Density-dependent processes in the transmission of ...

75 downloads 0 Views 1MB Size Report
(Collins et al. 1980). Skin loads reported by De Leon. & Duke (1966) were doubled to account for the fact that mff were counted after 025 h but that biopsies.
115

Density-dependent processes in the transmission of human onchocerciasis: intensity of micronlariae in the skin and their uptake by the simuliid host M. G. BASANEZ 1 2 *, M. BOUSSINESQ 3 , J. PROD'HON 3 , H. FRONTADO 4 , N. J. VILLAMIZAR 4 , G. F. MEDLEY 2 and R. M. ANDERSON2 1

Instituto de Medicina Tropical, Universidad Central de Venezuela, Caracas, Venezuela Department of Biology, Parasite Epidemiology Research Group, Imperial College, Prince Consort Road, London SW7 2BB, UK 3 Antenne ORSTOM aupres du Centre Pasteur, BP 1274, Yaounde, Cameroon 4 Centra Amazdnico de Investigacidn y Control de Enfermedades Tropicales (CAICET), Puerto Ayacucho, Venezuela. 2

Amazonas,

(Received 29 January 1993; revised 2 June 1993; accepted 5 June 1993) SUMMARY

The transmission success of Onchocerca volvulus is thought to be influenced by a variety of regulatory or density-dependent processes that act at various points in the two-host life-cycle. This paper examines one component of the life-cycle, namely, the ingestion of microfilariae by the simuliid vector, to assess the relationship between intake of larvae and the density of parasites in the skin of the human host. Analysis is based on data from three areas in which onchocerciasis is endemic and includes published information as well as new data collected in field studies. The three areas are: Guatemala {Simulium ochraceum s.l.), West and Central Africa (savanna members of the 5. damnosum complex), and South Venezuela (S. guianense). The data record experimental studies of parasite uptake by flies captured in the field and fed to repletion on locally infected subjects who harboured varying intensities of dermal microfilarial infection. Regression analyses of log transformed counts of parasite burdens ingested by the flies plotted against log transformed counts of microfilariae per mg of skin revealed little evidence for saturation in parasite uptake by thefliesas the intensity in the human host increased. There was a positive and highly significant rank correlation between both variables for the three blackfly species. In an alternative analysis a model was fitted to data on prevalence offlieswith ingested microfilariae (mflf) versus dermal mean intensities. The model assumed an overdispersed distribution of the number of mff/fly and a given functional relationship between intake and skin load. The results of both approaches were consistent. It is concluded that parasite ingestion by the vector host is not strongly density dependent in the three geographical areas and ranges of dermal loads examined. It therefore appears that this transmission process is of reduced importance as a regulatory mechanism in the dynamics of parasite population growth. Key words: Onchocerca volvulus, simuliid hosts, microfilarial intake, microfilarial concentration, density dependence.

INTRODUCTION

In the complex life-cycles of the filarial parasites, many factors influence the transmission dynamics of the parasite and observed epidemiological patterns of infection in human and vector populations. How these various factors combine and interact to determine observed patterns is difficult to assess in the absence of a mathematical framework that mimics the transmission dynamics of the parasite and the biological details of its life-cycle (Dietz, 1982; Anderson & May, 1991). Such a framework has proven to be very useful in shaping our understanding of the population biology of a variety of macroparasitic infections, including intestinal nematodes (Anderson, 1982 a; Anderson & May, * Reprint requests to Maria-Gloria Basanez, Department of Biology, Parasite Epidemiology Research Group, Imperial College, Prince Consort Road, London SW7 2BB.

1985), schistosome species (Macdonald, 1965; Cohen, 1976; May, 1977) and the filarial parasites (Hairston & Jachowski, 1968; Rochet, 1990; Grenfell & Michael, 1992). In the case of the filariases most attention has been directed towards the development of mathematical models of the transmission dynamics of human onchocerciasis (Dietz, 1982; Wada, 1982; Remme et al. 1986; Davies, Weidhaas & Haile, 1987; Plaisier et al. 1990). This is hardly surprising given the seriousness of the infection as a public health problem in equatorial Africa plus the volume and quality of epidemiological studies carried out under the umbrella of the Onchocerciasis Control Programme (OCP) in the western part of that continent (WHO, 1985, 1987). As noted by Dietz (1982), a variety of density-dependent processes acting at various points in the life-cycle of Onchocerca volvulus may have an important influence on the transmission

Parasitology (1994), 108, 115-127 Copyright © 1994 Cambridge University Press 9

PAR 108

M. G. Basdnez and others

116

Table 1. Sources of data and measures of central tendency used in this work Vector species

Locality

Simulium guianense

References

Williams Square root transformed Square root transformed Williams

DeLeon& Duke (1966) Collins et al. (1977).

Cote d'lvoire

Williams

Philippon (1977).

Central Africa North Cameroon South Venezuela

Williams

Boussinesq (1991).

Arithmetic

This work.

Guatemala

Simulium damnosum s.l. (savanna spp.) S. damnosum s.s. and 5. sirbanum S. damnosum s.s. and S. sirbanum

Type of mean used

West Africa Upper Volta

of the infection and on the long-term stability of parasite populations. In this context most emphasis has been placed on the adult parasite in the human host, given its long reproductive life-span (9-11 years on average, Plaisier et al. 1991 a), relative to the duration of the cycle within the simuliid host and the life-expectancy of infected vectors (measured in terms of weeks, Dietz, 1982; Anderson & May, 1985; Nelson, 1991). Less attention has generally been devoted to the development of models explicitly taking into account the possibility of density dependence in the population processes governing parasite transmission to and from the dipteran vector. However, density-dependent limitation of larval establishment within the fly is virtually the only source of regulation in the simulation model ONCHOSIM (Plaisier et al. 1990). As in other areas of ecological study, density dependence simply implies that a rate parameter controlling population size is a function of population density. If the relationship is such that a transmission rate, birth rate or survival rate decreases as population density rises, then the densitydependent mechanism will act as a regulatory constraint on population growth. With respect to the vectors of O. volvulus, density dependence may act on the capture of microfilariae from the skin of an infected human host during a bloodmeal, on the numbers ingested that gain access to the suitable site for further development in the insect (thoracic muscles), on the numbers maturing to the infective L3 stage for transmission to the next human host, on the size of the inoculum of L3s delivered to the next host during a subsequent feed, and on the survival of infected flies (parasite-induced vector mortality). In West Africa the vectors of O. volvulus are members of the Simulium damnosum complex of sibling species (Vajime & Dunbar, 1975). In Latin America, where the disease also occurs and is posing

