Synergistic impact of measles and diarrhoea on nutrition and mortality in Bangladesh.

Bulletin of the World Health Organization, 59 (6): 901-908 (1981) Synergistic impact of measles and diarrhoea on nutrition and mortality in Bangladesh* F. T. KOSTER,1 G. C. CURLIN,2 K. M. A. Aziz,3 & AzIZUL HAQUE4 The efficacy ofrecent efforts to improve the delivery ofavailable vaccines, especially in the developing countries, can be aided by surveys that identify populations sufferingfrom unduly high morbidity and mortality as a result ofaparticular disease. Thispaper gives the results ofaprospective household surveillance of5775 children in 12 villages in Bangladesh, in which the occurrence ofmeasles and diarrhoea, and changes in nutritional status, and the causes ofall deaths during a 12-monthperiod were monitored. The measles case-fatality rate was 3.7% among all children. Prolonged diarrhoea was the most common complication in fatal cases. The greatest weight loss was seen in children with measles complicated by pro- longed diarrhoea, and children under 4 years ofage in this group failed to achieve "catch- up" growth. Thus measles and diarrhoea appeared to interact synergistically to increase mortality and the irreversible effects of nutritional deprivation. This information should providefurther impetus to an extensive measles vaccinationprogramme in Bangladesh. Until then, nutrition intervention programmes should be aimed at children under 4 years ofage, convalescingfrom measles. In developing countries measles continues to be associated with much higher mortality than in indus- trialized countries (1, 2). The reasons for this vary from region to region, but protein-energy-malnutri- tion (especially kwashiorkor (3, 4)) and secondary bacterial and viral infections (1, 5) are probably the most important factors. Pneumonia (1, 5) is a well- recognized, often fatal, secondary infection. Diar- rhoea and dysentery frequently accompany measles at or before the appearance of the rash (2, 6), but their role in the exacerbation of malnutrition and mortality is not well documented. To explore the role of diarrhoea in contributing to measles associated morbidity and mortality, we conducted a prospective household surveillance of 5775 children from 1 month to 10 years of age, in 12 villages for a period of 12 months and monitored the onset and duration of any diarrhoeal episodes, * This research was supported by NIH Research Grant No. 5R07 AI 10048-15 and by the US Agency for International Development. Reprint requests should be addressed to Dr F. T. Koster, Division of Infectious Diseases, Department of Medicine, The University of New Mexico School of Medicine, Alburquerque, NM 87131, USA. I Fellow, The Johns Hopkins University International Center for Medical Research, Dacca, Bangladesh. 2 Director of Epidemiology, International Center for Diarrhoeal Disease Research, Bangladesh (ICDDR,B) Dacca, Bangladesh. 3Director, Field Studies, ICDDR,B, Dacca, Bangladesh. 4Sanitary Engineer, ICDDR,B, Dacca, Bangladesh. measles, changes in nutritional status, and the cause of all deaths. Measles was the single most important cause of death during that period and diarrhoea or dysentery was the most common complication of these fatal measles cases. In addition, measles complicated by prolonged diarrhoea was found to be associated with a large fall in body weight and a subsequent failure to achieve "catch-up" growth, especially in children under four years of age. METHODS The study was conducted in Matlab thana, a riverine province in southeast Bangladesh, from August 1975 to July 1976. All the 5775 children in the twelve villages, who had been under continuous demographic surveillance since 1966 as part of a field study for vaccine evaluation, were included in this prospective survey of diarrhoeal disease morbidity and nutritional status. Each family (after giving free and written consent) was visited once a week by a field worker who asked a standard series of questions on the frequency and consistency of daily bowel move- ments; for children under four years of age, the mothers were questioned. Diarrhoea was defined as three or more daily bowel movements consisting of watery stools or the presence