Early detection and response to meningococcal disease epidemics in sub-Saharan Africa: appraisal of the WHO strategy.

Research Early detection and response to meningococcal disease epidemics in sub-SaharanAfrica: appraisal of theWHOstrategy J.A.D. Leake,1 M.L. Kone,2 A.A. Yada,2 L.F. Barry,3 G. Traore,3 A. Ware,3 T. Coulibaly,4 A. Berthe,4 H. Mambu Ma Disu,4 N.E. Rosenstein,1 B.D. Plikaytis,1 K. Esteves,5 J. Kawamata,5 J.D. Wenger,5 D.L. Heymann,5 & B.A. Perkins1 Objective To assess the sensitivity, specificity and predictive value positive of the WHO threshold strategy for detecting meningococcal disease epidemics in sub-Saharan Africa and to estimate the impact of the strategy on an epidemic at district level. Methods Data on meningitis cases at the district level were collected weekly from health ministries, WHO country and regional offices, and nongovernmental organizations in countries where there were epidemics of meningococcal disease in 1997. An epidemic was defined as a cumulative district attack rate of at least 100 cases per 100 000 population from January to May, the period of epidemic risk. The sensitivity, specificity and predictive value positive of the WHO threshold rate were calculated, and curves of sensitivity against (1 – specificity) were compared with alternatively defined threshold rates and epidemic sizes. The impact of the WHO strategy on a district epidemic was estimated by comparing the numbers of epidemic cases with cases estimated to have been prevented by vaccination. Findings An analysis was made of 48 198 cases reported in 174 districts in Benin, Burkina Faso, the Gambia, Ghana, Mali, Niger, and Togo. These cases were 80.3% of those reported from Africa to WHO during the 1997 epidemic period. District populations ranged from 10 298 to 573 908. The threshold rate was crossed during two consecutive weeks in 69 districts (39.7%) and there were epidemics in 66 districts (37.9%). Overall, the sensitivity of the threshold rate for predicting epidemics was 97%, the specificity was 95%, and the predictive value positive was 93%. Taken together, these values were equivalent or better than the sensitivity, specificity and predictive value positive of alternatively defined threshold rates and epidemics, and remained high regardless of district size. The estimated number of potential epidemic cases decreased by nearly 60% in the age group targeted for vaccination in one district where the guidelines were followed in a timely manner. Conclusion The use of the WHO strategy was sensitive and specific for the early detection of meningococcal disease epidemics in countries of sub-Saharan Africa during 1997 and had a substantial impact on a district epidemic. Nevertheless, the burden of meningococcal disease in these countries remains formidable and additional control measures are needed. Keywords Meningitis, Meningococcal/epidemiology; Disease outbreaks/prevention and control; Meningococcal vaccines/therapeutic use; Immunization programs; Sensitivity and specificity; Predictive value of tests; Africa South of the Sahara (source: MeSH, NLM ). Mots cle´s Me´ningite me´ningococcique/e´pide´miologie; Epide´mie/pre´vention et controˆle; Vaccins antime´ningococciques/usage the´rapeutique; Programmes de vaccination; Sensibilite´ et spe´cificite´ (Epide´miologie); Valeur pre´dictive tests; Afrique subsaharienne (source: MeSH, INSERM). Palabras clave Meningitis meningoco´cica/epidemiologı´a; Brotes de enfermedades/prevencio´n y control; Vacunas meningoco´cicas/uso terape´utico; Programas de inmunizacio´n; Sensibilidad y especificidad; Valor predictivo de los tests; A´frica del Sur (fuente: DeCS, BIREME ). Bulletin of the World Health Organization 2002;80:342-349. Voir page 347 le re´sume´ en franc¸ais. En la pa´gina 347 figura un resumen en espan˜ol. Introduction Meningococcal disease epidemics occur throughout the world (1–5) but are most severe in the sub-Saharan African meningitis belt, where they have recurred since the early 20th century (6). This belt extends from Senegal to Ethiopia and includes 15 countries with an estimated population of over 270 000 000. Epidemics almost invariably take place during the dry season, i.e. from December until May, and end when the rainy season begins (7). During the dry season, meningococcus is presumed to cause meningitis. The terms ‘‘meningitis’’ and ‘‘meningococcal disease’’ are usually used synonymously. As yet it has been impossible to predict epidemics during a given year in particular districts or countries. Recently, epidemics of meningococcal disease have severely affected many countries of sub-Saharan Africa. The intervals between large-scale epidemics may be decreasing (8, 9). In 1996 and 1997,WHO reported 189 690 and 71 339 cases 1 Centers for Disease Control and Prevention, Atlanta, Georgia, USA. Correspondence should be sent to Dr John A.D. Leake, Division of Infectious Diseases, Children’s Hospital and Health Center, 3020 Children’s Way, MC 5041, San Diego, CA 92123, USA (email: jleake@chsd.org). 2 Office of the WHO Representative, Abidjan, Coˆte d’Ivoire. 