Campbell et al. (1980). DeLeon & Duke (1966)

an increasing public health problem in countries such as Ecuador (Guderian et al. 1988), the diversity of vectors is far greater and each species is likely to be also a complex of species (Crosskey, 1990; Shelley, 1991). Detailed epidemiological studies are presently required in Latin America because, with the recent advent of Ivermectin as a safe chemotherapeutic agent, control programmes are being started in the affected countries. The strategies devised for West Africa will undoubtedly need modification to take account of the different epidemiological and entomological settings (Shelley, 1991; Cupp, 1992). Although there are many papers on the relationship between Onchocerca and Simulium, few attempts have been made to gather the data needed in order to examine in detail the evidence for density dependence across different geographical areas, vector species and parasite loads. In this paper we investigate the first component of this relationship, namely, the rate of uptake of microfilariae from the skin of infected human hosts by the simuliid vector. Microfilarial intake by the fly may depend on the density of parasites in the skin because, among other factors, dermatological changes associated with increasing intensity of infection may impair larval uptake and hinder transmission to the vector (Kershaw, Duke & Budden, 1954; Duke, 19626). We examined data from Guatemala (S. ochraceum s.l.), West and Central Africa (S. damnosum s.l.) and South Venezuela (S. guianense) and based our analyses on published and unpublished material as well as on new field studies in Latin America. MATERIALS AND METHODS

Sources of data

Table 1 summarizes the sources of data used in this work. Only data of the savanna species of

117

Onchocerca volvulus microfilarial intake by simuliid vectors the 5 . damnosum complex (S. damnosum s.s. and

5. sirbanum) are included since they are the vectors of the parasite variety that more commonly induces blindness in the human host (WHO, 1987) and hence have been the focus of much research by OCP. Parasitological and entomological procedures

All the results analysed in this paper are based on flyfeeding experiments in which samples of wild flies of the local vector species were engorged to repletion on subjects harbouring different microfilarial skin loads of the local population of O. volvulus. Assessment of parasite load in the human host. The

methods of quantifying microfilardermic levels in the participant subjects prior to allowing the flies to feed on them vary among the cited studies, but generally they all follow the well-known procedures of skin-snipping (Holth- or Walser-type of punch), incubating the snips in aqueous medium, counting the emerging larvae, and expressing the counts as numbers of microfilariae (mff) per skin snip (ss) or per milligram (mg). Standardization of the results for comparative purposes is described below. In South Venezuela (Yanomami localities of Parima B and Orinoquito, Amazonas State), a minimum of 5 biopsies were taken with a Holth sclerocorneal punch from the particular body zone on which the flies were subsequently engorged. The sensitivity and reliability of the skin-snip procedure has been shown to increase with the number of snips taken up to a figure of 6 biopsies (Taylor et al. 1987, 1989; WHO, 1987). The snips were incubated for 24 h in buffered saline solution, scored and weighed according to the procedure of Yarzabal et al. (1985), and results expressed as the number of mff/mg. Details of the geographical location and ecological characteristics of the Parima area have been reported by Basanez & Yarzabal (1988) and Basafiez et al. (1988). Standardization of units and incubation times. In this

paper, all microfilarial counts in the human host are expressed as numbers of mff/mg of skin. Unless this figure is specifically provided by the authors, an average of 2-84 mg/ss is applied to the African data when the biopsies are taken with a Holth punch (Prost & Prod'hon, 1978; Boussinesq, 1991). When skin snips were examined well before 24 h, the reported results were multiplied by a factor (estimated from Collins et al. 1980) that varied according to the incubation time and procedure. The period of 24 h was chosen as the maximum because snips incubated for this time release around 85 % of their total microfilarial content (estimated by digestion of the biopsies with collagenase), the remaining 15 % being located deep in the dermis and hence unlikely to be available for transmission

(Collins et al. 1980). Skin loads reported by De Leon & Duke (1966) were doubled to account for the fact that mff were counted after 025 h but that biopsies had been teased before incubation (as described by Kershaw et al. (1954) and Duke (19626). The rest of the Guatemalan skin figures (Collins et al. 1977; Campbell et al. 1980) were corrected following Collins et al. (1980) and Boussinesq (1991), who considered that an incubation time of 0-5 h in tissue culture medium resulted in an emergence of approximately 50-60 % of the 24 h microfilarial count. Likewise, Philippon (1977) incubated the snips for 0-5 h and his results were doubled. Boussinesq (1991) used the same method described for the South Venezuelan data set. Fly collection and dissection. In the different studies, groups of wild flies were fed to repletion on the particular body regions of the participants from which dermal intensities had been determined. The biting females were collected at random in such areas (all Guatemalan and African studies) or forced to feed on them if necessary by dislodging the insects located on other parts of the body and inducing them to engorge on the selected areas (this proved to be feasible with S. guianense in South Venezuela). Although wild populations are prone to be a mixture offliesof different ages, Philippon & Bain (1972) and Boussinesq (1991) have not detected any significant difference between the microfilarial intakes by nulliparous and parous females. The engorged insects were dissected either immediately or several hours after the bloodmeal (usually within the first 12 h to avoid loss of mff through digestion). Ingested counts corresponded to the total number of mff found in the body of the fly, either from 'fresh' dissections or 'thick-smears'. Analysis of data Measures of the average number of mff in the human

and vector hosts. Due to the various ways in which mean intensities are reported in the different data sources, it was not possible to choose a single measure of central tendency. The means most commonly used in the published literature were the arithmetic mean (AM), the Williams' mean (WM) and the Square Root Transformed Mean (SQRTM). The usage of the last two reflects the fact that the numbers of mff/ss or mg, or/fly are very seldom normally distributed. Strictly speaking, Williams' means (a modification of the geometric mean to take into account zero records; Williams (1937)) are appropriate when the logarithmic transformation normalizes the distribution of the counts, which are originally overdispersed (variance/mean > 1), whilst Square root transformations are adequate when the square root of the counts normalizes their distribution (previously Poisson with variance/mean 9-2

118

M. G. Basdnez and others Table 2. Ingestion of microfilariae of Onchocerca volvulus by Simulium ochraceum s.l. from subjects with different skin intensities in Guatemala (See Table 1 and text for type of means used.) Study participants Flies

Patient code no.* Cl C7 CC3 CC1 C2 C3 C8 CC5

Dl-a CC10 CC2

C5 CC6 CC7 C6 C4 CC8

ecu CC4 CIO DII C9

CC12 CC13 CC9

Dl-b

No of skin snips 2 2 6 6 2 2 2 6

+4 6 6

2 6 6 2 2 6 6 6 2

+4 2 6 6 6

+4

(no. mff/ mg)t

M 1 d (no. mff/ mg)J

No. of flies

Mean intake (no. mff/fly)

0000 0-500 1-256 2-363 2-800 6-300 6-500 6-883 6000 12-816 15-550 17-700 18-618 25-821 28-500 30-300 35-222 48-492 53-219 60-800 63-000 83-600 86-792 97-277 111-824 183000

0000 0-833 2093 3-938 4-667 10-500 10-833 11-472 12000 21-360 25-917 29-500 31030 43035 47-500 50-500 58-703 80-820 88-698 101-333 126000 139-333 144-653 162128 186-373 366000