of blood or mucus in the 4132 901- F. T. KOSTER ET AL. stool, and was recorded on a calendar on each day of occurrence. Dysentery was not distinguished from diarrhoea. The length of each episode was defined as the number of consecutive days of diarrhoea with not more than one "non-diarrhoea" day interposed. Pro- longed diarrhoea was arbitrarily defined as that persisting for more than seven days. At 60-day intervals, the nutritional status of each child was assessed by height and weight measure- ments. These data were reported either as a percentage of the Harvard weight-for-height standard (7), or as the change in body weight over a two-month interval. Although the exact age of each child was known and weight-for-age percentages were calculated, different degrees of nutritional stunting in children over two years old rendered weight-for-age insensitive to acute changes in nutrition, and these data are not reported. The diagnosis of measles was based largely on recognition of the typical rash. The "atypical" des- quamating violaceous rash was not uncommon and was recognized as measles by both villagers and field workers. Serological support for the presumptive diagnosis of measles was obtained during an epidemic in village M, 160 children with a rash observed by the field worker within the last four months and 110 age- matched controls being bled using the fingerstick tech- nique two weeks after the last case of measles had been reported in the village. Each 0. I-ml blood sample was diluted 1:10 with physiological saline and stored at -50°C until assayed by the haemagglutination inhibition (HI) method (8). When the independent demographic surveillance team reported the death of a child, a "special survey" field worker interviewed the parents to identify specific symptoms prior to death. A standard ques- tionnaire covered the following symptoms: fever, cough, tachypnoea, diarrhoea, dysentery, anorexia, vomiting, anasarca, seizures, and events such as accidents or trauma resulting in death. No reliable retrospective distinction could be made between diar- rhoea and "dysentery" and these were combined as "diarrhoea". The data from this questionnaire agreed with the data from the weekly diarrhoea surveillance in all 11 fatal measles cases in villages M and L, for which all surveillance data were available at the time of analysis. A "measles-associated" death or complication was defined as an event occurring within 30 days after the onset of the rash. Data related to mortality were obtained from the questionnaire completed after each death. Otherwise, data on nutritional consequences of measles and diarrhoea were obtained from diarrhoea surveillance. Statistical significance was tested by the non-para- metric Wilcoxon rank sum test when comparing small populations (n< 50) and the chi-square test or the Student t test when comparing larger populations. RESULTS Epidemiology Measles surveillance in 12 villages, from August 1975 to July 1976, found cases in all the villages and two of these villages (M and L) experienced major epidemics. The incidence reached a peak in March (Fig. 1) and coincided with the period of lowest rain- fall and lowest relative humidity. While in any one village measles cases clustered during a three-month period, in the surveillance area measles cases appeared throughout the entire year and decreased to low inci- dence during the three monsoon months, when the rainfall was greatest and humidity highest. A total of 923 cases of measles were recorded during the 12-month period of surveillance. Of these, 896 (97%) occurred in children between the ages of 1 and 120 months (Tablel). Twenty-seven cases occurred in adolescents or adults. Seven cases were in infants 1-6 months of age, all of whom survived. Although the infant cases were not confirmed serologically, since all of these cases occurred during local measles epidemics (usually in the same household) they were presumed to be cases of measles and were included in the calcu- lations of morbidity and mortality. In village M among the 160 children with a presumptive history of measles, only 4 had convalescent HI titres of less than 1:20- of whom 2 were 9 months of age. Thus, 98% of the children reported as having a measles rash showed evidence of prior measles. Among the 110 controls, 30 had titres of less than 1: 20. Mortality Deaths were defined as measles-associated if the ill- ness leading to death began within 30 days of the MEAN PERCENT HUMIDITY 200 _10 140 120 100 o; s .J co 60 20 87 84 ^. 