3 World Health Organization and Ministry of Health, Burkina Faso. 4 World Health Organization and Ministry of Health, Mali. 5 Division of Emerging and Other Communicable Diseases and Expanded Programme on Immunization/Childhood Vaccine Initiative, World Health Organization, Geneva, Switzerland. Ref No. 01-1289 342 # World Health Organization 2002 Bulletin of the World Health Organization 2002, 80 (5) ofmeningitis respectively inAfrica (9). Case-fatalities exceeded 10% during both years. The epidemic in 1996 was the largest ever recorded in a single year, and substantial underreporting of both cases and deaths probably occurred (10). Both epidemics were caused by the serogroup A III.1 clone, which has circulated throughout the region for over a decade (11, 12), but serogroup C meningococci have also caused or significantly contributed to other recent epidemics (13). The WHO strategy for the control of epidemic meningo- coccal disease links the early prediction of epidemics with rapidly conducted mass vaccination campaigns (1). A threshold model of epidemic detection is used for early prediction. The threshold rate is 15 cases per 100 000 population per week averaged over two successive weeks. When the threshold strategy was first evaluated it was found to have reasonably good sensitivity (72– 93%), specificity (92–100%) and predictive value positive (65– 97%), and the population size considered most appropriate for epidemic prediction ranged from 30 000 to 100 000 (14). The potential impact of the strategy on epidemics depends on rapid detection and prompt mass vaccination campaigns. It has frequently been difficult to coordinate these activities in a timely manner in the meningitis belt (15). Meningococcal polysaccharide vaccines have been avail- able for several decades (16). Serogroups A and C are present in most such vaccines used in the region. These vaccines have repeatedly been shown to be more than 85% efficacious in the prevention of invasive meningococcal disease in Africa (17) and elsewhere. At least in developed countries, immunity lasts five years or longer among most healthy vaccinated adults (18). Since the introduction of the threshold strategy in 1995 an intensive multinational effort has promoted its use through the improvement of epidemic detection and response capacity. However, severe epidemics of meningococcal disease have continued to occur and the effectiveness of the strategy has therefore been questioned. The aims of the present study were to determine whether the strategy was still sufficiently sensitive and specific for detecting epidemics in sub-Saharan Africa and to estimate its impact on an epidemic at the district level. Methods Calculation of sensitivity, specificity and predictive value positive In 1997 we collected data weekly at the district level from health ministries, WHO country and regional offices, and nongovernmental organizations in every country where there were significant numbers of meningitis cases between January and May, the period of highest epidemic risk 1997. Incidence rates were calculated using denominators obtained from health ministries at the district, regional or national level. The districts were stratified into population quartiles. On the basis of the threshold rate of 15 cases per 100 000 population per week and using data from all districts included in the analysis, we calculated sensitivity, specificity and predictive value positive as aggregate values. The WHO definition of a district epidemic as an overall attack rate of at least 100 per 100 000 population per year was used. Districts with incomplete surveillance data were excluded. Two sets of curves of sensitivity against (1 – specificity) were constructed for different weekly incidence thresholds of meningitis. Data from all districts included in the analysis were used for the first set, whereas data from districts stratified by population quartile were used for the second. In order to assess whether alternative thresholds or epidemic definitions im- proved sensitivity or specificity relative to the WHO guide- lines, sensitivity and specificity were compared graphically for threshold values ranging from 2 to 26 cases per 100 000 po- pulation per week in increments of 2 and for alternative epidemic definitions of 50 and 75 per 100 000 population per year. Sensitivity, specificity and predictive value positive were also calculated in relation to a threshold definition of515 per 100 000 per week for two successive, not averaged, weeks. Cases prevented We collected data on weekly attack rates, vaccine coverage rates, and dates of a vaccination campaign in a medium-sized district (Reo District, Koudougou Region, Burkina Faso) in order to estimate the impact of the threshold strategy on an epidemic at this level. Meningococcal disease was defined as being present in persons with compatible symptoms and signs of meningitis and turbid cerebrospinal fluid (1). Confirmation by culture was not available. Coverage rates during the vaccination campaign were estimated on the basis of the proportion of control subjects vaccinated in a case-control study of vaccine efficacy conducted during April andMay 1997 (19). The controls were persons aged at least 5 years who were matched to cases by age category and place of residence. Interviews with subjects were conducted in their native language 7–14 days after a mass vaccination campaign with bivalent A/C meningococcal polysaccharide vaccine. The subjects were asked about their meningococcal vaccination status, which was confirmed by the examination of vaccination cards if these were available. For comparison the proportion of cases prevented per epidemic was estimated on the basis of vaccine coverage rates obtained from district ministry of health sources. The number of cases prevented by vaccination was estimated in the age group targeted during the vaccination campaign. This involved comparing attack rates among vaccinated and unvaccinated persons by means of the following equations (17). Weekly overall AR = (VAR x PV)/(NVAR x PNV); VAR= 0.15 xNVAR; 0.15 = (1 – VE). AR is attack rate; VAR, attack rate