Sensitivity, Specificity, and Public-Health Utility of Clinical Case Definitions Based on the Signs and Symptoms of Cholera in Africa

Johara Nadri Agence de Medecine Préventive, Paris, France;

Search for other papers by Johara Nadri in
Current site
Google Scholar
PubMed
Close
,
Delphine Sauvageot Agence de Medecine Préventive, Paris, France;

Search for other papers by Delphine Sauvageot in
Current site
Google Scholar
PubMed
Close
,
Berthe-Marie Njanpop-Lafourcade Agence de Medecine Préventive, Paris, France;

Search for other papers by Berthe-Marie Njanpop-Lafourcade in
Current site
Google Scholar
PubMed
Close
,
Cynthia S. Baltazar Instituto Nacional de Saúde, Ministry of Health, Maputo, Mozambique;

Search for other papers by Cynthia S. Baltazar in
Current site
Google Scholar
PubMed
Close
,
Abiba Banla Kere Institut National d’Hygiène, Lomé, Togo;
Ministry of Health, Lomé, Togo;

Search for other papers by Abiba Banla Kere in
Current site
Google Scholar
PubMed
Close
,
Godfrey Bwire Control of Diarrheal Diseases, Community Health Department, Ministry of Health, Kampala, Uganda;

Search for other papers by Godfrey Bwire in
Current site
Google Scholar
PubMed
Close
,
Daouda Coulibaly Institut National d’Hygiène Publique, Abidjan, Côte d’Ivoire;

Search for other papers by Daouda Coulibaly in
Current site
Google Scholar
PubMed
Close
,
Adele Kacou N’Douba Institut Pasteur de Côte d’Ivoire, Abidjan, Côte d’Ivoire;

Search for other papers by Adele Kacou N’Douba in
Current site
Google Scholar
PubMed
Close
,
Atek Kagirita Central Public Health Laboratory, Ministry of Health, Kampala, Uganda;

Search for other papers by Atek Kagirita in
Current site
Google Scholar
PubMed
Close
,
Sakoba Keita Division Prévention et Lutte contre la Maladie, Ministry of Health, Conakry, Guinea;

Search for other papers by Sakoba Keita in
Current site
Google Scholar
PubMed
Close
,
Lamine Koivogui Institut National de Santé Publique, Conakry, Guinea;

Search for other papers by Lamine Koivogui in
Current site
Google Scholar
PubMed
Close
,
Dadja E. Landoh Division de l’Epidémiologie, Ministry of Health, Lomé, Togo;

Search for other papers by Dadja E. Landoh in
Current site
Google Scholar
PubMed
Close
,
Jose P. Langa Instituto Nacional de Saúde, Ministry of Health, Maputo, Mozambique;

Search for other papers by Jose P. Langa in
Current site
Google Scholar
PubMed
Close
,
Berthe N. Miwanda Institut National de Recherche Biomédicale, Kinshasa, Democratic Republic of Congo;

Search for other papers by Berthe N. Miwanda in
Current site
Google Scholar
PubMed
Close
,
Guy Mutombo Ndongala Division Provinciale de la Santé du Nord Kivu, Goma, Democratic Republic of Congo;

Search for other papers by Guy Mutombo Ndongala in
Current site
Google Scholar
PubMed
Close
,
Elibariki R. Mwakapeje Epidemiology and Diseases Control Section, Preventive Department, Ministry of Health, Community Development, Gender, Elderly and Children, Dar es Salaam, Tanzania;

Search for other papers by Elibariki R. Mwakapeje in
Current site
Google Scholar
PubMed
Close
,
Jacob L. Mwambeta Curative Department, National Health Laboratory Quality Assurance and Training Center, Ministry of Health, Community Development, Gender, Elderly and Children, Dar es Salaam, Tanzania

Search for other papers by Jacob L. Mwambeta in
Current site
Google Scholar
PubMed
Close
,
Martin A. Mengel Agence de Medecine Préventive, Paris, France;

Search for other papers by Martin A. Mengel in
Current site
Google Scholar
PubMed
Close
, and
Bradford D. Gessner Agence de Medecine Préventive, Paris, France;

Search for other papers by Bradford D. Gessner in
Current site
Google Scholar
PubMed
Close

During 2014, Africa reported more than half of the global suspected cholera cases. Based on the data collected from seven countries in the African Cholera Surveillance Network (Africhol), we assessed the sensitivity, specificity, and positive and negative predictive values of clinical cholera case definitions, including that recommended by the World Health Organization (WHO) using culture confirmation as the gold standard. The study was designed to assess results in real-world field situations in settings with recent cholera outbreaks or endemicity. From June 2011 to July 2015, a total of 5,084 persons with suspected cholera were tested for Vibrio cholerae in seven different countries of which 35.7% had culture confirmation. For all countries combined, the WHO case definition had a sensitivity = 92.7%, specificity = 8.1%, positive predictive value = 36.1%, and negative predictive value = 66.6%. Adding dehydration, vomiting, or rice water stools to the case definition could increase the specificity without a substantial decrease in sensitivity. Future studies could further refine our findings primarily by using more sensitive methods for cholera confirmation.

INTRODUCTION

Cholera remains a major public-health issue in developing countries. In Africa, 3,221,050 suspected cases were notified to the World Health Organization (WHO) from 1970 to 2011, which represented 46% of all suspected cases reported worldwide.1 During 2014, Africa reported 105,287 cases, which represented an increase of 87% compared with the previous year2; The Democratic Republic of Congo (DRC), Ghana, and Nigeria recorded 83% of all African cases.3 Case fatality ratios were higher than 5% only in African countries and included Cameroon, Côte d’Ivoire, Guinea Bissau, and Kenya and Africa recorded 84% of deaths globally. The high values for disease burden in sub-Saharan Africa, however, may underestimate substantially the total disease burden. For example, some countries likely under-notified cases possibly because of fear of stigmatization: Gabon and Central African Republic did not report any cases from 2008 and 2004, respectively, even though their neighboring countries were regularly affected with cholera epidemics.4,5

From 75% to 80% of persons infected with cholera do not develop symptoms, but may still represent an important source of transmission.6 The primary symptoms among affected persons are acute watery diarrhea and vomiting, which may lead to severe dehydration.7,8 The mean incubation period varies from several hours to 5 days.8,9

Efficient cholera surveillance and accurate disease burden estimation rely on a clear case definition with high sensitivity and specificity. The current case definition for cholera in epidemic settings recommended by WHO was shown in Haiti to have high sensitivity (91%; 95% confidence interval [CI]: 90–93%) but low specificity (43%; 95% CI: 40–47%).10 Moreover, information on the performance of various case definitions is lacking in Africa and may differ from other areas due to differences in cholera presentation (e.g., through cholera interaction with other common acute or chronic diseases,1 antibiotic pretreatment that may alter disease course, and different intestinal microbial flora). Because of the common lack of laboratory diagnostic capacity in Africa outside the national capital cities,1113 clinicians and public health staff must rely more on clinical case definitions.

The African Cholera Surveillance Network (Africhol) was launched in 2009 with funding from the Bill & Melinda Gates Foundation. African Cholera Surveillance Network was implemented in 11 countries through 2015 and was designed to assess cholera burden in enhanced surveillance sites and outbreak sites using clinical diagnosis and subsequent laboratory confirmation.14 To assess the performance of different cholera case definitions in Africa, we used data from the Africhol surveillance network and outbreak sites to analyze the clinical signs of suspected cholera cases by age group using positive bacterial culture as the gold standard.

MATERIALS AND METHODS

Study design.

African Cholera Surveillance Network was implemented in the first countries during 2011 with the collaboration of national ministries of health (MoH). One or more enhanced surveillance areas were identified in each country. Eligibility for inclusion as an Africhol-enhanced surveillance zone was based on meeting all of the following requirements: recent history of high cholera incidence; yearly outbreaks or identification of cholera; hospitals or cholera treatment centers (CTCs) that received most of the patients with cholera; and a laboratory that could perform stool culture for Vibrio cholerae. Patients with suspected cholera that were hospitalized or treated in a CTC were included in the analysis. In most sites and most years, cholera was seasonal and consequently CTCs were established only after the first cases during the year were identified; at this point, all patients with suspected cholera usually were referred to these centers.