24 24 16 16 24 20 24 16 69 16 16 15 16 16 20 20 16 16 16 20 67 20 16 16 16 64

1-700 44-700 3-924 0168 2-600 12-500 51100 14-578 9000 15-791 25-561 26-900 15-734 37-984 43100 24-200 83-698 193-291 301-850 117-800 170000 82-200 137-397 83-722 249-677 390000

M

1 A

* D, DeLeon & Duke (1966); C, Collins et al. (1977); CC, Campbell et al. (1980); Dl-a and Dl-b, same participant, different body zones (a, legs, b, torso). + 4: More than four biopsies taken. f Skin loads as published. | Skin loads as corrected for 24 h incubation. « 1). For comparative purposes it is necessary to back transform in order to obtain meaningful figures (Elliot, 1977; Armitage & Berry, 1987). All three means are used in this paper (Table 1) and they are calculated as follows: WM = Antilog {27 [log(mff count + 1)]/«}-1, (Williams, 1937) SQR TM = {X |y(mff count + 1 )]/M} 2 - 1 , (Collins et al. 1977) where n is sample size. Statistical methods. The relationship between mean microfilarial intake per fly and mean skin load was explored by linear regression (Bartlett's three-group method and least-squares procedure) and nonparametric correlation (Spearman rank correlation test). These choices were based on the following criteria (Bailey, 1981; Sokal & Rohlf, 1981; Armitage & Berry, 1987). The two variables in question are both subject to random variation; as a consequence, the methods for fitting the regression line will depend

on whether this line is going to be used mainly for purposes of prediction (least squares generally applicable), or for investigating the functional relationship between the variables (other methods such as Bartlett's are required since estimates of slopes of y on x when both are subject to error are usually biased). Since we are interested in comparison of slopes as well as in prediction, we report the results of Bartlett's three-group method (Sokal & Rohlf, 1981) and of least-squares regression weighted by the sample size of each data point (Armitage & Berry, 1987). Whilst the functional relationship is discussed here, predictions of the numbers of mff ingested from a given skin load will be used in a subsequent publication. Data points for which the absolute standardized residual values were greater than 3 were considered 'outliers' and excluded from the analysis (Schlotzhauer & Litell, 1987). Since parametric correlation analysis requires a normal bivariate distribution for the pair of measurements in question, and it has been

119

Onchocerca volvulus microfilarial intake by simuliid vectors

Table 3. Ingestion of microfilariae of Onchocerca volvulus by Simulium damnosum s.l. from subjects with different skin intensities in West and Central African savanna Flies

Study participants Patient : No. of skin code snips no.* B3 B2 Bl B5 B4 B6 B8 B7 B9 B12 BIO B13 Bll B14 B15 B16 B17 B18 B20 B19 B22 B21 B23 B24 B26 DD1 B25 PHI B27 B28 B29 B30 PH2 B31

4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 2 4 4 4

+4 2 4 2 1 1 4 4 2

(no. mff/ mg)t

No. of flies

Mean intake (no. mff/ fly)J

000 000 0-00 0-07 007 015 0-27 0-51 0-57 0-78 0-92 112 1-30 1-74 1-77 1-82 2-00 2-83 3-81 4-50 917 14-37 14-54 16-70 18-75 30-00 31-62 35-50 4602 9401 104-23 11304 19300 43806

147 115 104 150 32 95 86 156 100 155 100 182 159 151 99 116 142 137 157 220 151 128 101 211 192 51 139 187 85 44 42 93 168 29

018 1-44 001 007 014 2-65 0-07 2-35 0-29 0-59 107 018 5-47 2-61 0-45 2-75 1-44 3-66 9-59 102 4-83 616 2-80 5-91 19-45 110 7-32 10-60 5-12 12-65 22-49 137-81 16200 141-82

Mean load

Flies with mff (%) 816 46-96 0-96 3-33 6-25 6316 4-65 61-54 1900 4000 5000 13-74 75-47 62-91 30-30 65-52 60-56 77-37 90-45 55-45 84-77 84-38 66-34 81-99 9115 90-65 88-80 84-71 95-45 97-62 98-92 98-90 96-55

* DD, De Leon & Duke (1966); PH, Philippon (1977); B, Boussinesq (1991). + 4: More than four biopsies. f Skin loads corrected for 24 h incubation. %, f Williams' means (see Table 1 and text).

mentioned already that counts of mff in the skin or in the flies seldom have this property, the nonparametric Spearman rank correlation test was chosen (Campbell et al. 1980; Bailey, 1981; Sokal & Rohlf, 1981). These analyses were performed with CSS: Statistica® (Complete Statistical System) software. An alternative approach to the investigation of the relationship between ingested and dermal mff was attempted by analysing the relationship between the prevalence of the infection in the flies and the mean intensity in the skin snips. Maximum likelihood estimation procedures were used to calculate the parameters and their asymptotic confidence limits (Cox & Hinkley, 1974) as described by Guyatt et al. (1990). The likelihood ratio statistic (which has

approximately a chi-squared distribution with degrees of freedom equal to the difference between the number of parameters in the models under testing) was applied to compare the values of the log likelihood functions obtained under different assumptions (Sokal & Rohlf, 1981; Cox & Oakes, 1984; Armitage & Berry, 1987).

RESULTS

Tables 2, 3 and 4 summarize the three data sets from Guatemala, West and Central Africa, and South Venezuela respectively. A raw scatter plot of the data points for the three data sets indicated an increase in the spread of the

M. G. Basdnez and others

120

Table 4. Ingestion of microfilariae of Onchocerca volvulus by Simulium guianense from subjects with different skin intensities in Southern Venezuelan Amazonas Study participants

Flies

Patient code no.

No. of skin snips

Mean load (no. mff/ mg)*

No. of flies

Mean intake (no. mff/ fly)t

Flies with mff (%)

COYI DAVIDI COYII MAYUBA CECILIO COYIII COYIV DAVIDII SHACO

16 7 14 5 6 12 16 8 6

19-569 22-897 47-745 65040 135138 193-387 322-311 463-667 734-654

28 69 31 10 22 23 62 101 26

15-929 9-246 37-290 80-700 79182 108130 163-226 149-564 356039

71-43 59-42 80-65 10000 90-91 91-30 8710 9901 10000

* Skin loads for 24 h incubation. f, * Arithmetic means (see Table 1 and text). observed average microfilarial intakes with mean skin load (not shown). A log-log transformation (having previously added 1 to each mean count in order to allow inclusion of the zeros) stabilized the variance of the residuals. T h e log-log transformed data and the subsequent Bartlett's regression line are presented in Fig. 1 A (S. ochraceum s.L), Fig. 1 B