among vaccinated persons; NVAR, attack rate among unvaccinated persons; PV, proportion of population vaccinated; PNV, proportion of population not vaccinated; VE, vaccine efficacy, defined as 85% for the purposes of this analysis. The estimated numbers of cases prevented were summed two weeks after the end of the vaccination campaign in order to allow for the development of post-vaccination antibody responses (16). The proportion of the epidemic prevented was calculated as: estimated cases prevented/epidemic cases + estimated cases prevented. Statistical analysis EpiInfo (version 6.04) was used for data entry and analysis of weekly district-level meningitis cases, the case-control study of vaccine efficacy, and chi-square calculation for comparison of proportions of districts crossing the epidemic threshold stratified by population quartile. SAS (Release 6.12, SAS Institute, Cary, NC, USA) was used to construct tables of districts and thresholds, and curves of sensitivity against (1 – specificity). 343Bulletin of the World Health Organization 2002, 80 (5) Response to meningococcal disease epidemics Results An analysis was made of 48 198 cases in 174 districts in Benin, Burkina Faso, the Gambia, Ghana, Mali, Niger, and Togo (Table 1). These cases amounted to 80.3% of those reported from Africa to WHO during the 1997 epidemic period (2). A total of 69 districts (39.7%) crossed the WHO threshold and 66 districts (37.9%) had epidemics, i.e. at least 100 cases per 100 000 population. Three districts (1.7%) with incomplete data were excluded from the analysis. District populations ranged from 10 298 to 573 908 per- sons (median = 176 291); 33 districts (19%) had populations in the range 30 000 to 100 000; 75% had populations of at least 100 000, and 45% had populations of at least 200 000. The quartiles of district population in thousands were: <111 (n=40), 111–<176 (n=46), 176–<248 (n=47), and5 248 (n=41). The proportions of districts crossing the epidemic threshold for at least two consecutive weeks were 18%, 43%, 49% and 39% in the respective quartiles (P = 0.04) (Fig.1). Sensitivity, specificity and predictive value positive Using districts with cumulative attack rates of at least 100 per 100 000 population as the standard, the overall sensitivity of the threshold was 97.0%, the specificity was 95.4% and the predictive value positive 92.8% (Table 2). These parameters were similar when data from each country, not shown here, were examined separately. The examination of weekly thresh- old values ranging from 2 to 26 cases per 100 000, averaged over two weeks, revealed that the threshold of 15 per 100 000 population had comparable or better specificity and predictive value positive for a given sensitivity than other thresholds for predicting yearly epidemics of 100 cases/ 100 000 population (Fig. 2). The values did not change significantly when the sensitivity, specificity and predictive value positive were calculated on the basis of a threshold rate of 15 per 100 000 population for two successive but not averaged weeks (Table 3). For districts of the second largest and largest quartiles of population size the sensitivity decreased to 88% and 80% respectively, whereas the specificity remained uniformly good for all districts regardless of population size (Table 4). Cases prevented A total of 2881 cases (attack rate for whole district = 1119/ 100 000) and 316 deaths (case-fatality rate = 11%) occurred during a 12-week meningococcal disease epidemic in Reo District, Koudougou Region, Burkina Faso (estimated 1997 Table 1. African countries with large-scale meningococcal disease epidemics, 1997 Country Cases reported Number of cases Number of to WHO (up to analysed (% of districts 1 June 1997) total reported cases)a analysed Burkina Faso 21 504 18 377 (84.5) 50 Ghana 18 551 12 531 (67.5) 24 Mali 10 960 10 178 (92.9) 45 Niger 3922 4198 (107)a 38 Togo 2619 2072 (79.1) 3 The Gambia 913 807 (88.4) 1 Benin 360 35 (9.7) 13 Other countries 1181 – – Totala 60 010 48 198 (80.3) 174 Source: reference 2. a Includes cases reported to WHO for the period after 1 June 1997. For complete 1997 case figures, see reference 9. Table 2. Numbers of districts above and below weekly threshold of 15 per 100 000 averaged over two consecutive weeks vs yearly attack rate of 100 cases/100 000 Yearly attack rate Threshold rate 5 100 < 100 Total 515 64 5 69 <15 2 103 105 66 108 174 Sensitivity: 64/(64+2) = 97.0%; 95% confidence interval = 88.5–99.5%. Specificity: 103/(103+5) = 95.4%; 95% confidence interval = 89.0–98.3%. Predictive value positive: 64/(64+5) = 92.8%; 95% confidence interval = 83.29–7.3%. Predictive value negative: (103/103+2) = 98.1%; 95% confidence interval = 92.6–99.7%. 344 Bulletin of the World Health Organization 2002, 80 (5) Research population = 257 367). The epidemic threshold was crossed in early February and the weekly attack rate peaked at 159 per 100 000 in mid-March. Vaccination campaigns were per- formed over a period of six days and were completed within two weeks after the epidemic threshold was crossed for the second consecutive week. The epidemic subsided in late April. Among the 120 control subjects in a case-control study of vaccine efficacy living in the village of Reo, vaccine coverage amounted to 86%. A total of 4217 cases were estimated to have been prevented during the six weeks following the vaccination campaign while the epidemic subsided, allowing two weeks after the end of the campaign for the development of humoral immunity among vaccinated persons. This number repre- sented nearly 1.5 times that of observed cases, or 59.4% of the