In addition to surveillance for endemic disease within the enhanced surveillance zones, outbreak investigations were conducted when the national surveillance system reported cholera cases in sites not included in the enhanced surveillance zones. The same methodology (tools, case definition, and training) as in the enhanced surveillance areas was used. Because of budgetary constraints, we focused our efforts on selected outbreaks that occurred in locations not yet studied but having adequate laboratory facilities available. A total of 27 sites in the seven target countries provided data.

Data collection.

Data were collected on a standardized case report form at each participating treatment center by trained clinical staff used by the National MoH; forms included epidemiological, clinical, behavioral, and laboratory information. Participating treatment centers were all specialized diarrheal disease treatment facilities, either CTCs or diarrheal treatment wards at regional referral hospitals.

We provided standard training to clinical staff at study sites in the evaluation of patients during for the presence of dehydration, difficulty breathing, altered consciousness, dry mucous membranes, and coma. Vomiting, diarrhea, watery stools, rice water stools, and mucous or bloody stools were assessed at the health facility through clinical examination. Patients self-reported symptoms such as nausea, abdominal pain, leg cramps, and number of stools during the 24 hours before admission. Deaths were only ascertained when they occurred in the clinic. All data were entered into the Africhol database.

Culture confirmation.

In principle, our study aimed to collect rectal swabs from each suspected cholera case. Practical constraints in the field (e.g., clinical staff becoming overwhelmed during the peak of an outbreak) prevented realization of this goal. National public health laboratories in each country performed culture confirmation of suspected cases. In case of work overload during large outbreaks, clinicians were advised to collect rectal swabs from the first 10 patients admitted in the health facility each day, as 10 samples per day was the maximum amount that each laboratory could process. Stool samples were enriched in alkaline peptone water and plated on thiosulfate-citrate-bile-salt-sucrose (TCBS) agar. When characteristic yellow colonies were identified on TCBS and the oxidase test was positive, the patient was considered positive for V. cholerae.15

Cholera case definition for the Africhol network.

In areas with known cholera, a suspected case was defined as a patient aged 2 years or more that developed acute watery diarrhea, with or without vomiting. In any particular site, the actual use of the case definitions was unknown even though yearly training was providing to the clinical staff during monitoring visits on sites and questionnaire guidelines were provided. A confirmed cholera case was a person in whom V. cholerae was identified by culture and positive oxidase test.

Participating countries.

Seven countries had more than 35 confirmed cholera cases and were included in the analysis: Côte d’Ivoire, DRC, Guinea, Mozambique, Tanzania, Togo, and Uganda. Prospective surveillance started on different dates for each country. Outside the surveillance zones, cases were identified during outbreak investigations (Table 1). The cut-off point for this analysis was July 2015.

Table 1

Participating countries, continuous surveillance sites, and outbreak investigation sites

CountriesAreasHealth facilitiesStartEnd
SurveillanceCôte d’IvoireKoumassi-Port Bouet-Vridi districtInfectious disease and pediatric departments of Port Bouet and Koumassi HospitalsAugust 2011January 2015
Vridi Health Center Temporary CTC
DRCGoma and Karimsibi districtsCTCs in General Provincial HospitalAugust 2011July 2015
Buhimba CTC
Kiziba temporary cholera treatment unit
GuineaFive districts of ConakryInfectious disease and pediatric departments of Donka hospitalJuly 2011November 2013
Additional CTC in Ratoma neighborhood opened during the 2012 epidemic
MozambiqueBeira cityPonta-Gea health centerOctober 2011August 2014
Macurrungo health center
Munhava health center
Macurrungo health center
Beira central hospital
TanzaniaMagu districtMagu HospitalJanuary 2012March 2013
Mwanza districtMwanza hospital
TogoFive districts of Lome and Golfe districtInfectious disease and pediatric departments of the Center Hospitalier Universitaire—Be HospitalJune 2011April 2015
Other district health centers in which a temporary CTC
Lake district of the Maritmie regionInfectious disease and pediatric departments of Aneho Hospital
Health centers with temporary CTCs
UgandaButaleja districtNamatela health centerDecember 2011June 2015
Manafwa districtBukigai health center
Mbale districtBusiu health center
OutbreakCôte d’IvoireAdiake districtAdiake general hospitalMay 2012October 2012
Temporary CTC
Adjame-Plateau-Attecoube, Cocody, East Yopougon, West YopougonAdjame-Plateau-Attecoube, Cocody health centersOctober 2014January 2015
Yopougon general hospital
DRCKinshasa cityKingabwa CTCAugust 2011February 2012
Malaku CTC
Massina Cholera treatment unit
GuineaBoffa, Coyah, Dubreka Forecariah, Kindia prefecturesBoffa health centerFebruary 2012March 2013
Coyah CTC
Dubreka hospital
Forecariah CTC
Manke CTC
MozambiqueCuamba districtCuamba CTCJanuary 2012February 2012
Montepuez districtMontepuez CTCFebruary 2012February 2012
Nampula districtNampula CTCFebruary 2013February 2013
January 2014May 2014
Pemba cityTemporary CTCJanuary 2013May 2014
TanzaniaDar-Es-Salam and Temeke districtTemeke CTCDecember 2011April 2012
UgandaKasese districtBwera hospitalOctober 2011October 2012
Kayangi health center
Kagando hospital
Kinyamaseke health center
Kitholhu health center
Other temporary treatment centers

CTC = cholera treatment center; DRC = Democratic Republic of Congo.

Statistical analysis.

Overall and within countries, we analyzed clinical symptoms stratified by the cumulative number of suspected, culture-tested, and culture-confirmed cases. Results are presented by age group and duration of hospitalization. We did not analyze data by site as with 27 different sites case counts were often low, increasing the risk of type 1 errors.

To assess the association between individual symptoms and a positive stool culture for V. cholerae, we conducted an analysis limited to patients with a culture result available and aged 1 year or more. Children aged less than 1 year were excluded because of a small sample size (Table 3). All symptom variables were dichotomous. We present unadjusted results from univariate analysis for the association between each individual symptom or sign and culture-confirmed cholera. For each individual symptom or sign that predicted culture positivity on univariate analysis at a significance level of P ≤ 0.10, we created a multivariate adjusted model that simultaneously adjusted for gender, age (entered as a linear variable), and country (using DRC as the referent category as a large majority of cases were reported from DRC). We created a separate model for each symptom or sign as symptoms and signs were highly correlated with each other. We considered a multilevel analysis but decided against it because of the small number of countries in our dataset16; as noted earlier, instead we opted for stratification to assess the exact degree to which results varied. Missing data represented between 2.4% and 18.6% of cases for individual signs and symptoms (Table 2) and these missing data were evenly distributed over time and across analyzed age groups. Based on this, we decided not to apply multiple imputations for missing data. Data analysis was conducted by Stata version 12 (StataCorp, College Station, TX).

Table 2

Repartition of missing data in clinical symptoms and signs of suspected, culture-tested, and culture-confirmed cholera cases in seven countries of the Africhol

Clinical symptom or signSuspected = 9,391, n (%)Culture tested = 5,084, n′ (%)Culture confirmed = 1,816, n″ (%)
Watery stool629 (6.7)121 (2.4)42 (2.3)
Vomiting1,248 (13.3)411 (8.1)163 (8.9)
Dehydration1,306 (13.9)476 (9.4)185 (10.2)
Rice water stools914 (9.7)215 (4.2)64 (3.5)
Nausea2,616 (27.9)917 (18.0)349 (19.2)
Dry mucous membranes2,855 (30.4)944 (18.6)367 (20.2)
Abdominal pain1,852 (19.7)639 (15.6)250 (13.8)
Leg cramp2,075 (22.1)697 (13.7)259 (14.3)
Mucous stool1,029 (11.0)270 (5.3)84 (4.6)
Altered consciousness2,298 (24.5)711 (14.0)268 (14.8)
Difficulty breathing2,285 (24.3)701 (13.8)268 (14.8)
Coma2,314 (24.6)712 (14.0)270 (14.9)
Bloody stool969 (10.3)221 (4.4)58 (3.2)

Africhol = African Cholera Surveillance Network. n, n′, n″ = total of missing data for symptoms.