(S. damnosum s.L) and Fig. 1 C (S. guianense). All of them show a positive relationship between the logarithms of both mean counts. The results of the linear regression and Spearman correlation analyses on the transformed variables are presented in Table 5. Since Bartlett's procedure does not yield a conventional least squares regression line, special methods are required for the computation of confidence limits (Sokal & Rohlf, 1981). Significance testing on the regression coefficients is based upon comparison of their confidence intervals which, due to the reasons aforementioned, are shown for Bartlett's method only. A b coefficient equal to 1 in the log-log transformed regression is indicative of linearity or proportionality between the original variables when the process is reversed, implying absence of density-dependence. Regression slopes smaller than 1 translate into non-linear relationships (log y = a' + b log x, then y (x) = axb, where log a = a'). T h e 95 % confidence limits for the slope of the regression on the Guatemalan and Venezuelan data sets include 1, suggesting that the relationship between the two variables in question is not significantly different from linear for these two simuliid species. Although the confidence limits around the b coefficient for S. damnosum s.l. do not include 1, the interval overlaps with those estimated for both S. ochraceum s.l. and S. guianense, implying that the relationship between the original variables is probably weakly non-linear in the case of the African

savanna data set. In all three cases the Bartlett's confidence interval included the least squares b estimate. In the case of the least-squares method the chosen regression line was forced through the origin because the initial estimates of the intercept a were not significantly different from zero (95 % confidence limits around a included zero in all three blackfly species) and because the residuals were observed to become more normally distributed under this assumption. The relationship derived from the least squares method will be used elsewhere. Spearman correlations were positive and highly significant in all three species. As an alternative method of estimation of the parameters governing the relationship between ingested and skin mff, the proportion of flies harbouring parasites in their bloodmeal was plotted against the intensity of the infection in the skin of the human host for the African savanna and Venezuelan data sets (only data points with sample sizes > 20 were included). There were no data of this sort in the papers on S. ochraceum s.L, since their authors reported the proportion of flies with thoracic but not with ingested mff. This kind of analysis has proved to be useful in understanding the relationship between prevalence, mean intensity and degree of aggregation in other infections (Anderson, 1987; Guyatt et al. 1990; Lwambo et al. 1992; Medley et al. 1993). T h e number of mff ingested/fly has been observed to be highly overdispersed (data not shown), and we assume that the relationship between the proportion of flies infected and the mean microfilarial intake (y(x)) is strongly non-linear with the form: Proportion of flies infected = 1 —{[1 +y(x)/k]~k}, where k is an inverse measure of the severity of

Onchocerca volvulus microfilarial intake by simuliid vectors

121

overdispersion and y(x) a given functional relationship between the mean intake (y) and the mean skin load (x). As a result, different models can be fitted to the data. Of the several functional forms of y(x) tested, two were finally chosen, the power model (y(x) = axb, with a = 1) and a linear model not forced through zero (y(x) = a + bx), both with k independent of the mean. The results of the maximum likelihood estimation of the parameters, their 95 % asymptotic confidence limits and the corresponding values of the log likelihood function are shown in Table 6. Fig. 2 illustrates the relationship between the percentage of flies with ingested parasites and the mean worm burden in the skin of the human host as well as the lines fitted according to these models. The percentage of flies with ingested mff increased with host skin density following the very non-linear pattern already described for the relationship between prevalence and intensity in other parasitic infections (Fig. 2 A.). The small values of k confirm the high degree of overdispersion present in the microfilarial counts from gut squashes (see literature cited in Table 1 and also Anderson (19826), Anderson & May (1985) and Wenk (1991)). The estimates of the b coefficient and its confidence limits in the power model obtained from the log-log regressions on the directly measured mean intakes versus mean skin loads (Table 5), and from the fit to the prevalence versus intensity in the flies (Table 6) were found to be consistent. However, the likelihood analysis suggests that the linear model provides a better fit to the data points (likelihood ratio statistic = 22-1308, P < 00005 with 1 D.F., see also Fig. 2B). DISCUSSION

0-5 1-0 1-5 20 2-5 Log (mean no. of skin mff/mg + 1)

30

Fig. 1. (A) Scatter plot of the log transformed mean microfilarial counts ( + 1) for Simulium ochraceum s.l. in the Guatemalan data sets. The straight line is the regression line computed with Bartlett's three-group method: logCy + l) = a' + 6.1og(x+ 1) excluding patient C7. Coefficient values are given in Table 5 and criteria for 'outliers' in the text. A, De Leon & Duke (1966); • . Collins et al. (1977); Q, Campbell et al. (1980); C7 excluded. (B) Scatter plot of the log ( + 1) transformed microfilarial counts for S. damnosum s.l. in the West and Central African savanna data sets. The Bartlett's regression line has been calculated excluding patient DDL Regression results are shown in Table 5. A, De Leon & Duke (1966); • , Philippon (1977); D, Boussinesq (1991); DD1 excluded. (C) Scatter plot of the log (+ 1) microfilarial counts for S. guianense in the South Venezuelan data set. Coefficients of the regression line (computed with Bartlett's procedure) are as in Table 5. D, This work.

In all three data sets (S. ochraceum s.l. from Guatemala, S. damnosum s.l. from savanna localities in West and Central Africa, and S. guianense from South Venezuela) there was a positive and highly significant rank correlation between the mean number of mff/mg of skin in the subjects the flies fed upon and the mean number of mff ingested by these flies. A standard linear regression procedure was not possible for the analysis of the data sets because, on the one hand, the variance of the observed mean intakes increased with rises in the mean skin load and, on the other, because both are random variables. The most suitable variance stabilizing transformation was found to be log (mean + 1) on both variables, after which a regression line estimated by Bartlett's method was found to be an appropriate model to explore the functional relationship between the transformed data. According to the regression analysis, there seems to be little evidence of densitydependent regulation of the microfilarial intake by the insects with increasing skin loads, at least across

M. G. Basdiiez and others

122

Table 5. Results of linear regression and Spearman correlation analyses of the log (mean c o u n t + 1 ) transformation of ingested skin microfilariae Vector species (Guatemala)

S. guianense S. damnosum s.l. (West and Central (South Venezuela) African savanna)

25

33

9

DD1



0-1456 0-6663 (0-4671-0-8517)

0-0789 0-8525 (0-6481-10900)

0-7972

0-8553

0-8773 101795 00000

0-9500 8-0495 0-0000

S. ochraceum s.l.

Statistics n

'Outliers'* C7 Patient no. Regression Bartlett's method a 00432 b 0-9853 95% C.L. (b) (0-7738-1-2001) Least squares b 10112 Correlation r(S) 0-8992 t(n-2 D.F.) 9-8579 P 00000

* Criteria for 'outliers' described in the text. Table 6. Maximum likelihood estimates for the parameters in the function y(x) (y = Mean microfilarial intake/fly and x = mean microfilarial load/mg of skin. The estimations derive from data on the prevalence of flies with ingested mff vs mean intakes assuming two different models for the relationship between y(x) and x. Prevalence of infection in the flies = 1 —{1/[1 +y(x)/k]lc} (see text).) Parameters Vector species:

S. damnosum s.l. and

S. guianense

Model:

y(x) = a + bx

y(x) = xb

95 % C.L. (a) b 95% C.L. {b) k 95% C.L. (k) Log likelihood

0-2893 (01899-0-3970) 0-4481 (0-3234-0-6226) 0-6323 (0-5329-0-7597) -2129-9338*'1'

0-76O2 (O-6550-O-8779) 0-5876 (0-5016-0-6971) -2140-9991*12'

* Likelihood ratio statistic = 2 [(l)-(2)] = 221306, P < 00005 when compared to a chi-squared distribution with 1 D.F. the data sets and ranges of dermal parasite burdens explored in this study. Duke (19626), working in a forest area of West Cameroon, found that at very high skin levels, ^burnt-out' cases, with severe cutaneous pathology and thickening, were less able to act as efficient reservoirs of mff available to the flies, producing a plateau or even a decrease in the relationship between ingested and skin mff (see also Kershaw et al. 1954, 1956). These studies were the basis of previous claims about the existence of density dependence in this stage of the incorporation of the parasites by the vector host (Dietz, 1982).

