Epidemiology and population dynamics of Ascaris lumbricoides and Trichuris trichiura infection in the same community

TRANSACTIONS OF THE ROYAL SOCIETY OF TROPICAL MEDICINE AND HYGIENE (1987) 81, 987-993 987 Epidemiology and population dynamics of Ascaris lumbricoides and Trichuris trichiura infection in the same community D. A. P. BUNDY *, E. S. COOPERS, D. E. THOMPSON*, J. M. DIDIER~ AND I. SIMMONS~ Parasite Epidemiology Research Group, Department of Pure 6 Applied Biology, Imperial College, University of London, Prince Consort Road, London, SW7 ZBB; Parasite Epidemiology Project, P.O. Box 306, Cashies, St Lucia; 3Minisrry of Health, Casks, St Lucia Abstract The gastrointestinal hehninth infection status of an age-stratified sample from a single Caribbean community was assessed using anthelmintic expulsion techniques. The same sample was re-assessed in a similar manner after a 17 month period of re-infection. The age-prevalence profile of Ascaris lumbricoides was convex while that of Trichuris trichiura was asymptotic. The age-intensity profiles of both species were convex. These differing patterns are attributed to differences in the absolute worm burdens of the 2 species. The frequency distributions of infection intensity were similar for both species, and largely independent of host age. The basic reproductive rate of A. lumbricoides (& = l-1.8) was similar to that recorded elsewhere and much lower than that of T. trichiura (& = 4-6), implying that the latter is intrinsically more resistant to control. Individual hosts were predisposed to high (or low) intensity infection with either species, although predisposition to both species simultaneously was not conclusively demonstrated. Further studies are required to determine the cause of these observations. Introduction Ascaris lumbricoides and Ttihuris trichiura are among the most ubiquitous of human parasites, infecting more than one billion (104 people and affecting the health of communities throughout the world (PETERS, 1978). Safe and effective anthelmintics are available for both helminthiases, yet attempts to control endemic infection by chemotherapy have, with some notable exceptions (APCO, 1980), met with limited success. Improved understanding of the population dynamics of infection in endemic communities may facilitate the design of rational strategies for control (WALSH & WARREN, 1979; ANDERSON & my, 1982). For hehninth parasite populations the intensity is the central statistic determining both the morbidity of infection and the dynamics of transmission (ANDER- SON & MEDLEY, 1985). Meaningful population-based analysis of transmission is therefore dependent on reliable field estimation of this statistic. Intensitv estimates based on indirect measures, such as the density of parasite eggs in stool, are a rather crude approximation (SARLES, 1929) and more reliable analysis can be achieved using direct estimates of gastrointestinal nematode worm burdens obtained by anthelmintic expulsion (SMILLIE, 1924). This procedure is logistically difficult in community applications (CROLL et al., 1982), particularly in the case of T. trichura (BUNDY et al., 1985a, b), but has been successfully used in field studies. These studies have tended to focus on one or other of the 2 hehninth species, principally A. lumbricoides. In a previous article (BUNDY et al., 1987a) we provided quantitative data on the population dynamics of T. trichiura in St Lucia. The present study aims to provide comparative data on the population *Author for correspondence. dynamics of A. lumbrkides and T. trichiura in the same endemic focus in the Americas. Methods The study was conducted in the village of Anse-La-Raye (oooulation anoroximatelv 2500) situated on the Caribbean coast of the is&d of St Lucia. The clinical and demographic characteristics of the study population have been described nreviouslv (BUNDY et nl., 1987a; COOPER & BUNDY, 1986;? HOMPS~N et al., 1986); Faecal specimens were collected from an age-stratified sample (n=244) of the population and examined using the Kato thick smear technique in order to identify infected individuals and to estimate prevalence (see BUNDY er al., 1987a, for details of the sampling procedure). Ah infected individuals were offered mebendaxole therapy (100 mg twice daily for 3 d: Vermoxo, Janssen Pbarmaceutica, Belgium). Total 24h stools were collected for 4 d after treatment from a randomly selected subsample (a= 113) of infected persons recruited from each age class. Counting the worms expelled provides an estimate of infection intensity which is