Considering the availability of symptoms for each suspect cholera case recorded in the Africhol database, we used three different categories of case definitions:

  1. 1.World Health Organization case definition: WHO has two case definitions for cholera, one for areas where disease is not known to be present and one for areas where there is a cholera epidemic. As Africhol sites were selected based, in part, on the known presence of cholera, we assessed only the second, defined as a patient aged more than 5 years old and who develops acute watery diarrhea, with or without vomiting.17
  2. 2.Ad hoc case definitions developed from combinations of different clinical signs that were identified through multivariate analysis.
  3. 3.Country-specific national integrated disease surveillance and response definitions:
    • Côte d’Ivoire, DRC: a patient with severe diarrhea and dehydration or death caused from acute watery diarrhea.

    • Togo: a patient aged five or more years with severe diarrhea and dehydration or death caused from acute watery diarrhea.

    • Mozambique: a patient aged one or more years with acute diarrhea of sudden onset, with or without vomiting or dehydration and regardless of the appearance of the diarrhea.

    • Guinea: a patient aged one or more years with watery diarrhea, with or without vomiting.

    • Tanzania and Uganda: a patient aged two or more years with acute watery diarrhea of sudden onset, with or without vomiting.

Based on culture results as the gold standard, we assessed the sensitivity and specificity of different cholera case definitions by country, identification in an enhanced surveillance zone or outbreak site, and age group. For illustrative purposes, we also present positive and negative predictive values, recognizing that these depend on the a priori likelihood of disease in the study population.

Ethics.

African Cholera Surveillance Network provided technical and financial resources to national MoHs to support cholera surveillance. Cholera is part of the national public health surveillance through the integrated disease surveillance and response system supported by WHO. The Africhol protocol was approved and implemented by the MoH of each country. The Togolese government further elected to submit the protocol for approval to the national institutional review board (IRB). The remaining countries did not seek IRB approval as they considered that they were conducting epidemic disease surveillance and response covered by national public health laws as an integral part of the public health mandate of the MoH and associated executing agencies.

RESULTS

A total of 9,391 suspected cholera cases were enrolled in the study from June 2011 to July 2015, with almost half (43.4%) from the DRC, followed by Guinea (20%) and Mozambique (14%) (Table 3). More than 54% of the suspected cases (5,084) had a culture test done, of which 1,816 (35.7%) were confirmed for V. cholerae. Goma and its suburbs in DRC represented 60% of culture-confirmed cases. The proportion of men and women was approximately the same within the categories of suspected and culture-tested cases.

Table 3

Characteristics of suspected, culture-tested, and culture-confirmed cholera cases in seven countries of the Africhol (N = 9,391)

CountryNotification siteSuspectedCulture tested (percentage of suspected)Culture confirmed (percentage of culture tested)
Côte d’IvoireTotal193136 (70.5)45 (33.1)
Surveillance1126618
Outbreak817027
DRCTotal4,0743,278 (80.5)1,177 (35.9)
Surveillance3,4682,9381,041
Outbreak606340136
GuineaTotal1,865261 (14.0)95 (36.4)
Surveillance1,32014352
Outbreak54511843
MozambiqueTotal1,317430 (32.6)79 (18.4)
Surveillance3942952
Outbreak92313577
TanzaniaTotal15163 (36.4)39 (61.9)
Surveillance11780
Outbreak345539
TogoTotal630573 (91.0)293 (51.1)
Surveillance587540282
Outbreak433311
UgandaTotal1,161343 (29.5)88 (25.7)
Surveillance39815923
Outbreak76318465
Sex
Female4,5222,532 (56.0)842 (33.3)
Male4,5432,510 (55.2)955 (38.0)
Missing32642 (12.9)19 (45.2)
Age group
< 13122 (71.0)3 (13.6)
1–2312226 (72.4)53 (23.4)
2–3461342 (74.2)114 (33.3)
3–4424279 (65.8)122 (43.7)
4–5324234 (72.2)95 (40.6)
(5–14)1,9561,188 (60.7)469 (39.5)
(15–59)5,1332,499 (48.7)874 (35.0)
> 60584266 (45.5)78 (29.3)
Unknown16628 (17.4)8 (28.6)
Total9,3915,084 (54.1)1,816 (35.7)

Africhol = African Cholera Surveillance Network; DRC = Democratic Republic of Congo.

Within the category of diarrhea, 91.2% of suspected cases and 92.4% of confirmed cases were classified as watery diarrhea, whereas 65.3% of suspected cases and 68.4% of confirmed cases were classified as rice water diarrhea (Table 4). Culture-tested and culture-confirmed cases had approximately the same distribution of symptoms, including a high proportion with watery stools and dehydration, whereas suspected cases were less likely to report any particular symptom. The occurrence of vomiting was different according to age group with 83.5% among children less than age 5 years and 87.0% among persons over age 5 years. Among suspected cases, those < 5 and at least 5 years of age, respectively, had different levels of dry mucous membranes (57.7% versus 50.7%), abdominal pain (55.4% versus 39.2%), and leg cramps (55.6% versus 31.3%) (Table 5). The same difference between age groups was observed for culture confirmed cases except for vomiting (Table 6).

Table 4

Clinical symptoms and signs of suspected, culture-tested, and culture-confirmed cholera cases in seven countries of the Africhol; P values are shown for associations between symptoms or signs among suspected or culture-tested cases using culture confirmed cases as the referent category

Clinical symptom or signSuspectedCulture testedCulture confirmed
n/N (%) (P value)n′/N′ (%) (P value)n″/N″ (%)
Watery stool7,984/8,762 (91.2) (< 0.001)4,585/4,963 (92.4) (0.18)1,651/1,774 (93.1)
Vomiting7,030/8,143 (86.3) (0.003)3,972/4,673 (85.0) (< 0.001)1,503/1,653 (90.9)
Dehydration6,331/6,334 (78.3) (< 0.001)4,071/4,608 (88.4) (< 0.001)1,550/1,631 (95.0)
Rice water stools5,538/8,477 (65.3) (< 0.001)3,331/4,869 (68.4) (< 0.001)1,327/1,752 (75.7)
Nausea4,864/6,775 (71.8) (< 0.001)3,228/4,167 (77.5) (< 0.001)1,193/1,467 (81.3)
Dry mucous membranes3,654/6,536 (55.9) (< 0.001)2,716/4,160 (65.6) (< 0.001)1,037/1,449 (71.6)
Abdominal pain3,953/7,316 (52.4) (< 0.001)2,471/4,445 (55.6) (0.55)861/1,566 (55.0)
Leg cramp3,758/7,316 (51.4) (< 0.001)2,312/4,387 (52.7) (< 0.001)928/1,557 (59.6)
Mucous stool461/8,362 (5.5) (< 0.001)340/4,814 (7.1) (0.969)122/1,732 (7.0)
Altered consciousness354/7,093 (5.0) (0.54)194/4,373 (4.4) (0.019)84/1,548 (5.4)
Difficulty breathing262/7,106 (3.7) (0.084)148/4,383 (3.4) (0.064)63/1,551 (4.0)
Coma81/7,077 (1.1) (0.11)54/4,372 (1.2) (0.16)24/1,546 (1.6)
Bloody stool48/8,422 (0.6) (0.055)31/4,863 (0.6) (0.004)2/1,758 (0.1)