Duke's data, however, were not included in this analysis because his work was carried out with other species of the damnosum complex present in the forested zones of Cameroon (likely to have been 5 . mengense and 5 . squamosum, Boussinesq (1991); Crosskey (1990)). The microfilarial intakes by S. soubrense and S\ sanctipauli (both forest members) in areas of West Africa (Cote d'lvoire) were also less than those by their savanna counterparts (Prod'hon et al. 1986, 1987; Boussinesq, 1991). Analyses of these latter data, not presented in this paper, suggest that the intake levels may indeed reach a plateau

Onchocerca volvulus microfilarial intake by simuliid vectors

100

200 300 400 500 600 Mean no. of skin mff/mg

0-1

1 10 100 Mean no. of skin mff/mg

700

800

1000

Fig. 2. (A) Percentage of flies with ingested mff versus mean dermal intensities for the African and Venezuelan species. The lines correspond to the maximum likelihood estimation of the relationship between prevalence (P) and intensity (y) in the flies, assuming that this relationship has the form P = 1 —{1/[1 +y(x)/k]lc} with constant k. All data points have ) Linear relation between sample sizes n > 20. ( mean intake (y) and mean skin burden (x): y(x) = a + bx, ) power model: y(x) = xb. The parameter values ( are as given in Table 6. • , Boussinesq (1991); • , Philippon (1977); A, this work. (B) Percentage of savanna Simulium damnosum s.l. and S. guianense specimens with ingested mff plotted against the logarithm of the mean skin load in the human reservoir. Lines are as described in (A) and in Table 6. The log scale permits a better appreciation of the goodness of fit of the different models at lower skin densities. Deviance (log likelihood ratio) as in Table 6. • , Boussinesq (1991); • , Philippon (1977); A, this work. Empty and filled squares: 5. damnosum s.l.; filled triangle: S. guianense.

at high dermal burdens in the forest zones. A detailed discussion of the origin of such differences between West African 'savanna' and 'forest' Onchocerca-Simulium complexes is not presented here, but differences in the clinical manifestations of the disease (more ocular involvement in the savanna in contrast with more cutaneous pathology in the forest), and in the depth at which the parasites are situated in the skin as the individual mean intensity

123

changes (deeper at heavier loads in the forest; Bain et al. (1986); Vuong et al. (1988)), are possible explanations (Boussinesq, 1991). According to WHO (1987), onchocerciasis in the savanna zones of Cote d'lvoire and Cameroon is characterized by higher skin microfilarial concentrations and more dermal atrophy than in the forest. However, the participants in Boussinesq's study very rarely presented with signs of pachydermy. The case for the lack of strong density dependence in the relationship between ingested and dermal parasites is supported by the results of the alternative analysis on data of prevalence of mff in the flies versus intensity in the skin of the human reservoir. Not only did this analysis provide very similar parameter estimates and confidence limits to those obtained from the regression methods applied to the direct mean counts, but also it suggested that the relationship between dermal burden and microfilarial intake could be better described by a linear rather than by a power model. Such a discrepancy may be partly due to the small size of some samples used for the prevalence versus intensity approach. This sort of analysis relies on accurate estimates of both prevalence and intensity in the field, which have been shown to depend on sample size and the degree of aggregation; the more overdispersed the distribution of the counts of parasites/host, the greater the sample size required to estimate reliably these two indices (Medley et al. 1993; Billingsley et al. 1994). The estimated values of k were less than 1, indicating a high degree of overdispersion (Anderson & May, 1985) and an agreement with direct observations of the distribution of the numbers of ingested mff/fly (data not shown). The interpretation of the apparent absence of density dependence in the relationship between parasite uptake by the vector and parasite load in the human host may be complicated by the presence of parasite aggregation in both host species. Aggregated patterns of animal or plant abundance can act to increase the efficacy of regulatory constraints on the total population since they act most severely in densely populated habitats (Anderson & Gordon, 1982; Anderson et al. 1982; Rosewell, Shorrocks & Edwards, 1990). In the context of O. volvulus transmission, the densely populated patches are the heavily infected individuals in the human or vector populations. Subtle changes in skin structure in heavily infected people could act to decrease the efficiency of transmission contributed from the small fraction of the population with heavy burdens. In the case of the vector, aggregated distribution could imply that parasite-induced vector mortality related to worm load (Duke, 1962 a; De Leon & Duke, 1966; Omar & Garms, 1977; Takaoka et al. 1984) would act on a small proportion of the total vector population. Although the results recorded here do not support the former conjecture, the latter may be

M. G. Basdnez and others

an important regulatory constraint. It will be examined in a future publication on the impact of parasite aggregation within the vector population. The phenomenon of so-called 'microfilarial concentration', in which flies have been reported to ingest more parasites than those apparently available as judged by skin snip counts, has been ascribed to the possible existence of chemo-attractant factors in the saliva of the insects gathering mff in the site of the bite (Crosskey, 1990). Concentration has been claimed to explain the higher intakes of mff by S. ochraceum s.l. and other Guatemalan species when compared with S. damnosum s.l. for similar skin loads of the homologous parasite strain in the human host (Dalmat, 1955; De Leon & Duke, 1966). This phenomenon has also been observed in species of the S. amazonicum group in the Brazilian Amazonas (Shelley et al. (1979), most likely to be S. oyapockense s.l./roraimense, Shelley (1988)). However, these and other studies (Omar & Garms, 1975 and those cited in Table 1) have relied on dermic microfilarial counts performed on the basis of variable numbers of skin snips, various incubation media and different incubation times, the latter generally being considerably shorter than the recommended minimum period of 8 h for enhanced skin snip sensitivity (Collins et al. 1980; WHO, 1987). The results presented here show that after standardizing the data for assessed skin burdens to a 24 h incubation period the phenomenon of microfilarial concentration by S. ochraceum s.l. becomes less evident, although the Guatemalan species does seem to ingest more mff than S. damnosum s.l. for similar levels of mff in the skin of the human reservoir. It is interesting to note that although S1. ochraceum s.l. and S. oyapockense s.l. may ingest comparatively more mff, they have in common the possession of a well-developed cibarial armature which damages a high proportion of the larvae (Omar & Garms, 1975; Shelley et al. 1987) and considerably diminishes the numbers effectively available for further development within the fly. In all considerations of the intake of skin-dwelling microfilariae by dipteran vectors it must be remembered that skin snip counts are a crude way of estimating the true density of parasite transmission stages in the human host. A much more direct assessment of the relationship between parasite density in the human host and fly intake of larvae is possible for the filarial worms with blood-dwelling mff (lymphatic filariases and mansonelliasis ozzardi). Here it is possible to get accurate estimates of numbers of mff per bloodmeal volume ingested by the insects and relate them to parasite densities per corresponding volumes of peripheral or venous blood from the vertebrate hosts (Bryan & Southgate, 1988). In the case of skin-dwelling mff, the mechanism by which they are gathered into the wound produced by the bite is not well understood. It is assumed that the depth at which the buccal apparatus

