. These results are in line
with previous studies conducted prior to vaccine introduction in other
countries [15]. Thus, the WHO’s recommended dosing is applicable
in Myanmar [16].
Table 2
Rotavirus strain distribution in Myanmar, January 2009 to December 2014 (N =577).
January to June
2009 N (%)
July 2009-June
2010 N (%)
July 2010-June
2011 N (%)
July 2011-June
2012 N (%)
July 2012-June
2013 N (%)
July 2013-June
2014 N (%)
July to December
2014 Total
G1P[8] 15 (34.9)7 (8.6) 2 (1.5) 30 (17.3) 0 31 (41.9) 3 (7.7) 88
(15.3)
G1P[6] 3 (7) 0 0 2 (1.2) 0 0 0 5 (0.9)
G2P[4] 1 (2.3) 1 (1.2) 14 (10.2) 9 (5.2) 22 (73.3) 3 (4.1) 1 (2.6) 51
(8.8)
G2P[6] 0 0 0 0 2 (6.7) 2 (2.7) 0 4 (0.7)
G2P[8] 0 0 0 0 0 0 1 (2.6) 1 (0.2)
G3P[8] 5 (11.6) 0 0 0 0 0 0 5 (0.9)
G9P[4] 0 0 0 1 (0.6) 1 (3.3) 2 (2.7) 0 4 (0.7)
G9P[8] 0 0 0 7 (4) 1 (3.3) 20 (27) 20 (51.3) 48
(8.3)
G12P[8] 6 (14) 50 (61.7) 103 (75.2) 45 (26) 1 (3.3) 0 0 205
(35.5)
G12P[6] 3 (7) 7 (8.6) 8 (5.8) 37 (21.4) 2 (6.7) 0 0 57
(9.9)
Mixed 6 (14) 5 (6.2) 2 (1.5) 11 (6.4) 0 0 1 (2.6) 25
(4.3)
Partially
typed
4 (9.2) 4 (4.9) 7 (5.1) 26 (15) 1 (3.3) 14 (18.9) 10 (25.6) 66
(11.4)
Untypable 0 7 (8.6) 1 (0.7) 5 (2.9) 0 2 (2.7) 3 (7.7) 18
(3.1)
Total 43 (1 00) 81 (1 00) 137 (1 00) 173 (1 00) 30 (1 00) 74 (1 00) 39 (1 00) 577
(1 0 0)

UNCORRECTED PROOF
4 Vaccine xxx (2017) xxx-xxx
Regarding the clinical presentations, vomiting, fever and dehydra-
tion were found to be significantly associated with rotavirus positiv-
ity. This finding is also in line with the findings of other studies where
RVGE cases presented with more severe clinical manifestations com-
pared to rotavirus negative cases [17]. Of enrolled children, two pa-
tients expired: both were 10 months-old female and had some degree
of dehydration; one case was rotavirus positive and the other was ro-
tavirus negative.
This study identified a different profile of genotype distribution by
seasonal years. G1P[8] was predominant in the first 6 months of 2009,
whereas G12P[8] was predominant from 2009-2010, 2010–2011 and
2011–2012. However, G2P[4] became the most predominant strain in
2012–2013; while G1P[8] was predominant in 2013–2014 and G9P[8]
from July to December 2014. This changing pattern is in agreement
with the WHO’s genotype distribution reported both globally and re-
gionally with higher resemblance to regionally reported strains [18].
From 2009-2011, the globally dominant genotype was G1P[8], how-
ever G12P[9] strain emerged in Southeast Asian Region starting from
2010. The dominant strain globally as well as regionally in 2013 and
2014 was G1P[8] and G9P[8] respectively [12,17,19,20]. Diversity of
circulating rotavirus strains is reported regardless of vaccine use [21].
Enhanced surveillance is needed from the perspective of devising fu-
ture vaccine strategies and monitoring strain specific vaccine effec-
tiveness.
This analysis was subject to limitations. We conducted surveil-
lance at one site only and therefore our data may not be representative
for Myanmar as a whole. However, although surveillance was con-
ducted at YCH only, this is the largest children’s hospital and the ma-
jor pediatric referral centre in Myanmar, and likely captures the epi-
demiological profile of a large portion of our population of interest.
5. Conclusion
RVGE burden in Myanmar warrants consideration of rotavirus
vaccine introduction. The high proportion of RVGE in chil-
dren < 5 years hospitalized with AGE, the diversity of circulating
genotypes and epidemiological patterns reported in this study provide
vital input for vaccine programmers in planning vaccine introduction
and monitoring vaccine impact and effectiveness in Myanmar.
Conflict of interest
None to declare.
Acknowledgement
We would like to thank the World Health Organization for funding
this project and the Christian Medical College, Vellore, India for pro-
viding PCR primers for genotyping. We are also grateful to the Board
of Directors (DMR) for encouraging conduct this project and special
thanks are to the medical superintendent and AGE patients at YCH for
their permission to collect specimens.