within 90% of the true worm burden (BUNDY et al., 1985a, b). The efficacy of treatment was confirmed by examination of stool specimens collected 30 d after treatment. The rate of reinfection of the population was monitored by monthly stool examination. After 17 months the prevalence and intensity of hehninttt infection approached pre-treatment levels. Treatment was again offered to all infected individuals and, where possible, 4 x 24h post-treatment stools were collected from the same individuals as were examined following the first intervention. Total gastrointestinal hehninth worm burdens acquired during the 17 month period of reinfection were estimated. This study was conducted with the agreement of the Ministry of Health (St Lucia), the St Lucia Medical Association and the Anse-La-Raye Community Association. Permission was received from the individuals concerned or, in the case of children, their guardians. Results Horizontal age-prevalence and age-intendy profiles The age-prevalence relationships of A. lumbricoides

988 EPIDEMIOLOGY AND POPULATION DYNAMICS OF A. lumbricoides AND T. trichiura and T. trichiura in this population are recorded in Fig. 1. A. lumbricoides prevalence rises rapidly with age, maintains a plateau at approximately 80% between the ages of 5 and 15 years, and declines to less than 50% in adults. The age-prevalence profile of T. trichiura is essentially similar in the child age classes, but there is no decline in adulthood. For both species, mean infection intensity increases with age at a similar rate to that of prevalence (Fig. 2). A. lumbricoides attains a peak of 6-S worms per child (variance: 1.75 X lo*; maximum: 59 worms per child) in the 5-10 year age-class, and then declines in adults to a mean of 4.1 worms per person. T. trichiura attains a eak of 96.2 worms per child (variance: 1.57? x 10 ; maximum: 1016 worms) in the 5-10 year age class, and then declines to a mean of 27.3 worms per adult. Frequency distributions of worm burdens The frequency distribution of T. trichiura in this population has been reported previously (BUNDY et ae., 1987a). A. Zumbricoides worm numbers per host were highly aggregated for all the age classes examined. The variance greatly exceeded the mean intensity within each age class, and the negative binomial distribution provided a good description (significant at the 5% level) of observed trends for the total population (Fig. 3). The value of the statistic k (the negative binomial exponent, which varies inversely with the severity of parasite aggregation) was less than one for all ageclasses (Table 1). Fig. 4 records the dependency of k on host age for A. lumbricoides and T. trichiura. Fig. 1. Relationship between A. lumbricoides (&Cl) and T. tichiura (+-+) infection prevalence and host age. Rate of infection in children ana the ba&c repoductive rate, R, From Fig. 2 it is apparent that the rate at which individuals acquire worms (A, the per capita rate of infection) is age-dependent over the age range 0 to 10 years. The rate of change in mean worm burden, M(a), with age, a, may be represented for the child age-classes by the equation (1): dm(a)/da = A -@f(a), where u, the average parasite death rate, and A are assumed to be constant and independent of age. The precise value of p for A. lumbricoides is unknown, but intuitive approximations (CROLL et al., 1982; ANDERSON & MAY, 1985) and estimates based on the ratio of immature:mature worms (ELKINS et al., 1986) Fig. 2. Relationship between mean worm burden and host age: (A) A. lumbkn.des (B) T. nichiura. d 25.--- 1 Fig. 3. Frequency distribution of parasite numbers per person in the total sample. suggest that the life expectancy (l/u) of A. lumbricoides is of the order of one year. Fitting the model by least squares analysis to the observed rate of change in mean worm burden over the age range O-10 years (Fig. 5) yields an estimated A of 4-6 worms per year. Using methods described by ANDERSON (1982) a crude estimate of the basic reproductive rate, R,, can be derived from estimates of A, u, k and the severity of density-dependent constraints on fecundity (.a=o.98). These analyses suggest that R, has a value of l-l.8 (defined per generation time of the parasite), and thus that a mature female worm produces, on so+