124

of Simulium spp. penetrates the skin determines the stratum of mff available for transmission. Although this depth can be remarkable considering their short proboscis (0-4 mm in the case of S. damnosum s.l., Wenk (1981)), it still will be limited to the superficial parts of the subepidermal layer (Kershaw et al. 1956; Kershaw, 1958), but it is as yet unclear whether mff come from adjacent or more distant locations, whether they are passively dragged in by the capillary blood irrigating the particular area of skin or actively recruited, and what is the relationship between the blood volume in the well of blood created by the mouth-parts of the pool-feeder blackflies and the surface/weight of skin at the site of the bite, etc. In conclusion, within the range of parasite loads in the skin of infected patients recorded in the three onchocerciasis endemic areas explored in this study, the rate of parasite acquisition by the simuliid vector does not appear to decline significantly at high microfilardermic intensities. Transmission success to the fly seems to be roughly proportional to the density of transmission stages of the parasite in the skin of the human host. However, once the larvae are ingested many other factors could act to reduce the fraction which actually reach the infective L3 stage as parasite density within the host rises. One of them is the migration of the mff out of the bloodmeal towards the thoracic muscles of the insect (or ' LI uptake', Plaisier et al. 1990, 19916). This process, plus the interplay between parasite aggregation in the dipteran host and the rate of parasite-induced vector mortality, will be examined in subsequent publications. M.G.B. acknowledges CONICIT of Venezuela (Consejo Nacional de Investigaciones Cienti'ficas y Tecnologicas), grant No. SI-1473 for financing the field work carried out among the Yanomami, and the British Council/FCO for a research training scholarship at Imperial College. M.B. is indebted to the Special Programme UNDP/World Bank/ WHO for Research and Training in Tropical Diseases (TDR/FIL Contract No. 870336) for sponsoring data collection in North Cameroon. G. F.M. is a Royal Society University Research Fellow. R. M. A. thanks the Wellcome Trust for financial support. Dr W. Sawyer helped with the statistical analyses, Drs D. A. P. Bundy, H. Townson, and A. J. Fulford, as well as Mr J. Williams, made helpful comments. Finally, the invaluable assistance of M. Bolivar, J. A. Gomez, I. Narbaiza and V. Park in many aspects of the field work at CAICET in Venezuela, and J. M. Prud'hom and P. Enyong in Cameroon, is greatly appreciated. REFERENCED

(1982a). The population dynamics and control of hookworm and roundworm infections. In Population Dynamics of Infectious Diseases (ed. Anderson, R. M.), pp. 67—108. London: Chapman and Hall. ANDERSON, R. M. (19826). Epidemiology. In Modern Parasitology (ed. Cox, F. E. G.), pp. 204-51. Oxford: Blackwell Scientific Publications. ANDERSON, R. M.

Onchocerca volvulus microfilarial intake by simuliid vectors ANDERSON, R. M. (1987). Determinants of infection in human schistosomiasis. In Bailliere's Clinical Tropical Medicine and Communicable Diseases, Vol. 2, No. 2, (ed. Mahmoud, A. F.), 2nd Edn, pp. 279-300. London: Bailliere Tindall. ANDERSON, R. M. & GORDON, D. M. (1982). Processes influencing the distribution of parasite numbers within host populations with special emphasis on parasite-induced host mortalities. Parasitology 85, 373-98. ANDERSON, R. M., GORDON, D. M., CRAWLEY, M. J. &

HASELL, M. p. (1982). Variability in the abundance of animal and plant species. Nature, London 296, 245-8. ANDERSON, R. M. & MAY, R. M. (1985). Helminth infections of humans: mathematical models, population dynamics, and control. Advances in Parasitology 24, 1-101. ANDERSON, R. M. & MAY, R. M. (1991). Infectious Diseases of Humans. Oxford: Oxford University Press. ARMITAGE, p. & BERRY, G. (1987). Statistical Methods in Medical Research, 2nd edn. Oxford: Blackwell Scientific Publications. BAILEY, N. T. j . (1981). Statistical Methods in Biology, 2nd edn. London: Hodder and Stoughton. BAIN, O., VUONG, P . , PETIT, G., PROD'HON, J., RANQUE, P.

& CHABAUD, A. G. (1986). Differences dans la localisation des microfilaires d'O. volvulus en savane et en foret: Implications cliniques eventuelles. Annales de Parasitologie Humaine et Compare'e 61, 125-6. BASANEZ, M. G. & YARZABAL, L. (1988). Onchocerciasis in the Sierra Parima and Upper Orinoco regions, Federal Territory of Amazonas, Venezuela. In Parasitic Diseases : Treatment and Control (ed. Miller, M. J. & Love, E. J.), pp. 231-56. Boca Raton, FL: CRC Press.

125

parasite system. In Theoretical Ecology (ed. May, R. M.), pp. 237-56. Oxford: Blackwell Scientific Publications. COLLINS, R. C , CAMPBELL, C. C , W I L T O N , D. P . & NEWTON,

L. (1977). Quantitative aspects of the infection of Simulium ochraceum by Onchocerca volvulus. Tropenmedizin und Parasitologie 28, 235-43. COLLINS, R. C , BRANDLING-BENNETT, A. D., HOLLMAN, R. B., CAMPBELL, C. C. & DARSIE, R. F. ( 1 9 8 0 ) .

Parasitological diagnosis of onchocerciasis: comparison of incubation media and incubation times for skin snips. American Journal of Tropical Medicine and Hygiene 29, 3 5 ^ 1 . cox, D. R. & HINKLEY, D. v. (1974). Theoretical Statistics. London: Chapman and Hall. cox, D. R. & OAKES, D. (1984). Analysis of Survival Data. London: Chapman and Hall. CROSSKEY, R. w. (1990). The Natural History of Blackflies. Chichester: John Wiley & Sons. CUPP, E. w. (1992). Treatment of onchocerciasis with Ivermectin in Central America. Parasitology Today 8, 212-14. DALMAT, H. T. (1955). The Blackflies (Diptera, Simuliidae) of Guatemala and their Role as Vectors of Onchocerciasis. Smithsonian Miscellaneous Collection 125. Washington: Smithsonian Institute. DAVIES, ] . B., WEIDHAAS, D. E. & HAILE, D. G. ( 1 9 8 7 ) .