References
[1] GBD 2013 Mortality and Causes of Death Collaborators. Global, regional, and
national age-sex specific all-cause and cause-specific mortality for 240 causes of
death, 1990-2013: a systematic analysis for the Global Burden of Disease Study
2013. Lancet 2015; 385: p. 117–71.
[2] J.E. Tate, A.H. Burton, C. Boschi-Pinto, U.D. Parashar, Global, regional, and na-
tional estimates of rotavirus mortality in children <5 years of age, 2000–2013,
Suppl Clin Infect Dis 62 (2) (2016) S96–S105.
[3] World Health Organization. Weekly epidemiological record. Rotavirus vaccines
WHO position paper- January 2013: 49–64.
[4] World Health Organization, Immunization, vaccines and biologicals, WHO,
Geneva, Switzerland, 2010.
[5] National Health Plan (2011-2016) Planning series 3. Yangon, Myanmar: Depart-
ment of Planning and Statistic, Ministry of Health, Union of Myanmar, 2006.
[6] Kyaw Moe, Hlaing Myat Thu, Win Mar Oo, Khin May Aye, Thin Thin Shwe,
Win Mar, C.D. Kirkwood, Genotyping of rotavirus isolates collected from chil-
dren less than 5 years of age admitted for diarrhoea at the Yangon Children’s
Hospital, Myanmar, Vaccine 27 (Suppl) (2009) 89–92.
[7] Bresee J, Parashar U, Holman R, et al. Generic protocol for hospital-based sur-
veillance to estimate the burden of rotavirus gastroenteritis in children under five
years of age. Field test version. 2002. Available at: www.who.int/vaccines-docu-
ments/pdf.Accessed 3 May 2008.
[8] The Wellcome Trust Research Laboratory, Manual of RT-PCR for genotyping
Group A Rotavirus strains, Christian Medical College, Vellore, India, 2012.
[9] Mullick S, Mandal P, Nayak MK et al. Hospital based surveillance and genetic
characterization of rotavirus strains in children with acute gastroenteritis in
Kolkata. 11th International Rotavirus Symposium, New Delhi (3-5 September)
2014; p. 106.
[10] D. Wu, C. Yen, Z.D. Yin, et al., The public health burden of rotavirus disease in
children younger than five years and considerations for rotavirus vaccine intro-
duction in China, Pediatr Infect Dis J 35 (12) (2016 Dec) e392–e398.
[11] World Health Organization, Vaccine preventable diseases surveillance, Global
Rotavirus Surveillance Inform Bulletin 9 (2014) 2–3.
[12] S.J. Jyotsna, Rajiv Sarkar, Denise Castronovo, et al., Seasonality of rotavirus in
South Asia: a meta-analysis approach assessing associations with temperature,
precipitation, and vegetation index, PLoS ONE. 5 (2012 May 7) e38168.
[13] Afzal MF, Sultan MA. Disease frequency and seasonality of rotavirus diarrhea in
children younger than 5 years of age in Lahore. 11th International Rotavirus
Symposium, New Delhi (3-5 September) 2014; p. 109.
[14] Bakainaga A, Anyua H, Kisakye A, et.al. Rotavirus infection among under-five
year old children admitted in Mulago National Referral hospital, June
2006-2013, Uganda. 11th International Rotavirus Symposium, New Delhi (3-5
September) 2014; p. 158.
[15] S. Mehendale, S. Venkatasubramanian, C.P. Girish Kumar, G. Kang, M.D.
Gupte, R. Arora, Expanded Indian national rotavirus surveillance network in the
context of rotavirus vaccine introduction, Indian Pediatr 53 (7) (2016 Jul 8)
575–581.
[16] World Health Organization, Rotavirus vaccines, WHO position paper January
2013. No. 5, 2013, 88, p. 49–64 http://www.who.int/wer.
[17] Fodha I, Fredj BH, Bouraoui H, et.al. Influence of children on rotavirus illness
severity. In: 11th International Rotavirus Symposium, New Delhi (3-5 Septem-
ber) 2014; p. 101.
[18] World Health Organization, Independent strategic review of the global rotavirus
surveillance network, Global Rotavirus Inform Surveillance Bulletin 8 (2013) 4.
[19] World Health Organization, Vaccine Preventable Diseases Surveillance, Global
Rotavirus Surveillance Inform Bulletin 6 (2012) 12–15.
[20] World Health Organization, Vaccine preventable diseases surveillance, Global
Rotavirus Surveillance Inform Bulletin 10 (2015) 2–4.
[21] E. Leshem, B. Lopman, R. Glass, J. Gentsch, K. Bányai, U. Parashar, M. Patel,
Distribution of rotavirus strains and strain-specific effectiveness of the rotavirus
vaccine after its introduction: a systematic review and meta-analysis, Lancet In-
fect Dis 14 (9) (2014 Sep) 847–856, https://doi.org/10.1016/
S1473-3099(14)70832-1, Epub 2014 Jul 28.