D. A. P. BUNDY et al. 989 Table l-frequency distributions of Ascaris Zumbricoides worm numbers per person: summary statistics. Age class 12-17 months 18-23 months 2-4 years S-10 years 1 l-29 years y+j= Mean age (years) 1.17 1.70 2.84 6.66 16.39 42.58 Sample size Fig. 4. Relationship between the population distribution of infection intensity, measured by the negative binomial exponent k (which varies inversely with the severity of aggregation), and host age of A. lumbricdes ( m ) and T. rrichiura (+-+). Fig. 5. A comparison of the predictions of equation (1) in the main text (solid lines) and observed changes in the mean worm load per person in the infant and child age classes (open squares). The assumption that A is independent of age results in the model overestimating worm burdens in infants (line passing through the origin). A better fit is achieved by assuming that no infection occurs in infancy (fine passing through the point M(a) = 0, (I = 1). average throughout her reproductive life span, 1 to 2 females that attain reproductive maturity in the human host (in the absence of density-dependent or other constraints on population growth). Similar analyses for T. trichiura in the same population indicate much higher rates of transmission than for A. lumbricoides: A = 90 worms/child/year; R, = 4-6 (BUNDY et al., 1987a). Total worms i; 129 123 120 4:; Mean burden 2.37 2.62 4.45 6.47 4.80 4.12 4.42 k 0.999 O-458 0.814 0.482 0.708 0.586 Predisposition to multiple species helminth infection In order to determine whether individuals are predisposed to a particular intensity of infection (either high or low relative to others in the population) it is necessary to compare the worm burdens of individuals at the start of the study with the worm burdens acquired by the same individuals after 17 months of reinfection. The convex relationship between intensity and host age introduces a potential source of error into this comparison, however, since age-related differences in the rate of acquisition of infection introduce a trend for children to acquire higher worm burdens than adults (BUNDY et al., 1987b). The data were therefore age-standardized about the means of 3 age classes [O-5-2 years (n= 15); 2-6 years (n=18); and 7+ years (n=20)] using the relationship y = (x - f)/s, where y = the age standardized worm burden; x = observed worm burden; f = mean worm burden for that age class; and s = standard deviation for that age class. These age-standardized data were then compared using the Kendall s rank correlation test (Table 2). Individuals who were intensely infected with A. lumbricoides at the first observation tended to reacquire heavier than average worm burdens of A. lumbricoides during the 17 month period of reinfection, while individuals who were initially lightly infected tended to reacquire light infections. Similarly, the initial worm burdens of T. trichiura were significantly correlated with the reinfection worm burdens of the same species. Thus, if the helminth species are considered separately, it is apparent that individuals are predisposed to heavier (or lighter) than average infection intensities. An examination of correlations between the 2 species, however, produced conflicting results. There was a significant correlation between the worm burdens of the two species at the time of the first observation (r = 0.1812; P<O*Ol; n = 103) and at the second (T = 0.2558; P<O*Ol; n=53). Hence an individual with an intense A. lumbricoides infection is likely simultaneously to harbour an intense infection with T. trichiura. The correlation is weak, however, since significance was lost when the data set was reduced to only those individuals whose worm burdens were assessed at both times of observation (T = 0.0472; not significant; n=53). Analysis of this restricted data set showed no significant correlation between the initial worm burden of one species and the intensity acquired after reinfection with the other (Table 2). An individual who has a heavier than average worm burden of A.

990 EPIDEMIOLOGY AND POPULATION DYNAMICS OF A. lumbricoides AND T. trichiura,o -co w _ 2 -CD -d -(v 111 IIll III I,, II,,,, -0 II I I II I I II, I, I I I,,, -I 1. I I I I I I 1 I I I I 1 I I I I I -- 0