Models as aids to understanding onchocerciasis. In Black-flies : Ecology, Population Management and Annotated World List (ed. Kim, K. C. & Merrit, R. W.), pp. 396^07. University Park and London: The Pennsylvania State University. DE LEON, j . & DUKE, B. o. L. (1966). Experimental studies

on the transmission of Guatemalan and West African strains of Onchocerca volvulus by Simulium ochraceum, S. metallicum and S. callidum. Transactions of the BASANEZ, M. G., YARZABAL, L., TAKAOKA, H., SUZUKI, H., Royal Society of Tropical Medicine and Hygiene 60, NODA, s. & TADA, I. (1988). The vectorial role of 735-52. several blackfly species (Diptera: Simuliidae) in DIETZ, K. (1982). The population dynamics of relation to human onchocerciasis in the Sierra Parima onchocerciasis. In Population Dynamics of Infectious and Upper Orinoco regions of Venezuela. Annals of Diseases (ed. Anderson, R. M.), pp. 209-41. London: Tropical Medicine and Parasitology 82, 597-611. Chapman and Hall. BILLINGSLEY, P . F., MEDLEY, G. F . , CHARLWOOD, ] . D. & DUKE, B. o. L. (1962a). Studies on factors influencing the SINDEN, R. E. (1994). Relationship between prevalence transmission of onchocerciasis. I. The survival rate of and intensity of Plasmodium falciparum infection in Simulium damnosum under laboratory conditions and natural populations of Anopheles mosquitoes. the effect upon it of Onchocerca volvulus. Annals of American Journal of Tropical Medicine and Hygiene (in Tropical Medicine and Parasitology 56, 130—5. the Press). DUKE, B. o. L. (19626). Studies on factors influencing the BOUSSINESQ, M. (1991). Etude epidemiologique de transmission of onchocerciasis. II. The intake of l'onchocercose en zone de savane camerounaise. Onchocerca volvulus microfilariae by Simulium Effets d'un traitement de masse par l'ivermectine. damnosum and the survival of the parasites in the fly Ph.D. thesis, University of Montpellier II. under laboratory conditions. Annals of Tropical BRYAN, j . H. & SOUTHGATE, B. A. (1988). Factors affecting Medicine and Parasitology 56, 255-63. transmission of Wuchereria bancrofti by anopheline ELLIOT, j . M. (1977). Some Methods for the Statistical mosquitoes. 1. Uptake of microfilariae. Transactions of Analysis of Samples of Benthic Invertebrates, 2nd edn. the Royal Society of Tropical Medicine and Hygiene 82, Freshwater Biological Association, Sci. Publ. No. 25. 128-37. Cumbria: Titus Wilson & Son Ltd. CAMPBELL, C. C , COLLINS, R. C , HUONG, A. Y. & FIGUEROA GRENFELL, B. T. & MICHAEL, E. (1992). Infection and MARROQUIN, H. (1980). Quantitative aspects of the disease in lymphatic filariasis: an epidemiological infection of Simulium ochraceum by Onchocerca approach. Parasitology 104, S81-S90. volvulus: the relation of skin microfilarial density to GUDERIAN, R. H., BECK, B. J., STONE, D. J . , ISABEL, K. & vector infection. Tropenmedizin und Parasitologie 31, MACKENZIE, c. D. (1988). Onchocerciasis in Ecuador: 475-8. recent observations in the province of Esmeraldas. COHEN, j . E. (1976). Schistosomiasis: a human-host Journal of Tropical Medicine and Hygiene 91, 161-8.

M. G. Basdnez and others GUYATT, H. L., BUNDY, D. A. P., MEDLEY, G. F. & GRENFELL,

B. T. (1990). The relationship between the frequency distribution of Ascaris lumbricoides and the prevalence and intensity of infection in human communities. Parasitology 101, 139-43. HAIRSTON, N. G. & JACHOWSKI, L. A. (1968). Analysis of the Wuchereria bancrofti population in the people of American Samoa. Bulletin of the World Health Organization 38, 29-59. KERSHAW, w. E. (1958). The population dynamics of infection with Onchocerca volvulus in the vector Simulium damnosum. Proceedings of the Xth International Congress of Entomology, Montreal, 1956 3, 499-501. KERSHAW, W. E., DUKE, B. O. L. & BUDDEN, F. H. ( 1 9 5 4 ) .

Distribution of microfilariae of O. volvulus in the skin. Its relation to the skin changes and to eye lesions and blindness. British Medical Journal 2, 724-9. KERSHAW, W. E., JAMISON, D. G., NUGENT, D. & DUKE, B. O.

L. (1956). Preliminary observations on the depth distribution of the microfilariae of Onchocerca volvulus in the skin and its relation to the reservoir of infection to the fly. Transactions of the Royal Society of Tropical Medicine and Hygiene 50, 6. LWAMBO, N . J. S., BUNDY, D. A. P . & MEDLEY, G. F. H.

(1992). A new approach to morbidity risk assessment in hookworm endemic communities. Epidemiology and Infection 108, 469-81. MACDONALD, G. (1965). The dynamics of helminth infections, with special reference to schistosomes. Transactions of the Royal Society of Tropical Medicine and Hygiene 59, 489-506. MAY, R. M. (1977). Togetherness among schistosomes: its effects on the dynamics of the infection. Mathematical Biosciences 35, 301-43. MEDLEY, G. F . , SINDEN, R. E., FLECK, S., BILLINGSLEY, P . F . , TIRAWANCHAI, N . & RODRIGUEZ, M. H. ( 1 9 9 3 ) .

Heterogeneity in patterns of malarial oocyst infections in the mosquito vector. Parasitology 106, 441—9. NELSON, G. s. (1991). Human onchocerciasis: notes on the history, the parasite and the life cycle. Annals of Tropical Medicine and Parasitology 85, 83-95. OMAR, M. s. & GARMS, R. (1975). The fate and migration of microfilariae of a Guatemalan strain of Onchocerca volvulus in Simulium ochraceum and S. metallicum, and the role of the buccopharyngeal armature in the destruction of microfilariae. Tropenmedizin und Parasitologie 26, 183-90. OMAR, M. s. & GARMS, R. (1977). Lethal damage to

Simulium metallicum following high intakes of Onchocerca volvulus microfilariae in Guatemala. Tropenmedizin und Parasitologie 26, 169-82. PHILIPPON, B. (1977). Etude de la transmission d'Onchocerca volvulus (Leuckart, 1893) (Nematoda, Onchocercidae) par Simulium damnosum Theobald, 1903 (Diptera, Simuliidae) en Afrique tropicale. Travaux et Documents de VORSTOM, (Paris) No. 63. PHILIPPON, B. & BAIN, o. (1972). Transmission de

l'onchocercose humaine en zone de savane d'Afrique Occidentale, passage des microfilaires A'Onchocerca volvulus Leuck. dans l'hemocele de la femelle de Simulium damnosum Th. Cahiers O.R.S.T.O.M., se'rie Entomologie me'dicale et Parasitologie 10, 251—61.