total number of epidemic cases predicted in the absence of vaccination (Fig. 3). On the assumption that the observed case- fatality rate remained constant at 11% it was estimated that 464 deaths were prevented by vaccination. Alternative data for the entire district obtained from the health ministry, viz. the number of doses given divided by the estimated target population, indicated vaccine coverage to be 64%. The use of this value resulted in a lower estimate of the proportion of epidemic cases prevented, namely 46%. Discussion The WHO threshold strategy was found to be sensitive and specific in the prediction of epidemicmeningococcal disease in seven African countries during 1997. No other threshold or epidemic definition performed as well or significantly better when sensitivity, specificity and predictive value positive were considered together. A number of thresholds were found to be reasonably sensitive and specific for predicting epidemics of various size, as has been suggested elsewhere (20, 21). It is possible that lowering the threshold from 15 to 10 per 100 000 per week in high-risk epidemic situations, e.g. for increases in incidence observed early during the meningitis season, would allow even higher sensitivity and earlier epidemic detection with minimal sacrifices in specificity and predictive value positive (21–24). However, the maximization of each of these parameters is critical. A reduction in sensitivity leads to delayed detection of epidemics and potentially reduces the number of cases and deaths averted through vaccination. Reductions in specificity and predictive value positive lead to vaccination campaigns of suboptimal impact in the absence of an ensuing, but predicted, epidemic. Such campaigns place considerable demands on human and material resources in countries where both are in very short supply. The most appropriate denominator range for the use of the threshold in districts has been suggested as 30 000 to 100 000 persons (1, 14). In the present study, however, it emerged that sensitivity and specificity were very good in districts with populations of up to 175 000 and reasonably good even in very large districts. We found that districts with over 110 000 people were more likely to cross the threshold, suggesting that the transmission ofmeningococcal diseasemay be comparatively rapid and/or sustainable in areas of high population density. This finding should be interpreted cautiously, however, since surveillance may be more complete and timely in districts with relatively large populations. By comparing actual and theoretical epidemic curves in one district experiencing a large-scale epidemic, we found that nearly three-fifths of meningococcal disease cases were prevented through the timely application of the threshold strategy. This number is close to a recently published estimate of the strategy’s potential impact (15). In the district analysed, vaccine coverage rates were higher and the vaccination campaign was conductedmore promptly than in other districts studied during the past decade (17, 25). This explains the higher estimated impact of the strategy in our study, and refutes the notion that delays in implementation of the threshold strategy are inevitable in the region (26). Although the threshold strategy was a sensitive and specific predictor of district epidemics and had a significant impact on an epidemic at the district level, it has both theoretical and practical limitations. Assuming a vaccine efficacy of at least 85% (16) and optimal levels of achievable Table 3.Numbers of districts above and belowweekly threshold of 15 cases per 100 000 for two consecutive weeks vs yearly attack rate of 100 cases/100 000 Yearly attack rate Threshold rate 5 100 < 100 Total 515 60 3 63 <15 6 105 111 66 108 174 Sensitivity: 60/(60+6) = 90.9%. Specificity: 105/(105+3) = 97.2%. Predictive value positive: 60/(60+3) = 95.2%. Predictive value negative: (105/105+3) = 97.2%. Table 4. Sensitivity and specificity of WHO threshold by district size District size (in 1000s) Sensitivity Specificity <111 (n = 40) 100 97 111–<176 (n = 46) 100 93 176–<248 (n = 47) 88 95 5 248 (n = 41) 80 93 345Bulletin of the World Health Organization 2002, 80 (5) Response to meningococcal disease epidemics vaccine coverage of 85%, as in the district studied, the threshold strategy’s maximum potential for epidemic reduc- tion is approximately 72% (85 x 0.85), starting days to weeks after vaccination, depending on the humoral immune response. In a sub-Saharan African district of average size and weekly epidemic activity, numerous cases per week might still occur even with strict adherence to the threshold strategy. Besides the theoretical limits, in Africa there are formidable practical barriers to the control of epidemic diseases. Delays in communication between peripheral and national authorities increase the time between the detection of epidemics and the performance of vaccination. The inadequacy of vaccine supplies and logistics may decrease the effectiveness of the threshold even when an epidemic is rapidly detected. The threshold strategy is difficult to implement among sparsely scattered populations that are inaccessible by road. The strategy’s effectiveness is probably diminished more by these and other practical difficulties than by the theoretical limits. Although routine immunization has been proposed as an alternative to the currently recommended WHO threshold strategy (27, 28), arguments have been made against the preventive use of meningococcal polysaccharide vaccines in developing and industrialized countries alike. Studies in