Africhol = African Cholera Surveillance Network. n, n′, n″ = total of presence of symptoms/N, N′, N″ = total of the cases with data available

Table 5

Distribution of clinical symptoms and signs by age group for all suspected cholera cases (N = 9,230)

Watery stool10+ stools in 24 hoursVomitingDehydrationRice water stoolsNauseaDry mucous membra-nesAbdominal painLeg crampMucous stoolAltered conscious-nessDifficulty breathingComaBloody stool
Age < 5 years1,3575081,1711,1931,004874633500391835244183
Percentage92.138.683.585.669.275.350.739.231.35.84.03.41.40.2
Age ≥ 5 years6,5232,0605,7675,0984,4653,9603,0033,4193,3233743012146345
Percentage91.135.487.077.464.671.457.755.455.65.55.33.71.10.7
P value0.1850.0280.001< 0.0010.0010.007< 0.001< 0.001< 0.0010.7050.0640.5410.3720.041
Table 6

Distribution of clinical symptoms and signs by age group for confirmed cholera cases (N = 1,808)

Watery stool10+ stools in 24 hoursVomitingDehydrationRice water stoolsNauseaDry mucous membra-nesAbdominal painLeg crampMucous stoolAltered conscious-nessDifficulty breathingComaBloody stool
Age < 5 years3611723323512992552191411153116770
Percentage94.848.390.797.079.584.264.642.535.08.34.52.02.00.0
Age ≥ 5 years1,2865621,1671,1951,025936816716812916856172
Percentage92.747.491.094.574.880.673.6558.266.36.75.74.11.40.2
P value0.1640.7480.8840.0540.0560.1510.001< 0.001< 0.0010.3040.3860.0250.4550.458

Of all suspected cases, those hospitalized for more than 1 day had more frequent stooling, (i.e., > 10 stools in the 24 hours before admission), nausea, vomiting, dehydration, dry mucous membranes, abdominal pain, and leg cramps than those hospitalized less than 1 day (Table 7). By contrast, among confirmed cases, only those hospitalized for more than 3 days were more frequently diagnosed with dehydration (91.7% versus 87.9%), as well as having between 10 and 25 stools in the last 24 hours (49.8% versus 28.1%), nausea (83.4% versus 56.7%), and dry mucous membranes (76.6% versus 50.0%) compared with those hospitalized less than a day (Table 8).

Table 7

Distribution of clinical symptoms and signs by duration of hospitalization for all suspected cases (N = 5,768)

Watery stool10+ stools in 24 hoursVomitingDehydrationRice water stoolsNauseaDry mucous mem-branesAbdominal painLeg crampMucous stoolAltered consciousnessDifficulty breathingComaBloody stool
Hosp. less than 1 day15528133931227363594868314
Percentage*90.118.582.157.471.455.344.137.131.23.55.31.90.70.6
Hosp. 1 day93424080063473355542841540267312769
Percentage94.326.084.167.175.266.648.645.343.66.93.42.90.70.9
Hosp. 2 days1,5495511,4261,2201,2511,018848776808885842136
Percentage94.735.889.577.177.773.157.551.753.15.53.82.80.90.4
Hosp. 3 days+2,7851,0642,5742,3642,2001,9011,5161,4781,514144153883314
Percentage94.941.491.684.476.177.463.655.558.75.05.93.41.30.5

Referral category.

P value < 0.05 for an association between the symptom and likelihood of one or more hospital days (i.e., odds ratio > 1).

Table 8

Distribution of clinical symptoms and signs by duration of hospitalization for culture-confirmed cholera cases (N = 1,253)

Watery stool10+ stools in 24 hoursVomitingDehydrationRice water stoolsNauseaDry mucous membranesAbdominal painLeg crampMucous stoolAltered consciousnessDifficulty breathingComaBloody stool
Hosp. less than 1 day3392929311715152037210
Percentage*100.028.190.687.994.056.750.045.569.09.123.36.53.30.0
Hosp. 1 day13959117128115102897765777420
Percentage99.344.783.692.882.180.367.456.247.55.15.22.91.50.0
Hosp. 2 days31314129528825925520917617226131130
Percentage98.146.293.792.681.786.771.658.557.18.44.33.61.00.0
Hosp. 3 days+73934969671362257152140642744403192
Percentage97.549.893.397.182.583.476.656.860.45.95.14.11.30.3

Referral category.

P value < 0.05 for an association between the symptom and likelihood of one or more hospital days.

P value < 0.05 for an association between the symptom and likelihood of less than one hospital day (i.e., odds ratio < 1).

Rice water stools, vomiting, nausea, severe dehydration, dry mucous membranes, leg cramps, and altered consciousness predicted positive stool culture in univariate analysis. Bloody stools predicted a negative stool culture, although this symptom was reported for only 31 patients (Table 9). When adjusting for gender, age, and country, all of these symptoms remained associated with culture positivity.

Table 9

Distribution of collected samples and culture confirmation status in seven member countries of the Africhol

CultureUnivariateMultivariate
SymptomsNPositive N (%)OR95% CIOR95% CI
Watery stoolNo378123 (32.5)Ref.
Yes4,5851,651 (36.0)1.170.93–1.46
Number of stools in 24 hours0–92,382814 (34.1)Ref.Ref.
10–251,951735 (37.6)1.171.03–1.321.171.02–1.34
Rice water stoolsNo1,538425 (27.6)Ref.Ref.
Yes3,3311,327 (39.8)1.731.52–1.981.861.60–2.17
Mucous stoolNo4,4741,610 (36.0)Ref.
Yes340122 (35.9)0.990.79–1.25
Bloody stoolNo4,8321,756 (36.3)Ref.Ref.
Yes312 (6.5)0.120.03–0.510.080.02–0.34
NauseaNo939274 (29.2)Ref.Ref.
Yes3,2281,193 (37.0)1.421.22–1.671.341.13–1.59
VomitingNo701150 (21.4)Ref.Ref.
Yes3,9721,503 (37.8)2.231.85–2.712.001.64–2.46
Dry mucus membranesNo1,424412 (28.9)Ref.Ref.
Yes2,7161,037 (38.2)1.511.32–1.741.571.35–1.82
DehydrationNo53781 (15.1)Ref.Ref.
Yes4,0711,550 (38.1)3.462.71–4.424.983.75–6.62
Abdominal painNo1,974705 (35.7)Ref.
Yes2,471861 (34.8)0.960.85–1.09
Leg crampsNo2,075629 (30.3)Ref.Ref.
Yes2,312928 (40.1)1.451.27–1.641.591.39–1.82
Difficulty breathingNo4,2351,488 (35.1)Ref.
Yes14863 (42.6)1.370.98–1.91
Altered consciousnessNo4,1791,464 (35.0)Ref.Ref.
Yes19484 (43.3)1.421.06–1.891.411.04–1.92
ComaNo4,3181,522 (35.3)Ref.
Yes5424 (44.4)1.430.86–2.52

Africhol = African Cholera Surveillance Network; CI = confidence interval; OR = odds ratio; Ref. = referent category. A separate multivariate model was created for each symptom that predicted culture positivity on univariate analysis by simultaneously adjusting from gender, age (for those at least age 1 year), and country.

For all countries combined, the WHO case definition had a high sensitivity (92.7%; 95% CI: 91.2–94.0) but low specificity (8.1%; 95% CI: 7.1–9.3) among persons aged at least 5 years. For those aged 1–4 years, the WHO case definition had a sensitivity of 94.0% (95% CI: 91.1–96.1) and specificity of 6.1% (95% CI: 4.4–8.1) (Tables 10 and 11). A case definition combining watery stools or rice water stools with either dehydration or vomiting substantially improved specificity without lowering sensitivity. The definition that achieved the highest specificity was rice water stools associated with the occurrence of less than 10 stools in the last 24 hours: 70.2% for cases aged more than 5 years and 65.1% for cases aged 1–4 years.