Table 2-Predisposition to infection with Ascaris lumbricoides and Trichuris trichiura. Kendall s rank correlation coefficient for age-standardized worm burdens (n=53) Initial infection ZFity intensity after 17 months of reinfection Trichuris Ascaris 0~1794(P<O~O5) O.l071(NS) Trichuris O.O958(NS) 0.6417(P<O.O1) NS: not significant. lumbricoides is not necessarily more likely to acquire a heavier than average infection with T. trickiura after a period of reinfection. This conclusion may indicate that the sample size is too limited to permit detection of weak underlying trends. Discussion The mean intensity of A. lumbricoides and T. trickiura is observed to increase in childhood and decline in adulthood (Fig. 2). This convex relationship between mean worm burden and age has been described for all the major gastrointestinal hehninths of humans (ANDERSON. 1982: BUNDY. 1986). The precise cause is unknown but is assumed to involve some form of age-related change in exposure or susceptibility (ANDERSON& MAY, 1985). The rate of increase in prevalence during childhood is very similar for both A. lumbricoides and T. trickiura, presumably reflecting their similar modes of infection. The rate of change in prevalence in the adult age classes differs for the 2 species: the prevalence of T. trichiura maintains a relatively constant value throughout adulthood, while the prevalence of A. lumbtioides declines with age. This marked difference in the age-prevalence profiles of the 2 species appears to be a consistent observation (cf. Figs 2 and 3 in ANDERSON & MAY, 1985), which may be attributable to differences in the absolute size of worm burden typical of the 2 species. The urecise relationshin (ANDERSON 1982) between prevalence and mean intensity is given by equation (2): p(a) = 1 - [l + M(a)/klmk, wherep(a) = rate of change of prevalence with host age, a; M(a) = rate of change of mean worm burden with host age, a; and k = the negative binomial exponent which varies inversely with the severity of aggregation. For the very low values of k observed for gastrointestinal hehninth populations, this relationship implies that major decreases in mean intensity result in modest declines in prevalence (see ANDERSON & MAY+, 1985, Fig. 10 for numerical illustration of this relatronship). The relationship is, however, nonlinear such that, for a given low value of k, prevalence is more sensitive to changes in mean worm burden when the absolute size of the worm burden is small. A practical consequence of this is that prevalence is more likely to show significant decline with decreasing worm burden for x lumbricoides infection (where mean worm burdens are typically of the order of S-10 worms) than for T. trickiuru infection (where mean worm burdens are at least an order of magnitude larger). It has been suggested that A. lumbricoides has 2 forms of age-prevalence distribution: one in which prevalence declines in adulthood and another in which prevalence remains high in all age groups D. A. P. BUNDY et al. 991 (FEACHEM et al., 1983). The authors suggested that these differing distributions reflect differences in environmental- and cultural factors. The present study, however, shows that both types of distribution may occur in the same environment and commtmity, for 2 parasite species with very similar modes of infection. It might be concluded that the different distributions observed for A. lumbricoides reflect differences in the mean intensity of infection at different localities. This study provides information on the relationship between host age and the frequency distribution of parasite numbers per host, for 2 parasite species at the same locality. Previous studies have examined this relationship for a single species (Fig. 6). Earlier analyses of this relationship suggested a trend for aggregation to decrease (K to increase) with host age (ANDERSON, 1982). Inspection of the data sets now available suggests instead that the degree of aggregation is largely independent of host age, although there is some indication of age-dependent increase in the younger age classes. Indeed! the value of k is remarkably consistent for the different age-classes and localities, given that this unit-less statistic has a potentially infinite range of values. This observation is in agreement with theoretical studies which have suggested that the degree of parasite aggregation is a crucial determinant of macroparasite population stability (MAY & ANDERSON, 1978), and have predicted that the quantities characterizing the interactions between natural populations would be a far-fromrandom set. A second surprising feature is the remarkable similarity of the k values for A. lumbricoides and T. frichiuru in the same community. Previous analyses, which compared mean k values obtained from a variety of different communities, had suggested that T. trickiura populations were more severely aggregated than those of A. lumbricoides (BUNDY et al., 1985~; ANDERSON & MAY, 1985). The present results suggest that, under the same environmental conditions, both species have very similar frequency distributions, and hence that the differences observed in different localities may be a function of environmental differences, The