sub- Saharan Africa and elsewhere have indicated that meningo- coccal polysaccharide vaccines are poorly immunogenic, perhaps especially so in areas with high rates of malnutrition, malaria and geohelminth infections (29, 30), and that they confer only evanescent protection among children aged under 4 years, among whom the risk is highest (31). They produce a limited booster effect and may attenuate serum bactericidal responses to meningococcal serogroups A and C upon revaccination (32, 33). In most settings they do not appreciably or durably alter nasopharyngeal carriage rates (34–36). Furthermore, even with antigens of longstanding use, routine EPI vaccine coverage rates rarely surpass 50–60% in sub- Saharan African countries (37). Controversy exists concerning the potential cost- effectiveness of preventive meningococcal vaccination in sub-Saharan Africa. A simulated epidemic model suggested that routine vaccination might be more cost-effective than mass vaccination strategies (38), but an analysis of data from Burkina Faso indicated that four preventive doses of meningococcal polysaccharide vaccine would prevent only 30% of meningitis cases as opposed to 42% through the WHO strategy (39). Strategies other than preventive vaccina- tion are clearly needed (40, 41). The present study had several limitations. The sensitiv- ity, specificity and predictive value positive of the threshold strategy were calculated using data collected during a large- scale epidemic; whether the strategy would be effective in the control of smaller epidemics or endemic disease is uncertain. Frequently, there is incomplete and late case-reporting from health centres, the most peripheral African health care access points, to district, regional and national health authorities. Over a quarter of the cases that occurred in six countries during 1997 were reported late, i.e. after a delay of at least three months, to national health authorities and WHO (2, 9). In order to maximize the number of districts analysed, data were obtained largely through national sources, which may not have reflected the true incidence of meningococcal disease at district level. This is evidently a problem in relation to other infectious diseases in Africa (42). In the almost complete absence of laboratory confirmation the strategy’s impact might have been underestimated if meningococcal disease cases were caused by other virulent serogroups, e.g. W-135, not included in the bivalent A+C vaccines used inmost campaigns at district level (43–45). Finally, our estimated impact of the strategy on the basis of data from a single district with high vaccine coverage and minimal delay in vaccination may not represent the strategy’s impact on other district epidemics, as discussed above. Measuring the impact of the WHO strategy in sub- Saharan Africa requires a longitudinal sample of representative districts experiencing diverse epidemic conditions and re- sponse capacities (46). Additional meningococcal disease control measures are needed in Africa. The identification of districts at high risk for epidemics (23) and the evaluation of preventive vaccination or other, e.g. treatment-based, strategies (47), are priorities for research. However, there is a paucity of data suggesting that preventive strategies significantly alter the epidemiology of meningococcal disease in actual African districts (48) and there is little evidence that improved control of epidemics can be expected through the use of alternatively defined thresh- olds or other strategies. Many investigators feel that strengthened implementation of the WHO threshold ap- proach is critical for improved meningococcal disease control in this region (49, 50). One of the most promising prospects for improving the control of meningococcal disease involves using meningo- coccal capsular polysaccharide-protein conjugate vaccines (50, 51). In clinical trials in the Gambia (52), Niger (53), and the USA (54), meningococcal conjugate vaccines produced up to 20 times more serum bactericidal activity than polysaccharide vaccines. Direct evidence and experience with Haemophilus influenzae type b conjugate vaccines suggests that immunol- ogical memory and herd immunity, mediated via the interruption of nasopharyngeal carriage, are also possible advantages (55–58). Efforts should be made to expedite the licensing of meningococcal conjugate vaccines and to reduce their cost in order to permit their earliest possible use in developing countries. n Acknowledgements The authors thank the following persons for technical assistance: Maria Santamaria, Evgeny Tikhomirov (WHO/ Geneva); Francis Varaine, Anne-Valerie Kaninda, Rosamund Lewis, Christophe Paquet (Epicentre, Paris); Felicity Cutts (London School of Hygiene and Tropical Medicine); Georges Soula (Centre Muraz, Bobo-Dioulasso, Burkina Faso); and Daniel Kertesz (WHO/Mali). Conflicts of interest: none declared. 