126 PLAISIER, A. P . , VAN OORTMARSSEN, G. J., HABBEMA, J. D. F . ,

REMME, j . & ALLEY, E. s. (1990). ONCHOSIM: a

model and computer simulation program for the transmission and control of onchocerciasis. Computer Methods and Programs in Biomedicine 31, 43-56. PLAISIER, A. P . , VAN OORTMARSSEN, G. J., REMME, J. & HABBEMA, j . D. F. (1991 a). The reproductive lifespan of Onchocerca volvulus in West African savana. Ada Tropica 48, 271-84. PLAISIER, A. P . , VAN OORTMARSSEN, G. J., REMME, J.,

E. s. & HABBEMA, j . D. F. (19916). The risk and dynamics of onchocerciasis recrudescence after cessation of vector control. Bulletin of the World Health Organization 69, 169-78. ALLEY,

PROD'HON, J., HEBRARD, G., PRUD'HOM, J. M. & COURET, D.

(1986). Action de l'ivermectine (MK-933) sur la transmission de l'onchocercose humaine par le groupe vecteur Simulium soubrense—Simulium sanctipauli en zone forestiere de Cote d'lvoire. Resultats immediatement et six mois apres traitement. Rapport OCCGE/ORSTOM non publie No. 03/IPR/Rap/86. PROD'HON, J., LARDEAUX, F . , BAIN, O., HEBRARD, G. & PRUD'HOM, J. M. (1987). Ivermectine et modalites de la reduction de l'infection des simulies dans un foyer forestier d'onchocercose humaine. Annales de Parasitologie Humaine et Compare'e 62, 590-8. PROST, A. & PROD'HON, J. (1978). Le diagnostique de l'onchocercose. Revue critique des methodes en usage. Medecine Tropical 38, 519-32. REMME, J., BA, O., DADZIE, K. Y. & KARAM, M. ( 1 9 8 6 ) . A

force-of-infection model for onchocerciasis and its applications in the epidemiological evaluation of the Onchocerciasis Control Programme in the Volta river basin area. Bulletin of the World Health Organization 64, 667-81. ROCHET, M.-J. (1990). A simple deterministic model for bancroftian filariasis transmission dynamics. Tropical Medicine and Parasitology 41, 225-33. ROSEWELL, J., SHORROCKS, B. & EDWARDS, K. ( 1 9 9 0 ) .

Competition on a divided and ephemeral resource: testing the assumptions. I. Aggregation. Journal of Animal Ecology 59, 977-1001. SCHLOTZHAUER, s. D. & LITELL, R. c. (1987). SAS System

for Elementary Statistical Analysis. Cary, NC: SAS Institute Inc. SHELLEY, A. j . (1988). Biosystematics and medical importance of the Simulium amazonicum group and the S. exiguum complex in Latin America. In Biosystematics of Haematophagous Insects (ed. Service, M . W.), pp. 203-20. The Systematics Association. Special Volume No. 37. Oxford: Clarendon Press. SHELLEY, A. j . (1991). Simuliidae and the transmission and control of human onchocerciasis in Latin America. Cadernos de Saude Ptiblica 7, 310-27. SHELLEYj A. J., PINGER, R. R., MORAES, M. A. P. & HAYES, J.

(1979). Concentration of microfilariae of Onchocerca volvulus by Simulium sanguineum during feeding: use in mapping parasite distribution in the skin. Journal of Medical Entomology 16, 48-51. SHELLEY, A. J., LUNA DIAS, A. P . A., MORAES, M. A. P . &

PROCUNIER, w. s. (1987). The status of Simulium oyapockense and S. limbatum as vectors of human onchocerciasis in Brazilian Amazonia. Medical and Veterinary Entomology 1, 219-34.

Onchocerca volvulus microfilarial intake by simuliid vectors SOKAL, R. R. & ROHLF, F. j . (1981). Biometry. The

Principles and Practice of Statistics in Biological Research. New York: W. H. Freeman and Company. TAKAOKA, H . , SUZUKI, H . , N O D A , S., TADA, I., BASANEZ, M .

G. & YARZABAL, L. (1984). Development of Onchocerca volvulus larvae in Simulium pintoi in the Amazonas region of Venezuela. American Journal of Tropical Medicine and Hygiene 33, 414—19. TAYLOR, H. R., KEYVAN-LARIJANI, E., NEWLAND, H. S.,

WHITE, A. T. & GREENE, B. M. (1987). Sensitivity of skin snips in the diagnosis of onchocerciasis. Tropical Medicine and Parasitology 38, 145-7. TAYLOR, H. R., MUNOZ, B., KEYVAN-LARIJANI, E. & GREENE,

B. M. (1989). Reliability of detection of microfilariae in skin snips in the diagnosis of onchocerciasis. American Journal of Tropical Medicine and Hygiene 41, 467-71. VAJIME, c. G. & DUNBAR, R. w. (1975). Chromosomal identification of eight species of the subgenus Edwardsellum near and including Simulium (Edwardsellum) damnosum Theobald (Diptera: Simuliidae). Tropenmedizin und Parasitologie 26, 111-38. VUONG, P . , BAIN, O., CABARET, J., PETIT, G., PROD'HON, J.,

P. & CHABAUD, A. G. (1988). Forest and savanna onchocerciasis: comparative morphometric histopathology of skin lesions. Tropical Medicine and Parasitology 39, 105-10. RANQUE,

127

WADA, Y. (1982). Theoretical approach to the epidemiology of onchocerciasis in Guatemala. Japanese Journal of Medical Science and Biology 35, 183-96. WENK, P. (1981). Bionomics of adult blackflies. In Blackflies : The Future for Biological Methods in Integrated Control (ed. Laird, M.), pp. 215—27. London: Academic Press. WENK, P. (1991). The vector host link in filariasis. Annals of Tropical Medicine and Parasitology 85, 139-47. WORLD HEALTH ORGANIZATION (1985). Ten years of onchocerciasis control in West Africa. WHO Document No. OCP/GVA/85-1B. WORLD HEALTH ORGANIZATION (1987). W H O Expert

Committee on Onchocerciasis. WHO Technical Report Series No. 752. Geneva. WILLIAMS, c. B. (1937). The use of logarithms in the interpretation of certain entomological problems. Annals of Applied Biology 24, 404-14. YARZABAL, L., BOTTO, C , ARANGO, M . , RAGA, L. M., WONG, F., ALLAN, R., JAIMES, J. L. & SANCHEZ-BEAUJON, R.

(1985). Epidemiological aspects of onchocerciasis in the Sierra Parima, Federal Territory of Amazonas, Venezuela. In La Oncocercosis en America (ed. Yarzabal, L., Botto, C. & Allan, R.), pp. 43-63. Caracas: Ediciones PROICET-Amazonas.

Suggest Documents