78 82 90 76 76 71 79 as 87 II':;lsi#;2 s H! I~~~~~~~~~~~~~~~~~~~~~U AUG SEP OCT NOV DEC JAN FEB MAR APR MY JUN JUL MONTHS >: TOTAL 12 VILLAGES MEHERAN CM LUDHUA Fig. 1. Monthly incidence of measles in 12 villages between August 1975 and July 1976, compared with monthly 10-year means for relative humidity (%) and rainfall (cm). 902 MEASLES AND DIARRHOEA IN BANGLADESH Table 1. Incidence of measles and mortality rates in the children of 12 Matlab villages, by age group, August 1975-July 1976 1-23 24-47 48-71 72-120 months months months months Total Population (n) 1421 1369 1148 1837 5775 Measles: No. of cases 203 307 193 193 896 Attack rate per 1000 143 224 168 105 155.2 No. of deaths 9 13 8 3 33 Case-fatality rate per 1000 44.3 42.4 41.5 15.5 36.8 Mortality (per 1000) All causes 18.3 27.0 23.5 3.3 21.1 Measles-associateda 5.6 9.5 7.0 1.6 5.7 Gastrointestinalb 2.1 8.0 7.8 0.5 5.6 Pneumonia 1.4 2.9 1.7 0.5 1.6 Other 9.0 6.6 7.0 0.7 8.2 * Deaths occurring within one month of rash onset. b All diarrhoea and dysentery, acute and chronic, not following measles. appearance of the rash. The case-fatality rate for measles-associated deaths was approximately the same among all age groups under 6 years of age (Table 1). The case-fatality rate among the 6-10-year-olds was one-third of that of the younger children, but measles accounted for a higher percentage of all deaths among children aged 6 years or more. Measles- associated deaths were distributed equally among males and females and measles was the single most important cause of death in these 12 villages, causing 27% of the deaths from all causes. More detailed information about the events and symptoms preceding all the 122 childhood deaths reported during this period was obtained by family interviews using the standard questionnaire (Table 2). Three groups of measles-associated deaths could be defined: those in whom diarrhoea was the pre- dominant complication prior to death; those in whom respiratory symptoms of cough and tachypnoea were the predominant complications prior to death; and those who had a combination of diarrhoea and respir- atory symptoms, or had seizures, anasarca, anorexia, or vomiting prior to death (labelled "other"). Among the measles deaths complicated by diarrhoea, the mean length of the diarrhoeal episode immediately prior to death was 13 days, compared with 25 days for lethal diarrhoea episodes not associated with measles. In this Bangladeshi population, mortality from measles-associated illness was not related to nutritional status. This conclusion was made from a case-control study comparing the weight-for-height percentage of age-, sex- and neighbourhood-matched controls to the nutrition of the 33 children who died of measles-associated illnesses (Table 2). The pre-morbid nutrition of all these children was comparable to that of their age-matched controls (P>0.5, rank sum test), and this comparison was valid for the groups of chil- dren dying of measles complicated by diarrhoea. In contrast, children dying of diarrhoea alone (no measles during the study period) had significantly poorer pre-morbid nutrition than those children dying of measles-associated illnesses (P< 0.01) or other causes (P<0.01). Temporal measles-diarrhoea relationships Measles substantially increased the number of epi- sodes of diarrhoea. In village M, during the months from October to January, 123 children without measles had a monthly diarrhoea incidence rate of 0.14, 0.15, 0.16, and 0.06 episodes per child, respec- tively. In the same village, 175 children experiencing the onset of measles during the same four-month period had monthly diarrhoea incidence rates of 0.23, 0.23, 0.43, and 0.19, respectively. If this group of 175 had not experienced measles but instead had had diar- rhoea rates comparable with those of their age- matched controls (given above) we calculate that 83 of 261 diarrhoeal episodes, or 32%, would not have