346 Bulletin of the World Health Organization 2002, 80 (5) Research Re´sume´ De´tection et re´ponse pre´coces en cas d’e´pide´mie de me´ningite a` me´ningocoque en Afrique subsaharienne : e´valuation de la strate´gie OMS Objectif Evaluer la sensibilite´, la spe´cificite´ et la valeur pre´dictive positive de la strate´gie OMS fonde´e sur l’utilisation d’un seuil e´pide´mique pour de´tecter les e´pide´mies de me´ningite a` me´ningocoque en Afrique subsaharienne, et estimer l’impact de cette strate´gie sur une e´pide´mie au niveau du district. Me´thodes Les donne´es sur les cas de me´ningite a` me´ningocoque au niveau du district ont e´te´ recueillies une fois par semaine aupre`s des ministe`res de la sante´, des bureaux re´gionaux de l’OMS, des bureaux de l’OMS dans les pays et des organisations non gouvernementales, dans les pays ou` des e´pide´mies de me´ningite a` me´ningocoque ont se´vi en 1997. Une e´pide´mie e´tait de´finie par un taux d’atteinte cumule´ au niveau du district d’au moins 100 cas pour 100 000 habitants de janvier a` mai, pendant la pe´riode de risque e´pide´mique. La sensibilite´, la spe´cificite´ et la valeur pre´dictive positive du seuil utilise´ par l’OMS ont e´te´ calcule´es, et on a compare´ les courbes de sensibilite´ en fonction de la valeur (1-spe´cificite´) pour diffe´rentes valeurs du seuil et d’autres de´finitions de l’e´pide´mie. L’impact de la strate´gie OMS sur une e´pide´mie survenue dans un district a e´te´ estime´ par comparaison entre le nombre de cas e´pide´miques et le nombre estime´ de cas e´vite´s graˆce a` la vaccination. Re´sultats L’analyse a porte´ sur 48 198 cas rapporte´s dans 174 districts du Be´nin, du Burkina Faso, de Gambie, du Ghana, du Mali, du Niger et du Togo. Ces cas repre´sentaient 80,3 % des cas rapporte´s par l’Afrique a` l’OMS pendant la pe´riode e´pide´mique de 1997. Les districts comptaient entre 10 298 et 573 908 habitants. Le seuil e´pide´mique a e´te´ franchi pendant deux semaines conse´cutives dans 69 districts (39,7 %) et 66 districts (37,9 %) ont e´te´ touche´s par une e´pide´mie. Globalement, la sensibilite´ du seuil e´pide´mique pour la pre´vision des e´pide´mies e´tait de 97 %, sa spe´cificite´ de 95 % et sa valeur pre´dictive positive de 93 %. Prises ensemble, ces valeurs e´taient e´quivalentes ou supe´rieures aux valeurs correspondantes pour des seuils diffe´rents et d’autres de´finitions de l’e´pide´mie, et restaient e´leve´es quelle que soit la taille du district. Le nombre estime´ de cas e´pide´miques potentiels avait diminue´ de pre`s de 60 % dans le groupe d’aˆge vise´ par la vaccination dans un district ou` les directives avaient e´te´ applique´es en temps voulu. Conclusion L’utilisation de la strate´gie OMS e´tait un moyen sensible et spe´cifique de de´tection pre´coce des e´pide´mies de me´ningite a` me´ningocoque qui ont touche´ les pays d’Afrique subsaharienne en 1997 et a eu un impact important sur l’e´pide´mie dans un district. Ne´anmoins, le poids de la me´ningite a` me´ningocoque dans ces pays reste e´norme et des mesures de lutte supple´mentaires sont ne´cessaires. Resumen Deteccio´n y respuesta precoces en las epidemias de enfermedad meningoco´cica en el A´frica subsahariana: evaluacio´n de la estrategia de la OMS Objetivo Evaluar la sensibilidad, la especificidad y el valor predictivo positivo de la estrategia de umbral de la OMS para la deteccio´n de las epidemias de enfermedad meningoco´cica en el A´frica subsahariana, y estimar el impacto de la estrategia en una epidemia a nivel de distrito. Me´todos Con periodicidad semanal se reunieron datos sobre los casos de meningitis a nivel de distrito a partir de los ministerios de salud, de las oficinas regionales y de paı´s de la OMS y de organizaciones no gubernamentales en paı´ses donde habı´a epidemias de enfermedad meningoco´cica en 1997. Se considero´ que habı´a epidemia cuando la tasa acumulativa de ataques por distrito alcanzaba al menos los 100 casos por 100 000 habitantes entre enero y mayo, el periodo de riesgo de epidemia. Se calcularon la sensibilidad, la especificidad y el valor predictivo positivo de la tasa umbral de la OMS y se compararon las curvas de sensibilidad frente al valor (1-especificidad) para diferentes umbrales y dimensiones de la epidemia. El impacto de la estrategia de la OMS en una epidemia de a´mbito distrital se calculo´ comparando las cifras de los casos epide´micos y los casos que segu´n las estimaciones se habı´an prevenido mediante la vacunacio´n. Resultados Se analizaron 48 198 casos notificados en 174 dis- tritos de Benin, Burkina Faso, Gambia, Ghana, Malı´, Nı´ger y Togo. Estos casos representaban el 80,3% de los notificados por A´frica a la OMS durante el periodo epide´mico de 1997. La poblacio´n de los distritos iba de 10 298 a 573 908 personas. La tasa umbral fue superada durante dos semanas consecutivas en 69 distritos (39,7%), y hubo epidemias en 66 distritos (37,9%). A nivel global, la sensibilidad de la tasa umbral para predecir las epidemias fue del 97%, la especificidad del 95%, y el valor predictivo positivo del 93%. Considerados conjuntamente, estos valores alcanzaban o superaban la sensibilidad, la especificidad y el valor predictivo positivo observados al definir de otra forma las tasas umbral y las epidemias, y seguı´an siendo altos con independencia del taman˜o del distrito. El nu´mero estimado de casos epide´micos potenciales disminuyo´ casi en un 60% en el grupo de edad elegido para vacunacio´n en un distrito donde se siguieron puntualmente las directrices. Conclusio´n La estrategia de la OMS funciono´ como una alternativa sensible y especı´fica para detectar precozmente las epidemias de la enfermedad meningoco´cica en los paı´ses del A´frica subsahariana durante 1997 y tuvo una repercusio´n sustancial en la epidemia de un distrito. No obstante, la carga de enfermedad meningoco´cica sigue siendo enorme en esos paı´ses, lo que hace necesarias medidas de control adicionales. 347Bulletin of the World Health Organization 2002, 80 (5) Response to meningococcal disease epidemics References 1. World Health Organization. Control of epidemic meningococcal disease: WHO practical guidelines. Geneva: World Health Organization; 1998. p. 1-82. Unpublished document WHO/EMC/BAC/98-3. 2. World Health Organization. Response to epidemic meningitis in Africa, 1997. Weekly Epidemiological Record 1997;72:313-8. 