basic reproductive rate of A. lumbricoides has a surprisingly similar range of values in four widely separated communities: 1.7 in Burma and Bangladesh; 1 to 2.1 in Korea; and 1 to 2 in St Lucia (THEIN-HLIANG et al., 1983; MARTIN et al., 1983; CHAI et al., 1985). The present estimates confirm that T. trickiuru has a higher transmission rate than A. lumbricoides even under the same environmental conditions. This clearly demonstrates that whipworm is intrinsically more resistant to control than roundworm. Children acauire T. trickiuru at a rate of 90 year- whereas the iate for A. lumbricoides is of the order of 4 to 6 worms year-, The results indicate that the mtensity of infection reacquired by an individual following treatment is significantly correlated with the intensity of infection before treatment. The correlation is indenendent of host age. These observations suggest th& an individual is predisposed to heavy (or light) infection with A. lumbricoides or T. trichiura. Previous studies have indicated predisposition of humans to hookworm (SCHAD& ANDERSON, 19851, A. lumbricoides (CROLL

992 EPIDEMIOLOGY AND POPULATION DYNAMICS OF A. lumbricoides AND T. trichiura et al., 1982), Enterobius vewnicularis (HASWELL- ELKINS et al., 1987), and T. trichiura infection (BUNDY et al., 1987b). It is not clear whether hosts are simultaneously predisposed to both hehninth species. Two previous studies failed to show any correlation between the infection intensities of A. lumbricoides and T. trichiura in multiple species infections (CROLL et al., 1982; COOPER & BUNDY, 1986). The present study indicates that such correlations may exist, but that they are weak and sensitive to sample size. A recent study of a relatively large population in India (n=525) has provided firm evidence for multiple species predisposition to A. lumbricoides, E. vermicularis and hookworm infection, although the evidence for a similar correlation with T. trichiura infection is less convincing due to the very low rates of worm recovery attained by the anthehnintic employed (HASWELL- ELKINS et al., 1987); The available evidence is therefore contradictory, although those studies which have involved large sample sizes and precise worm burden assessment tend to indicate a trend towards multiple species predisposition. This study is unusual in that it attempts to examine the population dynamics of 2 hehninth species endemic in the same community. The findings indicate that current population dynamical theory can adequately describe the observed behaviour, but further studies are required to explain the underlying causative mechanisms. Acknowledgements We thank the Ministry of Health (St Lucia) for technical assistance, the people of Anse-La-Raye for their enthusiastic cooperation, and Janssen Pharmaceutics for supplying the anthehnintic. The Parasite Epidemiology Project (St Lucia) is a collaborative activity of the Ministry of Health (St Lucia), the University of the West Indies, and Imperial College, Universitv of London. Major funding for the Project is provided-by the Wellcome Trust. References Anderson, R. M. (1982). The population dynamics and control of hookworm and roundworm infections. In: Pomdation Lhmamics of Infectious Diseases. R. M. Anderson (editor).*london: Chapman and Hall, pp. 67-106. Anderson, R. M. & May, R. M. (1982). Population dynamics of human helminth infections: control by chemotherapy. Nature, 297, 557-563. Anderson, R. M. &May, R. M. (1985). Helminth infections of humans: mathematical models, population dynamics and control. Advances in Parasitology, 24, l-101. Anderson, R. M. & Medley? G. F. (1985). Community control of helminth infections of man by mass chemotheraw. Parasitolow. 90. 629-660. APCO (l980). Colleczd Papers on the Control of Soiltransmitted Helminthiases. I. Asian Parasite Control Organization, Tokyo, Japan. Bundy, D. A. P. (1986). The epidemiology of Trichuris and trichuriasis in Caribbean communities. Transactions of the Royal SOCieN of Trotrical Medicine and Hygiene, _- 80, 706-718. - Bundy, D. A. P., Thompson, D. E., Golden, M. H. N., Cooper, E. S., Anderson, R. M.. & Harland,, P. S. E. (1985a). Population distribution of Trichuris zrtchiura in a community of Jamaican children. Transactions of the ;3?