occurred during this four-month period. The definition of a measles-associated complication as that occurring within 30 days after the onset of the rash, was derived from the analysis of 119 village M children who had both measles and diarrhoea during a 903 F. T. KOSTER ET AL. Table 2. Nutritional status of 122 children who died in 12 Matlab villages, August 1975-July 1976. Percentage of standard Percentage of Days between Days of weight-for- standard rash onset diarrhoea Cause of death No. Median age height' weight-for-age and death prior to death (months) (mean ± SE) (mean + SE) (mean + SE) (mean ± SE) Measles deaths 33 37 85.7 ± 3.0 62.6 ± 2.1 14.5 ± 1.7 7.9 ± 1.3 with diarrhoea 15 36 82.1 ± 3.4 60.5 ± 2.3 17.5 ± 3.1 12.9 ± 1.8 with pneumonia 10 37 91.8 ± 3.0b 66.0 ± 2.7 9.7 ± 1.3 2.3 ± 1.0 with otherc 8 43 84.9 ± 3.3 62.7 ± 3.4 13.9 + 1.9 6.2 ± 2.3 Non-measles deaths 89 38 82.5 ± 2.8 59.9 ± 4.9 - - diarrhoea alone 29 38 73.5 ± 4.0d 52.0 ± 6.4 - 24.7 ± 3.9 Uving controlse 33 38 88.2 ± 0.9 * Weight and height recorded prior to onset of illness leading to death. b Nutritional status of children dying of measles with pneumonia is significantly better than that of children dying of measles with diarrhoea or other, P = < 0.05, Wilcoxon rank sum test. c Deaths associated with a combination of diarrhoea and respiratory symptoms or with seizures, anasarca, oral ulceration, and anorexia. d Nutritional status of children dying of diarrhoea alone is significantly lower than that of controls or all measles-associated deaths, P < 0.01, Wilcoxon rank sum test. * Survivors without history of measles during 12-month study period, age- and sex-matched to measles-associated deaths. 6-month period in the winter. Fig. 2 displays the temporal occurrence of all 149 diarrhoea episodes in relation to the onset of the measles rash. Although the diarrhoeal episodes were distributed throughout the 6-month period, analysis of individual case data shows that there was a marked increase in diarrhoeal episodes during the 5-week period beginning 1 week before and continuing for 4 weeks after the onset of the rash. In addition, 51%o of the episodes that occurred during the 5-week risk period lasted at least 7 days, whereas only 25%o of the episodes outside the risk period lasted that long (P< 0.005, chi-squared test). ul X : X 2 0 Z1 1 O 21 mU = DIARRHEAL EPISODE 7 DAYS DURATION M DIARRHEAL EPISO7 E .7 DAYS DURATION - 7 6 -5 4 3 2 ; j BEFORE RASH RASH I I AFTER RASH Mortality after measles andprolonged diarrhoea To assess the impact of prolonged diarrhoea (more than 7 days) on mortality following measles, the rela- tive risk of death (%) from a diarrhoeal or measles epi- sode was calculated by dividing the number of deaths by the number of episodes of each type recorded in two large villages which contained 1888 children (Table 3). The calculation gave a value of 0.97% for prolonged diarrhoea episodes not associated with measles (line 2), and 2.01% for measles cases not complicated by prolonged diarrhoea (line 3 plus line 4). However, 11.94% of measles episodes complicated by prolonged diarrhoea (line 5) resulted in death, a rate significantly greater (P< 0.0005) than the rate due to measles alone. This four-fold increase in mortality, compared with the combined rates of both measles (without prolonged diarrhoea) and prolonged diar- rhoea without measles, indicates that measles and diarrhoea combine synergistically to increase mortality. The other ten villages experienced slightly lower mortality rates during the same period. Among 3887 children, 15 died from prolonged diarrhoea (3.86 per 1000) and 19 died from measles complicated by brief (10 cases) or prolonged (9 cases) diarrhoea (4.89/ 1000), but the case fatality rate for measles and complications in the two "epidemic" villages (38.36/ 1000) was comparable with that in the other ten non- epidemic villages (35.78/1000). WEEK OF ONSET OF DIARRHEAL EPISODE Fig. 2. Onset of 149 diarrhoeal episodes in 119 measles cases in relation to the onset of measles rash in village M. 1. *1 .2 .3 .4 .5 .6 -7 904 6 11 6 = MEASLES AND DIARRHOEA IN BANGLADESH Table 3. Mortality rates due to measles and diarrhoea, and case-fatality rates in 1888 children aged 1-120 