3. Cartwright K, editor. Meningococcal disease. Chichester: J. Wiley and Sons; 1995. 4. Riedo FX, Plikaytis BD, Broome CV. Epidemiology and prevention of meningococcal disease. Pediatric Infectious Diseases Journal 1995;14:643-57. 5. Schuchat A, Robinson K, Wenger JD, Harrison LH, Farley M, Reingold AL, et al. Bacterial meningitis in the United States in 1995. Active Surveillance Team. New England Journal of Medicine 1997;337:970-6 6. Lapeyssonnie L. La me´ningite ce´re´bro-spinale en Afrique.[Cerebrospinal meningitis in Africa.] Bulletin of the World Health Organization 1963; 28(Suppl.):1-103. 7. Greenwood BM, Blakebrough IS, Bradley AK, Waali S, Whittle HC. Meningococcal disease and season in sub-Saharan Africa. Lancet 1984;1:1339-42. 8. Wenger JD, Perkins BA. Patterns in the emergence of meningococcal disease. In: Scheld WM, Armstrong D, Hughes JB, editors. Emerging infections 1. Washington (DC): ASM Press;1998. p.125-36. 9. World Health Organization. WHO report on global surveillance of epidemic- prone diseases. Available from: URL: http://www.who.int/emc-documents/ surveillance/docs/whocdscsrisr2001.html/Meningitis/Meningitis.htm 10. Greenwood BM, Bradley AK, Smith AW, Wall RA . Mortality from meningococcal disease during an epidemic in the Gambia, West Africa. Transactions of the Royal Society of Tropical Medicine and Hygiene 1987;81:36-8. 11. Moore PS, Reeves MW, Schwartz B, Gellin BG, Broome CV. Intercontinental spread of an epidemic group A Neisseria meningitidis strain. Lancet 1989;2:260-3. 12. Riou JY, Djibo S, Sangare L, Lombart JP, Fagot P, Chippaux JP, et al. A predictable comeback: the second pandemic of infections caused by Neisseria meningitidis serogroup A subgroup III in Africa, 1995. Bulletin of the World Health Organization 1996;74:181-7. 13. Koumare´ B, Bougoudogo F, Cisse M, Doumbia T, Keita MM. Bacteriological aspects of purulent meningitis in Bamako district. A propos of 1541 bacterial strains collected from 1979 to 1991. Bulletin de la Socie´te´ de Pathologie Exotique 1993;86:136-40. 14. Moore PS, Plikaytis BD, Bolan GA, Oxtoby MJ, Yacta A, Zoubga A, et al. Detection of meningitis epidemics in Africa: a population-based analysis. International Journal of Epidemiology 1992;21:155-62. 15. Woods CW, Armstrong G, Sackey SO, Tetteh C, Bugri S, Perkins BA, et al. Emergency vaccination against epidemic meningitis in Ghana: implications for the control of meningococcal disease in West Africa. Lancet 2000;355:30-3. 16. Goldschneider I, Gotschlich EC, Artenstein MS. Human immunity to the meningococcus. I. The role of humoral antibody. Journal of Experimental Medicine 1969;129:1307-26. 17. Pinner RW, Onyango F, Perkins BA, Mirza NB, Ngacha DM, Reeves M, et al. Epidemic meningococcal disease in Nairobi, Kenya, 1989. Journal of Infectious Diseases 1992;166:359-64. 18. Zangwill KM, Stout RW, Carlone GM, Pais L, Harekeh H, Mitchell S, et al. Duration of antibody response after meningococcal polysaccharide vaccination in US Air Force personnel. Journal of Infectious Diseases 1994;169:847-52. 19. Leake JAD. CDC Trip report: meningococcal vaccine efficacy, Burkina Faso, April-May 1997. Unpublished document. 20. de Chabalier F, Hassane A, Chippaux JP. Evaluation of surveillance thresholds for prediction of meningitis epidemics using ongoing surveillance data at the district level, in Niger. Transactions of the Royal Society of Tropical Medicine and Hygiene 2000;94:251-2. 21. Lewis R, Nathan N, Diarra L, Belanger F, Paquet C. Timely detection of meningococcal meningitis epidemics in Africa. Lancet 2001;358:287-93. 22. World Health Organization. Detecting meningococcal meningitis epidemics in highly endemic African countries. Weekly Epidemiological Record 2000;38:306-9. 23. de Chabalier F, Djingarey MH, Hassane A, Chippaux JP. Meningitis seasonal pattern in Africa and detection of epidemics: a retrospective study in Niger, 1990-98. Transactions of the Royal Society of Tropical Medicine and Hygiene 2000;94:664-8. 24. Kaninda AV, Belanger F, Lewis R, Batchassi E, Aplogan A, Yakoua Y, et al. Effectiveness of incidence thresholds for detection and control of meningococcal meningitis epidemics in northern Togo. International Journal of Epidemiology 2000;29:933-40. 25. Varaine F, Caugant DA, Riou JY, Konde MK, Soga G, Nshimirimana D, et al. Meningitis outbreaks and vaccination strategy. Transactions of the Royal Society of Tropical Medicine and Hygiene 1997;91:3-7. 26. McDamien D, Boelaert M, Van Damme W, Van der Stuyft P. Meningococcal immunisation in Ghana. Lancet 2000;355:2252-3. 27. Robbins JB, Towne DW, Gotschlich EC, Schneerson R. Love’s labours lost: failure to implement mass vaccination against group A meningococcal meningitis in sub-Saharan Africa. Lancet 1997;350:880-2. 28. Robbins JB, Schneerson R, Gotschlich EC. A rebuttal: epidemic and endemic meningococcal meningitis in sub-Saharan Africa can be prevented now by routine immunization with group A meningococcal capsular polysaccharide vaccine. Pediatric Infectious Diseases Journal 2000;19:945-53. 29. Greenwood BM, Bradley AK, Blakebrough IS, Whittle HC, Marshall TF, Gilles HM. The immune response to a meningococcal polysaccharide vaccine in an African village. Transactions of the Royal Society of Tropical Medicine and Hygiene 1980;74:340-6. 30. Williamson WA, Greenwood BM. Impairment of the immune response to vaccination after acute malaria. Lancet 1978;1:1328-9. 31. Reingold AL, Broome CV, Hightower AW, Ajello GW, Bolan GA, Adamsbaum C, et al. Age-specific differences in duration of clinical protection after vaccination with meningococcal polysaccharide A vaccine. Lancet 1985;2:114-8. 