Table 10

Sensitivity, specificity, PPV, and NPV for tested cases aged more than 5 years

Case definitionSensitivity % (95% CI)Specificity % (95% CI)PPV % (95% CI)NPV % (95% CI)
World Health Organization epidemic cholera definition92.7 (91.2–94.0)8.1 (7.1–9.3)36.1 (34.5–37.7)66.6 (60.9–71.9)
Watery stool and dehydration92.1 (90.5–93.6)17.3 (15.8–18.9)38.2 (36.5–40.0)79.9 (76.1–83.3)
Watery stool and vomiting88.8 (86.9–90.5)21.0 (19.3–22.7)38.3 (36.5–40.1)77.2 (73.7–80.4)
Watery stool and ≤ 9 stools in last 24 hours51.2 (48.3–54.1)46.5 (44.4–48.7)34.5 (32.3–36.8)63.4 (61.0–65.8)
Watery stool and (10–25) stools in last 24 hours47.0 (44.1–49.9)57.3 (55.2–59.4)37.7 (35.2–40.2)66.3 (64.1–68.4)
Watery stool and dry mucous membranes72.2 (69.5–74.8)34.1 (32.0–36.2)37.2 (35.2–39.3)69.4 (66.4–72.2)
Watery stool and dehydration and vomiting84.9 (82.8–86.8)26.6 (24.8–28.5)39.3 (37.4–41.1)75.9 (72.7–78.9)
Watery stool and dehydration and ≤ 9 stools in last 24 hours46.7 (43.8–49.6)58.5 (56.3–60.6)38.7 (36.1–41.3)66.2 (64.0–68.3)
Watery stool and dehydration and (10–25) stools in last 24 hours46.7 (43.8–49.6)58.3 (56.2–60.4)37.9 (35.4–40.5)66.7 (64.5–68.9)
Watery stool and (dehydration or vomiting)96.7 (95.6–97.6)10.7 (9.4–12.1)37.7 (36.0–39.4)85.3 (80.6–89.2)
Watery stool and (dehydration or rice water stools)95.9 (94.7–96.9)14.6 (13.1–16.1)38.6 (36.9–40.4)86.4 (82.5–89.7)
Watery stool and (dehydration or rice water stools or vomiting)97.2 (96.1–98.0)10.0 (8.8–11.3)37.8 (36.1–39.6)86.2 (81.4–90.2)
Watery stool and vomiting and dry mucous membranes68.6 (65.8–71.3)40.1 (38.0–42.3)38.3 (36.2–40.5)70.2 (67.5–72.8)
Rice water stools74.8 (72.4–77.0)36.6 (34.7–38.6)40.1 (38.2–42.0)71.9 (69.3–74.4)
Rice water stools and dehydration75.6 (73.1–77.9)36.1 (34.2–38.2)39.8 (37.8–41.8)72.6 (69.9–75.2)
Rice water stools and vomiting73.4 (70.9–75.8)41.2 (39.2–43.3)41.1 (39.1–43.2)73.5 (70.9–75.9)
Rice water stools and ≤ 9 stools in last 24 hours40.2 (37.4–43.1)70.2 (68.2–72.1)42.7 (39.8–45.7)68.0 (66.0–69.9)
Rice water stools and (10–25) stools in last 24 hours40.3 (37.5–43.2)62.9 (60.8–65.0)37.5 (34.9–40.3)65.6 (63.5–67.6)
Rice water stools and dry mucous membranes60.9 (58.0–63.8)46.0 (43.9–48.2)38.1 (35.8–40.4)68.4 (65.9–70.9)
Rice water stools and dehydration and vomiting70.8 (68.2–73.4)42.5 (40.4–44.6)40.9 (38.8–43.0)72.2 (69.7–74.6)
Rice water stools and dehydration and ≤ 9 stools in last 24 hours38.5 (35.6–41.4)69.4 (67.2–71.5)43.0 (40.0–46.2)65.2 (63.1–67.3)
Rice water stools and dehydration and (10–25) stools in last 24 hours39.0 (36.2–41.9)66.0 (64.0–68.0)37.3 (34.6–40.1)67.6 (65.6–69.5)
Rice water stools and vomiting and dry mucous membranes58.1 (55.2–61.1)51.1 (48.9–53.3)39.3 (37.0–41.8)69.1 (66.7–71.4)

CI = confidence interval; NPV = negative predictive value; PPV = positive predictive value.

Table 11

Sensitivity, specificity, PPV, and NPV for tested cases aged between 1 and 4 years

Case definitionSensitivity % (95% CI)Specificity % (95% CI)PPV % (95% CI)NPV % (95% CI)
World Health Organization epidemic cholera definition94.0 (91.1–96.1)6.1 (4.4–8.1)36.1 (33.1–39.2)64.1 (51.1–75.7)
Watery stool and dehydration95.0 (92.2–97.0)19.0 (16.1–22.2)39.3 (36.1–42.7)87.3 (80.7–92.3)
Watery stool and vomiting87.9 (84.2–91.1)23.3 (20.1–26.7)39.1 (35.7–42.5)77.6 (71.1–83.2)
Watery stool and ≤ 9 stools in the last 24 hours50.4 (45.1–55.7)44.4 (40.5–48.4)33.7 (29.7–37.9)61.5 (56.9–66.0)
Watery stool and (10–25) stools in the last 24 hours47.9 (42.6–53.2)59.4 (55.5–63.3)39.8 (35.1–44.6)67.0 (63.0–70.9)
Watery stool and dry mucous membranes63.8 (58.5–68.9)57.4 (53.4–61.3)45.3 (40.8–49.9)74.2 (70.0–78.0)
Watery stool and dehydration and vomiting85.5 (81.4–89.0)28.8 (25.3–32.5)40.1 (36.6–43.7)78.1 (72.2–83.2)
Watery stool and dehydration and ≤ 9 stools in the last 24 hours47.9 (42.5–53.2)58.9 (54.9–62.7)39.3 (34.6–44.1)67.0 (62.9–70.9)
Watery stool and dehydration and (10–25) stools in the last 24 hours47.9 (42.5–53.2)59.1 (55.1–63.0)39.9 (35.1–44.7)66.7 (62.5–70.6)
Watery stool and (dehydration or vomiting)97.8 (95.6–99.0)12.6 (10.2–15.5)38.5 (35.3–41.7)91.0 (83.1–96.0)
Watery stool and (dehydration or rice water stools)97.2 (94.9–98.6)18.4 (15.4–21.6)39.8 (36.5–43.1)92.2 (86.1–96.2)
Watery stool and (dehydration or rice water stools or vomiting)98.0 (96.0–99.2)12.5 (10.0–15.3)38.6 (35.4–41.8)91.9 (83.9–96.7)
Watery stool and vomiting and dry mucous membranes57.5 (52.1–62.8)61.7 (57.7–65.6)45.6 (40.8–50.4)72.3 (68.2–76.1)
Rice water stools78.8 (74.4–82.8)30.0 (26.6–33.7)38.9 (35.4–42.5)71.5 (65.9–76.7)
Rice water stools and dehydration80.2 (75.7–84.2)29.8 (26.3–33.5)38.7 (35.2–42.3)73.1 (67.3–78.4)
Rice water stools and vomiting73.4 (68.5–77.9)35.9 (32.2–39.7)39.0 (35.3–42.8)70.7 (65.5–75.6)
Rice water stools and ≤ 9 stools in the last 24 hours39.5 (34.3–44.8)65.1 (61.2–68.8)38.3 (33.1–43.7)66.6 (62.6–70.4)
Rice water stools and (10–25) stools in the last 24 hours44.3 (39.1–49.7)62.7 (58.8–66.5)40.0 (35.1–45.1)66.7 (62.8–70.5)
Rice water stools and dry mucous membranes54.5 (49.0–60.0)61.0 (57.1–64.9)43.1 (38.3–48.0)71.3 (67.2–75.2)
Rice water stools and dehydration and vomiting71.8 (66.8–76.5)36.3 (32.5–40.1)38.6 (34.8–42.4)69.8 (64.5–74.7)
Rice water stools and dehydration and ≤ 9 stools in the last 24 hours40.3 (34.9–45.9)64.7 (60.7–68.5)38.3 (33.1–43.7)66.6 (62.6–70.4)
Rice water stools and dehydration and (10–25) stools in the last 24 hours41.5 (36.4–46.7)63.8 (59.9–67.5)39.8 (34.9–44.9)65.3 (61.4–69.1)
Rice water stools and vomiting and dry mucous membranes47.9 (42.5–53.4)64.8 (60.8–68.6)43.2 (38.1–48.4)69.0 (65.0–72.8)

CI = confidence interval; NPV = negative predictive value; PPV = positive predictive value.