months, in 2 Matlab villages during a 12-month period of surveillance Cause-specific Type of episode No. of episodes No. of deaths mortality Case-fatalities per 1000 children per 1000 episodes Diarrhoea, < 7 days 2786 3 1.59 1.08 Diarrhoea, > 7 days' duration 822 8 4.24 9.73 Measles, no diarrhoea 99 4 2.11 40.40 Measles plus diarrhoea < 7 days' duration 199 2 1.06 10.05 Measles plus diarrhoea > 7 days' duration 67 8 4.24 119.40a All measles 365 14 7.41 38.36 a The risk of death following measles plus prolonged diarrhoea was significantly greater than that following measles plus brief or no diarrhoea, P < 0.0005 (chi-square test). Impactiof measles and secondary diarrhoea on nutrition Measles had an impact on both mortality and the nutritional status of surviving children. In the prospective nutrition study (Fig. 3), the persistent nutritional deficit after measles with or without prolonged diarrhoea, was assessed by comparing the change in the percentage of the standard weight-for- height ratio attained between the October 1975 and the August 1976 measurements for 403 children from villages M and L. Children 7-23 months of age who had measles complicated by prolonged diarrhoea sus- tained a significant, persistent deficit of approx- imately 10% (P<0.01). Children 23-47 months of age who had measles complicated by prolonged diar- H I S2 H I r L) z L.J U~ 0r 7-23mo. 86( 90 84 F88 86 80 (22) 8 78~ 76- 8Q O9T AUG1975 1976 24-47mo, (64) (68) - 27) OT AUG 1975 1976 1975 1976 Fig. 3. Change in weight-for-height (percentage of stan- dard) between October 1975 (pre-measles) and August 1976 among three groups: measles complicated by diarrhoea last- ing less than 7 days ([]); measles complicated by diarrhoea lasting 7 days or more (A); and controls (o). The number of subjects in each group is shown in parentheses on the diagram. rhoea failed to improve their nutritional status com- pared with the controls and with children who had measles and a brief attack of diarrhoea (P = 0.07) - although this difference was not as striking as that found in children less than 2 years old. The combination of measles and a prolonged period of diarrhoea in children 4 years of age or older did not cause a persistent weight-for-height deficit. The added contribution of the diarrhoea compo- nent to weight loss after measles was examined in village-M children with measles in the October- December period (Fig. 4). Initially, children with measles and diarrhoea of at least 7 days duration lost AGE 24-47 MONTHS 4 AGE 48-120 MONTHS 1.0 0 "i -0.5~~~~~~~~~~~~~ OCT DEC FEB APR JUN AUG ocT DEC FEB APR JUN AUG MONTHS Fig. 4. Changes in the mean body weight (± SE) in children of village M who had measles between October and 15 December (. measles + > 7-day diarrhoea; measles +0-6-day diarrhoea; ------ controls). Minimum of 22 children per point. There were insufficient data for the children aged 7 - 23 months. 905 F. T. KOSTER ET AL. twice as much weight as the other measles cases. All of the accelerated "catch-up" growth occurred in the December-February interval and thereafter growth rates with or without prior prolonged diarrhoea were similar. DISCUSSION This study of measles in Bangladesh confirms many earlier observations on measles in the tropics. Incidence reached its peak in the winter season, which has the lowest humidity and rainfall. This is consistent with the known persistence of infective aerosolized droplets in air of low relative humidity (9).The presumptive diagnosis of measles on the basis of the rash alone was supported by positive serology in 98% of children reported to have a rash. A similar agreement between serological data and diagnosis based on the appearance of rash has been reported in other surveys (10). The mean age of measles infection in the present study, 36 months, is similar to that reported in India (30 months) (10) but older than the mean age found in Nigeria (17 months) (4). Each village appears to experience a major epidemic every 3-4 years. In this study only 4 of the 12 villages surveyed prospectively experienced attack rates greater than 30% for children under 6 years of age. The high case-fatality rate (3.7%) observed in this study is comparable with the high rates found