-;lSociety of Tropscal Medicine and Hygtene, 77, Bundy, D. A. P., Thompson, D. E., Cooper, E. S. & Blanchard, J. (198513). Rate of expulsion of Trichuris trichiura with multiple and single dose regimens of albendazole. Transactions of the Royal Society of Tropical Medicine and Hygiene, 79, 641-644. Bundy, D. A. P., Thompson, D. E., Cooper, E. S., Golden, M. H. N. & Anderson, R. M. (1985~). Population dynamics and chemotherapeutic control of Trichuris trichiura infection of children in Jamaica and St Lucia. Transactions of the Royal Society of Tropical Medicine and Hygiene, 79, 759-764. Bundy, D. A. P., Cooper, E. S., Thompson, D. E., Anderson, R. M. & Didier, J. M. (1987a). Age-related mevalence and intensitv of Trichuris trichiura in a St Lucian community. Transactions of the Royal Society of Tropical Medicine and Hygiene, 81, 85-94. Bundy, D. A. P., Cooper, E. S., Thompson, D. E., Didier, 1. M., Anderson, R. M. & Simmons. I. (1987b). Predisposition to Trichuris trichiura infection in humans. Epidemiology and Infection, 98, 65-71. Chai, J., Kim, K., Hong, S., Lee, S. & Seo, B. (1985). Prevalence, worm burden and other epidemiological narameters of Ascaris lumbricoides infection in rural communities in Korea. Korean Journal of Parasitology, 23, 241-246. Cooper, E. S. & Bundy, D. A. P. (1986). Trichuriasis in St Lucia. In: Diarrhoea and Malnuttition in Children, A. S. McNeish and J. A. Walker Smith (editors). London: Butterworths, pp. 91-96. Croll, N. A., Anderson, R. M., Gyorkos, T. W. & Ghadirian, E. (1982). The population biology and control of Ascaris lumbricotdes in a rural community in Iran. Transactions of the Royal Society of Tropical Medicine and Hygiene, 76, 187-197. Elkins. D. B., Haswell-Elkins. M. & Anderson. R. M. (1986). The epidemiology -and control of intestinal helminths in the Pulicat Lake region of Southern India. 1. Study design and pre- and post-treatment observations on Ascaris lumbricoides infection. Transactions of the Royal Society of Tropical Medicine and Hygiene, 80, 774-792. Feachem, R. G., Bradley, D. J., Garelick, H. & Mara, D. D. (1983). Sanitation and Disease: Health Aspects of Excreta and Wastewater Management. World Bank Studies in Water Supply and Sanitation No. 3. New York: John Wiley & Sons. Haswell-Elkins, M. R., Elkins, D. B. & Anderson, R. M. (1987). Evidence for predisposition in humans to infection with Ascaris, hookworm, Enterobius and Trichuris in a South Indian Fishing Community. Parasitology, 95, 323-337. Martin, J., Keymer, A., Isherwood, R. J. & Wainwright, S. M. (1983). The prevalence and intensity of Ascaris lumbricoides infections in Moslem children from northern Bangladesh. Transactions of the Royal Society of Tropical Medicine and Hygiene, 77, 702-706. May, R. M. & Anderson, R. M. (1978). Regulation and stability of host-parasite population interactions II. Destabilizing processes. Journal of Animal Ecology, 47, 249-267. Peters, W. (1978). Medical aspects - comments and discussion. In: The Relevance of Parasitology to Human Welfare Today, A. E. R. Taylor and R. Muller (editors). Symposia of the British Society for Parasitology, 16. Oxford: Blackwell Scientific Publications. Sarles, M. P. (1929). The effect of age and size of infestation on the egg production of the dog hookworm Ancylostoma caninum. American Journal of Hygiene, 10, 658-660. Schad, G. A. & Anderson, R. M. (1985). Predisposition to hookworm infection in man. Science, 228, 1537-1540. Smillie, W. G. (1924). Control of hookworm disease in south Alabama. Southern Medical Journal, 17, 494-502. Thein-Hlaing, Than-Saw, Htoy-Htay-Age, Myint-Lwin & Theia Muang Myint. (1983). Epidemiology and transmission dynamics of Ascaris lumbricoides in Okpo village, rural Burma. Transactions of the Royal Society of Tropical Medicine and Hygiene, 78, 497-504.

D. A. P. BUNDY et al. 993 Thomnson. D. E., Bundy, D. A. P., Cooper, E. S., Golden, health care: an interim strategy for disease control in B.-E. & Schan tz, P. (1986). Epidemiology of toxocariasis developing countries. New EnglandJoumal of Medicine, in St Lucia. Bulletin of the World Health Organization, 64, 301, 967-974. 283-290. Walsh, J. A. & Warren, K. S. (1979). Selective primary Accepted for publication 5 May 1987 ANNOUNCEhIENT European Congress on Palliative Care University of Milan, Italy, 23-25 April 1988 Main lectures: Palliative Care @alfour Mount, Canada); Ethical Issues (Abdr6 Churaqui, Israel); Pain Control (R. G. Twycross, USA); Emotion Problems (G. P. Maguire, UK); Dying and Death (Elisabeth Kubler-Ross, USA); Grief and Bereavement (Colin M. Parkes, UK). Information: AISC, via Domenichino 11, 20149 Milan, Italy. ANNOUNCEMENT Fist International Congress of Tropical Neurology Paris, 27-29 April 1988 Information: Prof. Gaosguen, Service de Neurologie, Hhpital d Instructions des Armies du Val de Grace, 74 boulevard de Port-Royal, 75230 Paris cedex 05, France.