Data were similar when stratified by country, although some modest and occasional differences were found (Supplemental Tables 111). For example, the WHO case definition’s sensitivity was lower in Mozambique (64.5%; 95% CI: 52.7–75.1) than that in other countries (from 87.5% to 98.9%). Similarly, the WHO case definition had a lower sensitivity in outbreak sites versus in the enhanced surveillance zones (75.8% versus 97.8%) (Supplemental Tables 1215). The WHO case definition sensitivity and specificity did not change by individual age year for children between 1 and 4 years old (Supplemental Table 16). Country-specific case definitions had variable sensitivity and specificity, with sensitivity varying from 46.5% in Côte d’Ivoire to 100% in Mozambique and specificity varying from 4.1% in Mozambique and Uganda to 63.0% in Togo (Supplemental Table 17).

DISCUSSION

In using data from 27 sites located in seven African countries, we found that the WHO case definition for cholera had high sensitivity and low specificity overall and among all subgroups. This is likely because almost all true cholera cases typically had acute watery diarrhea but most patients with acute watery diarrhea did not have cholera, even in outbreak sites. Use of a highly sensitive definition facilitates the timely detection of outbreaks and may be appropriate given the potential for cholera to rapidly cause large outbreaks or national epidemics.1821 However, the lack of specificity may lead to excessive interventions or inappropriate distribution of resources. Adding the presence of other symptoms to the WHO case definition used during outbreaks—such as dehydration, vomiting, or rice water stools—substantially increased specificity with only a minimal decrease in sensitivity. Other studies of cholera have shown that the same clinical signs were associated with culture positivity among suspected cholera cases.10 For example, in Bangladesh, diarrhea, vomiting, and dehydration were more frequent in culture-positive cases.22 As in Haiti,23 we found that persons with longer hospitalization periods were more likely to have dehydration and signs of dehydration compared with persons hospitalized less than a day.

The WHO cholera case definition excludes children aged less than 5 years old to avoid expending resources on investigation of routine infant diarrhea.1 This policy is based on the assumption that a much higher percentage of early childhood diarrhea will result from causes other than cholera when compared with older persons. However, this assumption may be incorrect. In our study, the WHO epidemic cholera case definition had low specificity and was similar for persons aged between 1 and 4 years and 5+ years of age, a finding that occurred across all settings and within individual year ages among children aged 1–4 years. We also found that 21% of all confirmed cases were aged 1–4 years, which is in agreement with data from other studies in cholera endemic areas that have found cholera is a primary cause of diarrhea in this age group.24 These data suggest that the WHO case definition should incorporate children aged between 1 and 4 years, as has been done with many national case definitions.

Our study had several limitations. We did not have complete knowledge of why persons did or did not have a stool culture performed, nor did we have complete knowledge of the source populations from which our cases derived. This limitation could reduce the ability to apply our results to other settings. Nevertheless, our data were derived from seven different countries and results were similar across sites. Moreover, our data reflect results from real-world settings where clinical and MoH staff have limited data availability, including on source populations and reasons for the lack of stool culture.

Because of laboratory limitations in African field settings, we used culture confirmation as the gold standard despite a reported sensitivity of 66%.25 Consequently, a substantial portion of cases reported as culture negative may have had true cholera.14 This limitation should be addressed in future studies by use of a combination of several more sensitive testing methods.26

Differing levels of antibiotic pretreatment might impact subsequent case definition performance. To address this issue, we originally had planned to collect data on antibiotic use; however, these data were usually not entered on the data collection form and where data were included were incomplete (for example, the name of the medication or whether it was an antibiotic). Future studies might address this issue by testing urine antibiotic activity. Ideally, with further resources and funding available, we would have chosen to test the clinical definitions also against polymerase chain reaction to increase sensitivity followed by determination of the optimal clinical definition across all methods. We would encourage funders and researchers to consider investing in such a study.

Our study was conducted in limited areas and thus may not be representative of the remainder of participating countries or of nonparticipating countries, although the relative consistency of results suggests that this was not a major limitation. During cholera outbreaks, clinical and public health staff can be overwhelmed and thus may not complete case report forms fully and rectal swabs could not be collected for all patients; moreover, it is possible that staff focused on symptoms they thought were associated with cholera.

To our knowledge, this is the only recent study specifically on cholera symptoms in Africa. Our results suggest that the WHO case definition could be made more specific without a substantial decrease in sensitivity by adding one or more symptoms. However, the observed heterogeneity in the adherence of clinical staff to an established case definition across countries raises concerns of whether a more complex and hence specific case definition could be reliably implemented at least in the sites where we worked.

Furthermore, our results support inclusion of children aged between 1 and 4 years old in the case definition when cholera is known to be present. Besides improving estimates of disease burden, this would have the benefit of sensitizing clinicians, public-health workers, and stakeholders to the risk of cholera among young children, allowing for the mobilization of additional prevention efforts. Future studies could refine our findings primarily by using more sensitive methods for cholera confirmation.

Supplementary Material

Acknowledgments:

We thank all the focal points involved in the project: Didier Bompangue, Mireille Dosso, Sakoba Keita, Lamine Koivogui, Abiba Kere Banla, Ilesh Jani, Issa Makumbi and Peter Mmbuji; all the country coordinators, surveillance officers, laboratory coordinators and technicians, and country data managers. We acknowledge all the physicians, nurses, and clinical staff of each site, all the patients and their families for taking the time to respond to our surveillance questionnaire. We also thank all the members of the Africhol consortium and the other members of the Africhol coordination team: Aline Munier, Souleymane Sakande, Issaka Ouedraogo, Themba Mhlanga, Leonard Heyerdahl, Richard Wood, and Alexandre Blake.

REFERENCES

  • 1.

    Mengel MA, Delrieu I, Heyerdahl L, Gessner BD, 2014. Cholera outbreaks in Africa. Curr Top Microbiol Immunol 379: 117144.

  • 2.

    WHO, 2014. Weekly Epidemiological Record, Vol. 89, 345356. Available at: http://www.who.int/wer/2014/wer8931/en/. Accessed June 2, 2016.

    • PubMed
    • Export Citation
  • 3.

    WHO, 2015. Weekly Epidemiological Record, Vol. 90, 517544. Available at: http://www.who.int/wer/2015/wer9040/en/. Accessed June 2, 2016.

    • PubMed
    • Export Citation
  • 4.

    WHO, 2005. Weekly Epidemiological Record, Vol. 80, 261268. Available at: http://www.who.int/wer/2005/wer8031.pdf?ua=1.

    • PubMed
    • Export Citation
  • 5.

    WHO, 2009. Weekly Epidemiological Record, Vol. 84, 309324. Available at: http://www.who.int/wer/2009/wer8431.pdf?ua=1.

    • PubMed
    • Export Citation
  • 6.

    Gangarosa EJ, Mosley WH, 1974. Epidemiology and surveillance of cholera. Barua D, Burrows W, eds. Cholera. Philadelphia, PA: W.B. Saunders, 381403.

    • PubMed
    • Export Citation
  • 7.

    WHO, 2016. Cholera. Available at: http://www.who.int/mediacentre/factsheets/fs107/en/. Accessed June 3, 2016.