in other developing countries, such as Guatemala (6.6%) (6), India (2.2%) (10), and Nigeria (5%) (4), and is in contrast to rates observed in developed countries, such as the USA (0.02%) (2). These high case-fatality rates are generally ascribed to a high prevalence of pre-existing malnutrition, and/or a high incidence of secondary infections. For example, in India complications were three times more common among malnourished than well-nourished children (5). Case-fatality rates of measles among infants with kwashiorkor exceed 10% in West Africa (1, 3). This prospective study demonstrated there was no significant difference in the pre-measles nutrition of those who died and the age-matched population that survived. Also, we confirmed that diarrhoeal mortality is higher among children with poorer nutrition (11). Our observation that the children who died from measles associated with prolonged diarrhoea were significantly better nourished than children who died from diarrhoea alone may in part be explained by seasonal variations in nutritional status (K. Brown & R. Black, unpublished data, 1979). The incidence of measles was highest between December and March when nutrition was improving, while diarrhoeal deaths were more frequent between August and December when the mean weight-for-height percent- age was up to six points lower. However, seasonal variations in weight-for-height measurements could account for only a portion of the 12-point difference in the pre-morbid levels of nutrition observed between children who died from diarrhoea alone and those who died from measles (Table 2). The high mortality and incidence of complications among the better nourished measles cases appear to have been associated with a high incidence of secondary infections, to which the measles patients are more susceptible as a result of the lowering of resistance. In previous studies, bronchopneumonia was found to be the most common complication leading to death following measles (2, 4, 5), and was caused by either a prolonged measles giant-cell pneumonitis (12) or bacterial or viral superinfections. Ten of the 33 measles deaths in this study were preceded by symptoms of severe pneumonia alone. Paradox- ically, these children had better nutrition than those children dying of measles and other complications, but the reason for this finding is not known. In this study diarrhoea was the most common complication contributing to death (15 cases). In village M, 63% of all measles cases had diarrhoea during the five-week period ending four weeks after the appearance of the rash. This five-week period was denoted as the "diarrhoea-risk period". In Nigeria (4) diarrhoeal epidsodes were frequent for at least two weeks after the onset of the rash. In Guatemala (6) diarrhoea was documented only in the week preceding and the week following the appear- ance of the rash. The high frequency of diarrhoea in measles patients has been noted elsewhere (3, 4, 6, 10, 13). Diarrhoea, alone and associated with measles, was clearly responsible for the largest share of the mortality recorded among children in our study. Our observation, based on nutritional surveys every two months, that malnutrition was significantly worse prior to the onset of diarrhoea in children dying of diarrhoea alone, confirms data collected elsewhere (11, 20). Owing to the remoteness of the villages and lack of medical training of the field workers, treat- ment was provided only by indigenous practitioners. Clearly, mortality rates would have been much lower if appropriate antibiotic treatment of dysentery and oral glucose-electrolyte therapy for dehydration had been consistently available. The etiology of diarrhoea associated with measles infections is unknown. As in previous studies (6, 13) bloody diarrhoea was frequently observed in this study. Shigella spp. were often isolated from faecal samples (approximately 50%) taken during diar- rhoeal episodes in the five-week diarrhoea-risk 906A MEASLES AND DIARRHOEA IN BANGLADESH period, but controlled studies to assess the signifi- cance of Shigella superinfection were not performed. In addition to shigellosis, other common intestinal pathogens, such as enterotoxigenic Escherichia coli or