32. Ceesay SJ, Allen SJ, Menon A, Todd JE, Cham K, Carlone GM, et al. Decline in meningococcal antibody levels in African children 5 years after vaccination and the lack of an effect of booster immunization. Journal of Infectious Diseases 1993;167:1212-5. 33. MacLennan J, Obaro S, Deeks J, Williams D, Pais L, Carlone G, et al. Immune response to revaccination with meningococcal A and C polysaccharides in Gambian children following repeated immunisation during early childhood. Vaccine 1999;17:3086-93. 34. Hassan-King MK, Wall RA, Greenwood BM. Meningococcal carriage, meningococcal disease and vaccination. Journal of Infection 1988;16:55-9. 35. Blakebrough IS, Greenwood BM, Whittle HC, Bradley AK, Gilles HM. Failure of meningococcal vaccination to stop the transmission of meningococci in Nigerian schoolboys. Annals of Tropical Medicine and Parasitology 1983;77:175-8. 36. Greenwood BM, Hassan-King M, Whittle HC. Prevention of secondary cases of meningococcal disease in household contacts by vaccination. British Medical Journal 1978;1:1317-9. 37. World Health Organization. EPI information system, global summary. Geneva: World Health Organization; 1998. Unpublished document WHO/EPI/GEN/ 98.10. 38. Bovier PA, Wyss K, Au HJ. A cost-effectiveness analysis of vaccination strategies against N. meningitidis meningitis in sub-Saharan African countries. Social Science and Medicine 1999;48:1205-20. 39. Miller MA, Wenger J, Rosenstein N, Perkins B. Evaluation of meningococcal meningitis vaccination strategies for the meningitis belt in Africa. Pediatric Infectious Diseases Journal 1999;18:1051-9. 40. Perkins BA, Broome CV, Rosenstein NE, Schuchat A, Reingold AL. Meningococcal vaccine in sub-Saharan Africa. Lancet 1997;350:1708. 41. Wenger J, Tikhomirov E, Barakamfitiye D, Bele O, Heymann DL. Meningococcal vaccine in sub-Saharan Africa. Lancet 1997;350:1709-10. 42. Centers for Disease Control and Prevention. Progress toward global eradication of poliomyelitis, 1996. Morbidity and Mortality Weekly Report 1997;46: 579-84. 43. Kwara A, Adegbola RA, Corrah PT, Weber M, Achtman M, Morelli G, et al. Meningitis caused by a serogroup W135 clone of the ET-37 complex of Neisseria meningitidis in West Africa. Tropical Medicine and International Health 1998;3:742-6. 44. Centers for Disease Control and Prevention. Serogroup W-135 meningococcal disease among travelers returning from Saudi Arabia—United States, 2000. Morbidity and Mortality Weekly Report 2000;49:345-6. 45. Popovic T, Saccchi CT, Reeves MW, Whitney AM, Mayer LW, Noble CA, et al. Neisseria meningitidis Serogroup W-135 isolates associated with the ET-37 complex. Emerging Infectious Diseases 2000;6:428-9. 348 Bulletin of the World Health Organization 2002, 80 (5) Research 46. Greenwood BM, Greenwood AM, Bradley AK, Williams K, Hassan-King M, Shenton FC, et al. Factors influencing susceptibility to meningococcal disease during an epidemic in the Gambia, West Africa. Journal of Infection 1987;14:167-84. 47. Veeken H, Ritmeijer K, Hausman B. Priority during a meningitis epidemic: vaccination or treatment? Bulletin of the World Health Organization 1998;76:135-41. 48. Hassan J, Massougbodji A, Chippaux JP, Massit B, Josse R. Meningococcal immunisation and protection from epidemics. Lancet 1998;352:407-8. 49. Kaninda AV, Varaine F, Henkens M, Paquet C. Meningococcal vaccine in sub-Saharan Africa. Lancet 1997;350:1708. 50. Greenwood B. Manson lecture. Meningococcal meningitis in Africa. Transactions of the Royal Society of Tropical Medicine and Hygiene 1999;93:341-53. 51. Perkins BA. New opportunities for prevention of meningococcal disease. JAMA 2000;283:2842-3. 52. Twumasi PA, Kumah S, Leach A, O’Dempsey TJ, Ceesay SJ, Todd J, et al. A trial of group A plus group C meningococcal polysaccharide-protein conjugate vaccine in African infants. Journal of Infectious Diseases 1995;171:632-8. 53. Campagne G, Garba A, Fabre P, Schuchat A, Ryall R, Boulanger D, et al. Safety and immunogenicity of three doses of a Neisseria meningitidis A + C diphtheria conjugate vaccine in infants from Niger. Pediatric Infectious Diseases Journal 2000;19:144-50. 54. Lieberman JM, Chiu SS, Wong VK, Partidge S, Chang SJ, Chiu CY, et al. Safety and immunogenicity of a serogroups A/C Neisseria meningitidis oligosaccharide- protein conjugate vaccine in young children: a randomized controlled trial. Journal of the American Medical Association 1996;275:1499-1503. 55. Leach A, Twumasi PA, Kumah S, Banya WS, Jaffar S, Forrest BD, et al. Induction of immunologic memory in Gambian children by vaccination in infancy with a group A plus group C meningococcal polysaccharide-protein conjugate vaccine. Journal of Infectious Diseases 1997;175:200-4. 56. Centers for Disease Control and Prevention. Progress toward elimination of Haemophilus influenzae type b disease among infants and children–United States, 1987-1995. Morbidity and Mortality Weekly Report 1996;45:901-6. 57. Takala AK, Eskola J, Leinonen M, Kayhty H, Nissenen A, Pekkanen E, et al. Reduction of oropharyngeal carriage of Haemophilus influenzae type b (Hib) in children immunized with an Hib conjugate vaccine. Journal of Infectious Diseases 1991;164:982-6. 58. Mohle-Boetani JC, Ajello G, Breneman E, Deaver KA, Harvey C, Plikaytis BD, et al. Carriage of Haemophilus influenzae type b in children after widespread vaccination with conjugate Haemophilus influenzae type b vaccines. Pediatric Infectious Diseases Journal 1993;12:589-93. 349Bulletin of the World Health Organization 2002, 80 (5) Response to meningococcal disease epidemics