    • PubMed
    • Export Citation
  • 8.

    Kaper JB, Morris JG, Levine MM, 1995. Cholera. Clin Microbiol Rev 8: 4886.

  • 9.

    Azman AS, Rudolph KE, Cummings DAT, Lessler J, 2013. The incubation period of cholera: a systematic review. J Infect 66: 432438.

  • 10.

    Lucien MAB, Schaad N, Steenland MW, Mintz ED, Emmanuel R, Freeman N, Boncy J, Adrien P, Joseph GA, Katz MA, 2015. Identifying the most sensitive and specific sign and symptom combinations for cholera: results from an analysis of laboratory-based surveillance data from Haiti, 2012–2013. Am J Trop Med Hyg 92: 758764.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11.

    Petti CA, Polage CR, Quinn TC, Ronald AR, Sande MA, 2006. Laboratory medicine in Africa: a barrier to effective health care. Clin Infect Dis 42: 377382.

  • 12.

    Debes AK, Ateudjieu J, Guenou E, Ebile W, Sonkoua IT, Njimbia AC, Steinwald P, Ram M, Sack DA, 2016. Clinical and environmental surveillance for Vibrio cholerae in resource constrained areas: application during a 1-year surveillance in the far north region of Cameroon. Am J Trop Med Hyg 94: 537543.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13.

    Njanpop-Lafourcade B-M, Hugonnet S, Djogbe H, Kodjo A, N’douba AK, Taha MK, Stoeckel P, Gessner BD, 2013. Mobile microbiological laboratory support for evaluation of a meningitis epidemic in northern Benin. PLoS ONE 8: e68401.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14.

    Sauvageot D et al. 2016. Cholera incidence and mortality in sub-Saharan African sites during multi-country surveillance. PLoS Negl Trop Dis 10: e0004679.

  • 15.

    Perilla MJ, Ajello G, Bopp C, Elliot J, Facklam R, Snapp JS, Popovic T, Wells J, Dowell SE, Centers for Disease Control and Prevention, 2003. Manual for the Laboratory Identification and Antimicrobial Susceptibility Testing of Bacterial Pathogens of Public Health Importance in the Developing World. Atlanta, GA: Centers for Disease Control and Prevention.

    • PubMed
    • Export Citation
  • 16.

    Bryan ML, Jenkins SP, 2016. Multilevel modelling of country effects: a cautionary tale. Eur Sociol Rev 32: 322.

  • 17.

    WHO, 2008. Prevention and Control of Cholera Outbreaks: WHO Policy and Recommendations. Available at: http://www.who.int/cholera/technical/prevention/control/en/index1.html. Accessed June 3, 2016.

    • PubMed
    • Export Citation
  • 18.

    WHO, 2009. Weekly Epidemiological Record, Vol. 84, 109116. Available at: http://www.who.int/wer/2009/wer8414/en/. Accessed June 3, 2016.

    • PubMed
    • Export Citation
  • 19.

    Ahmed S, Bardhan PK, Iqbal A, Mazumder RN, Khan AI, Islam MS, Siddique AK, Cravioto A, 2011. The 2008 cholera epidemic in Zimbabwe: experience of the icddr,b team in the field. J Health Popul Nutr 29: 541546.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 20.

    Barzilay EJ, Schaad N, Magloire R, Mung KS, Boncy J, Dahourou GA, Mintz ED, Steenland MW, Vertefeuille JF, Tappero JW, 2013. Cholera surveillance during the Haiti epidemic—the first 2 years. N Engl J Med 368: 599609.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 21.

    WHO, 2010. Cholera in Haiti—Update 4. Available at: http://www.who.int/csr/don/2010_11_24/en/. Accessed June 3, 2016.

    • PubMed
    • Export Citation
  • 22.

    Weil AA, Khan AI, Chowdhury F, LaRocque RC, Faruque ASG, Ryan ET, Calderwood SB, Qadri F, Harris JB, 2009. Clinical outcomes in household contacts of patients with cholera in Bangladesh. Clin Infect Dis 49: 14731479.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 23.

    Valcin C-L et al. 2013. Predictors of disease severity in patients admitted to a cholera treatment center in urban Haiti. Am J Trop Med Hyg 89: 625632.

  • 24.

    Kotloff KL et al. 2013. Burden and aetiology of diarrhoeal disease in infants and young children in developing countries (the Global Enteric Multicenter Study, GEMS): a prospective, case-control study. Lancet 382: 209222.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 25.

    Alam M et al. 2010. Diagnostic limitations to accurate diagnosis of cholera. J Clin Microbiol 48: 39183922.

  • 26.

    Page A-L, Alberti KP, Mondonge V, Rauzier J, Quilici M-L, Guerin PJ, 2012. Evaluation of a rapid test for the diagnosis of cholera in the absence of a gold standard. PLoS One 7: e37360.

    • PubMed
    • Search Google Scholar
    • Export Citation

Author Notes

Address correspondence to Johara Nadri, Agence de Médecine Préventive, 21 Boulevard Pasteur, Paris 75015, France. E-mails: jnadri@aamp.org or johara.nadri@gmail.com

Financial support: The Africhol project received funding from the Bill & Melinda Gates foundation (grant number: OPPGH5233), administered by the Agence de Medecine Preventive (AMP), Paris, France.

Disclosures: AMP receives grant-specific funding from Sanofi Pasteur, which through its subsidiary Shanta, manufactures cholera vaccine. However, AMP does not and has never received funding from Sanofi Pasteur for work on cholera or cholera vaccines. J. N., B. D. G., B.-M. N.-L., and D. S. report grants from Bill & Melinda Gates Foundation, during the conduct of the study; grants from Crucell, GSK, Hilleman Laboratories, Merck, Novartis, Sanofi, outside the submitted work.

Author contributions: B. D. G., M. A. M., D. S., and J. N. conceived and designed the experiments. B. M. N. L. implemented the laboratory network and followed all logistic and diagnostic activities. C. S. B., A. B. K., G. B., D. C., A. K. N., A. K., S. K., L. K., D. E. L., J. P. L., B. N. M., G. M. N., E. R. M., and J. L. M., performed the experiments. J. N., B. D. G., and D. S., analyzed the data. J. N. wrote the first draft of the manuscript. B. D. G. and M. A. M. has provided critical input into the first draft of the manuscript. D. S., B. D. G., M. A. M., B. M. N. L., and G. B. have contributed to the final draft of the current manuscript. B. D. G. performed this work while employed by AMP. He current works for Pfizer Vaccines, Collegeville, PA.

Authors’ addresses: Johara Nadri, Delphine Sauvageot, Berthe-Marie Njanpop-Lafourcade, and Martin A. Mengel, Agence de Medecine Preventive, Paris, France, E-mails: jnadri@aamp.org, dsauvageot@aamp.org, blafourcade@aamp.org, and mmengel@aamp.org. Cynthia S. Baltazar and Jose P. Langa, Instituto Nacional de Saúde, Ministry of Health, Maputo, Mozambique, E-mails: cynthiasema@yahoo.com and jplanga1@yahoo.com.br. Abiba Banla Kere, Institut National d’Hygiène, Ministry of Health, Lomé, Togo, E-mail: kerebanla@yahoo.fr. Godfrey Bwire, Control of Diarrheal Diseases, Community Health Department, Ministry of Health, Kampala, Uganda, E-mail: cddmoh@yahoo.com. Daouda Coulibaly, Institut National d’Hygiène Publique, Abidjan, Côte d’Ivoire, E-mail: daocoul@yahoo.fr. Adele Kacou N’Douba, Institut Pasteur de Côte d’Ivoire, Abidjan, Côte d’Ivoire, E-mail: knadele@yahoo.fr. Atek Kagirita, Central Public Health Laboratory, Ministry of Health, Kampala, Uganda, E-mail: akagirita@gmail.com. Sakoba Keita, Division Prévention et Lutte contre la Maladie, Ministry of Health, Conakry, Guinea, E-mail: sakoba_keita@yahoo.fr. Lamine Koivogui, Institut National de Santé Publique, Conakry, Guinea, E-mail: koivoguil@gmail.com. Dadja E. Landoh, Division de l’Epidémiologie, Ministry of Health, Lomé, Togo, E-mail: mlandoh@yahoo.fr. Berthe N. Miwanda, Institut National de Recherche Biomédicale, Kinshasa, Democratic Republic of Congo, E-mail: bmiwanda@yahoo.fr. Guy Mutombo Ndongala, Division Provinciale de la Santé du Nord Kivu, Goma, Democratic Republic of Congo, E-mail: guymutombon@yahoo.fr. Elibariki R. Mwakapeje, Epidemiology and Diseases Control Section, Preventive Department, Ministry of Health, Community Development, Gender, Elderly and Children, Dar es Salaam, Tanzania, E-mail: emwakapeje75@gmail.com. Jacob L. Mwambeta, Curative Department, National Health Laboratory Quality Assurance and Training Center, Ministry of Health, Community Development, Gender, Elderly and Children, Dar es Salaam, Tanzania, E-mail: j.lusekelo@gmail.com. Bradford D. Gessner, Agence de Medecine Preventive, Paris, France and Pfizer Vaccines, Collegeville, PA, E-mail: bgessner@aamp.org.