rotaviruses (14), may be the cause of diarrhoea. The measles virus itself infects the intestinal epi- thelium (15), is associated with substantial protein losses from the gut (16), and may persist in mal- nourished cases (12). Thus, intestinal measles may be a viral "agent" of diarrhoea, or it may account in part for the prolonged period of susceptibility to secondary diarrhoeal pathogens. In addition, measles infections are known to cause a severe depression of the systemic immune response (12, 17), which persists for at least six weeks (18), and thus further contributes to the increased susceptibility to inter- current infection experienced by the victims. The observed acute and severe impact of measles on growth confirms earlier observations that weight losses greater than 5% of the pre-morbid body weight occur frequently, (4, 6) and that diarrhoea delays "catch-up" growth (4). In this study the failure to achieve "catch-up" growth appeared to be related to the magnitude of the initial weight loss. The persistent nutritional deficit was greatest in chil- dren under two years of age in whom measles was complicated by more than seven days of diarrhoea, but children 24-47 months old with the same compli- cation also suffered a weight deficit. Initial weight loss during measles infections is probably due to anorexia, stomatitis, withdrawal of solid food, and catabolism caused by the infection. Nitrogen loss in urine and decreased nitrogen absorption are exag- gerated when measles is accompanied by diarrhoea (19). Nutrition supplementation programmes should aim to provide food for children under four years of age who are convalescing after measles, in order to counter the catabolic effects of infection and over- come cultural prohibitions on feeding convalescing children. The interaction between nutrition and infection is known to produce high morbidity and mortality rates (20), in part by promoting secondary infections in the host. Although a synergistic effect of two inter- current infections leading to death has been demon- strated in a number of animal models (21), few examples have been documented in man. This study demonstrates synergy between measles infection and accompanying prolonged diarrhoea, in which the combined effect was greater than the sum of the impact of each infection. Prolonged diarrhoea is probably a result of infection by a heterogeneous mixture of bacterial and viral pathogens, and further studies to identify the superinfecting diarrhoeal agents that produce the greatest impact on nutrition and mortality is important for the design of future treatment regimens. A previous survey reported a considerably lower mortality due to measles in Bangladesh (22), but did not consider secondary infections to be a direct consequence of measles. Extrapolation of the mortality rates from measles in Matlab thana (375/100 000) to the remainder of Bangladesh, would indicate that approximately 75 000 children may die of measles and its complications each year. Therefore the introduction of measles vaccination to Bangladesh would have a major impact not only on survival in childhood, but also on nutritional status. ACKNOWLEDGEMENTS The authors are pleased to acknowledge the support and facilities provided by International Centre for Diarrhocal Disease Research. The study could not have been performed without the diligent work of the epidemiology field team of the Cholera Vaccine Trial. Special appreciation is expressed to Md Khan and J. Chakraborty for supervision of special survey teams, and to Dr J. Josephs of the Maryland State Department of Health for performing the measles serology. RtSUMt EFFET SYNERGIQUE DE LA ROUGEOLE ET DE LA DIARRHEE SUR LA NUTRITION ET LA MORTALITE AU BANGLADESH Afin de reunir des donn&es sur l'impact de la rougeole et des infections secondaires sur la mortalite et la nutrition parmi les enfants au Bangladesh, une enquete prospective sur les menages a et6 faite, portant sur 5775 enfants dans 12 villages: on a surveille pendant une periode de 12 mois les cas de rougeole (923 cas), de diarrhee, et les changements dans l'etat nutritionnel, ainsi que la cause de tous les deces. La rougeole et ses complications ont constitue la principale cause de dtc6s parmi les enfants de 1 A 47 mois. Si l'on considere la totalite des enfants ctudies, le taux de lttalit6 a ete de 3,7%. Chez ceux qui prtsentaient un mauvais etat nutritionnel avant l'episode de rougeole, les taux de mortalite n'ont pas et6 plus eleves que chez les enfants du meme Age dont l'etat de nutrition etait moyen. Une diarrh6e prolong6e (plus de 7 jours) etait la complication la plus frequente des cas de rougeole avec issue fatale. La perte de poids maximale a ett observee chez des enfants atteints de rougeole compliquee par une diarrh6e prolongee, et les 907 908 F. T. KOSTER ET AL. enfants de moins de 4 ans compris dans ce groupe ne sont pas parvenus A combler leur deficit de croissance. Le risque de deces par rougeole compliquee d'une diar- rh&e prolong6e (12%) etait d'environ quatre fois la somme des risques relatifs A la rougeole (2%o) et A la diarrh6e prolong6e (1Wo), ce qui met en evidence une synergie entre rougeole et diarrhee. Au Bangladesh, la rougeole est une maladie grave de l'enfance en raison, d'une part, de son taux de letalite eleve lie A de frequentes infections secondaires et, d'autre par, de l'importante perte de poids aigue, dont les enfants de moins de 4 ans ne parviennent pas A se remettre completement. REFERENCES 1. MORLEY, D. Severe measles in the tropics. British medical journal, 1: 297-300 (1969). 2. PUFFER, R. R. & SERRANO, C. V. Patterns ofmortality in childhood. Washington, DC, Pan American Health Organization, 1973, p.146 (Scientific Publication No. 262). 3. MORLEY, D. C. ET AL. Measles in West Africa. West African medical journal, 16: 24-30 (1967). 4. MORLEY, D. C. ET AL. Measles in Nigerian children. Journal of hygiene, 61: 115-134 (1963). 5. GHOSH, S. & DHATT, P. S. Complications of measles. Indian journal of child health, 10: 111-119 (1961). 6. SCRIMSHAW, N. S. ET AL. Studies of diarrhoea disease in Central America. VIII. Measles, diarrhoea and nutritional deficiency in rural Guatemala. American journal of tropical medicine and hygiene, 15: 625-631 (1966). 7. JELLIFFE, D. M. The assessment of the nutritional status of the community. Geneva, World Health Organization, 1966 (Monograph Series No. 53). 8. BLACK, F. L. In: Manual of clinical microbiology, 2nd ed., ASM, 1974, p.711. 9. DEJONG, J. G. The survival of measles virus in air, in relation to the epidemiology of measles. Archiv fiar die Gesamte Virusforschung, 16: 97 (1965). 10. SHAH, V. ET AL. A test survey of measles in a rural community in India. Bulletin of the World Health Organization, 46: 130-138 (1972). 11. GORDON, J. E. ET AL. Acute diarrheal disease in less developed countries. 2. Patterns of epidemiological behavior in rural Guatemalan villages. Bulletin of the World Health Organization, 31: 9-20 (1964). 12. DOSSETOR, J. ET AL. Persistent measles infection in malnourished children. British medical journal, 1: 1633-1635 (1977). 13. McGREGOR, I. A. Measles and child mortality in The Gambia. West African medical journal, 13: 251-257 (1964). 14. RYDER, R. W. ET AL. Entertoxigenic Escherichia coli and reovirus-like agent in rural Bangladesh. Lancet, 1: 659-663 (1976). 15. SHEENY, T. W. ET AL. Small intestinal mucosa in certain viral diseases. Journal of the American medical association, 190: 1023-1028 (1964). 16. AXTON, J. H. Measles: a protein-losing enteropathy. British medical journal, 3: 79-80 (1975). 17. COOVADIA, H. M. ET AL. An evaluation of factors associated with the depression of immunity and mal- nutrition in measles. American journal of clinical nutrition, 27: 665-669 (1974). 18. WESLEY, A. ET AL. Immunological recovery after measles. Clinical and experimental immunology, 32: 540-544 (1978). 19. SCRIMSHAW, N. S.-ET AL. Infection and kwashiorkor. Journal of tropical pediatrics and environmental child health, 6: 37-43 (1960). 20. SCRIMSHAW, N. S. ET AL. The interaction of nutrition and infection. Geneva, World Health Organization, 1968 (Monograph Series No. 57). 21. MACKOWIAK, P. A. Microbial synergism in human infections. New England journal of medicine, 298: 21-26 (1978). 22. CURLIN, G. C. ET AL. Demographic crisis: the impact of the Bangladesh civil war 1971 on births and deaths in a rural area of Bangladesh. Population studies, 30: 87-105 (1976).