These authors contributed equally to this work.

  • 1.

    Mengel MA, Delrieu I, Heyerdahl L, Gessner BD, 2014. Cholera outbreaks in Africa. Curr Top Microbiol Immunol 379: 117144.

  • 2.

    WHO, 2014. Weekly Epidemiological Record, Vol. 89, 345356. Available at: http://www.who.int/wer/2014/wer8931/en/. Accessed June 2, 2016.

    • PubMed
    • Export Citation
  • 3.

    WHO, 2015. Weekly Epidemiological Record, Vol. 90, 517544. Available at: http://www.who.int/wer/2015/wer9040/en/. Accessed June 2, 2016.

    • PubMed
    • Export Citation
  • 4.

    WHO, 2005. Weekly Epidemiological Record, Vol. 80, 261268. Available at: http://www.who.int/wer/2005/wer8031.pdf?ua=1.

    • PubMed
    • Export Citation
  • 5.

    WHO, 2009. Weekly Epidemiological Record, Vol. 84, 309324. Available at: http://www.who.int/wer/2009/wer8431.pdf?ua=1.

    • PubMed
    • Export Citation
  • 6.

    Gangarosa EJ, Mosley WH, 1974. Epidemiology and surveillance of cholera. Barua D, Burrows W, eds. Cholera. Philadelphia, PA: W.B. Saunders, 381403.

    • PubMed
    • Export Citation
  • 7.

    WHO, 2016. Cholera. Available at: http://www.who.int/mediacentre/factsheets/fs107/en/. Accessed June 3, 2016.

    • PubMed
    • Export Citation
  • 8.

    Kaper JB, Morris JG, Levine MM, 1995. Cholera. Clin Microbiol Rev 8: 4886.

  • 9.

    Azman AS, Rudolph KE, Cummings DAT, Lessler J, 2013. The incubation period of cholera: a systematic review. J Infect 66: 432438.

  • 10.

    Lucien MAB, Schaad N, Steenland MW, Mintz ED, Emmanuel R, Freeman N, Boncy J, Adrien P, Joseph GA, Katz MA, 2015. Identifying the most sensitive and specific sign and symptom combinations for cholera: results from an analysis of laboratory-based surveillance data from Haiti, 2012–2013. Am J Trop Med Hyg 92: 758764.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11.

    Petti CA, Polage CR, Quinn TC, Ronald AR, Sande MA, 2006. Laboratory medicine in Africa: a barrier to effective health care. Clin Infect Dis 42: 377382.

  • 12.

    Debes AK, Ateudjieu J, Guenou E, Ebile W, Sonkoua IT, Njimbia AC, Steinwald P, Ram M, Sack DA, 2016. Clinical and environmental surveillance for Vibrio cholerae in resource constrained areas: application during a 1-year surveillance in the far north region of Cameroon. Am J Trop Med Hyg 94: 537543.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13.

    Njanpop-Lafourcade B-M, Hugonnet S, Djogbe H, Kodjo A, N’douba AK, Taha MK, Stoeckel P, Gessner BD, 2013. Mobile microbiological laboratory support for evaluation of a meningitis epidemic in northern Benin. PLoS ONE 8: e68401.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14.

    Sauvageot D et al. 2016. Cholera incidence and mortality in sub-Saharan African sites during multi-country surveillance. PLoS Negl Trop Dis 10: e0004679.

  • 15.

    Perilla MJ, Ajello G, Bopp C, Elliot J, Facklam R, Snapp JS, Popovic T, Wells J, Dowell SE, Centers for Disease Control and Prevention, 2003. Manual for the Laboratory Identification and Antimicrobial Susceptibility Testing of Bacterial Pathogens of Public Health Importance in the Developing World. Atlanta, GA: Centers for Disease Control and Prevention.

    • PubMed
    • Export Citation
  • 16.

    Bryan ML, Jenkins SP, 2016. Multilevel modelling of country effects: a cautionary tale. Eur Sociol Rev 32: 322.

  • 17.

    WHO, 2008. Prevention and Control of Cholera Outbreaks: WHO Policy and Recommendations. Available at: http://www.who.int/cholera/technical/prevention/control/en/index1.html. Accessed June 3, 2016.

    • PubMed
    • Export Citation
  • 18.

    WHO, 2009. Weekly Epidemiological Record, Vol. 84, 109116. Available at: http://www.who.int/wer/2009/wer8414/en/. Accessed June 3, 2016.

    • PubMed
    • Export Citation
  • 19.

    Ahmed S, Bardhan PK, Iqbal A, Mazumder RN, Khan AI, Islam MS, Siddique AK, Cravioto A, 2011. The 2008 cholera epidemic in Zimbabwe: experience of the icddr,b team in the field. J Health Popul Nutr 29: 541546.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 20.

    Barzilay EJ, Schaad N, Magloire R, Mung KS, Boncy J, Dahourou GA, Mintz ED, Steenland MW, Vertefeuille JF, Tappero JW, 2013. Cholera surveillance during the Haiti epidemic—the first 2 years. N Engl J Med 368: 599609.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 21.

    WHO, 2010. Cholera in Haiti—Update 4. Available at: http://www.who.int/csr/don/2010_11_24/en/. Accessed June 3, 2016.

    • PubMed
    • Export Citation
  • 22.

    Weil AA, Khan AI, Chowdhury F, LaRocque RC, Faruque ASG, Ryan ET, Calderwood SB, Qadri F, Harris JB, 2009. Clinical outcomes in household contacts of patients with cholera in Bangladesh. Clin Infect Dis 49: 14731479.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 23.

    Valcin C-L et al. 2013. Predictors of disease severity in patients admitted to a cholera treatment center in urban Haiti. Am J Trop Med Hyg 89: 625632.

  • 24.

    Kotloff KL et al. 2013. Burden and aetiology of diarrhoeal disease in infants and young children in developing countries (the Global Enteric Multicenter Study, GEMS): a prospective, case-control study. Lancet 382: 209222.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 25.

    Alam M et al. 2010. Diagnostic limitations to accurate diagnosis of cholera. J Clin Microbiol 48: 39183922.

  • 26.

    Page A-L, Alberti KP, Mondonge V, Rauzier J, Quilici M-L, Guerin PJ, 2012. Evaluation of a rapid test for the diagnosis of cholera in the absence of a gold standard. PLoS One 7: e37360.

    • PubMed
    • Search Google Scholar
    • Export Citation
Past two years Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 1588 244 18
PDF Downloads 779 133 13
 
 
 
 
Affiliate Membership Banner
 
 
Research for Health Information Banner
 
 
CLOCKSS
 
 
 
Society Publishers Coalition Banner
Save