The effect of salt medicated with diethylcarbamazine in Bancroftian filariasis*

Bull. Org. mond. Sante 1969, 41, 195-208 Bull. Wld Hlth Org. The Effect of Salt Medicated with Diethylcarbamazine in Bancroftian Filariasis* A. DAVIS & D. R. BAILEY The paper describes a trial conducted in Tanzania of the effect on bancroftian micro- filaraemia ofcommon salt medicated with diethylcarbamazine at a 0.1 % (wlw) concentra- tion, when given to a closed population of 600-700 with a known salt intake. After good observer agreement in microfilarial counting had been demonstrated, quantitative blood surveys were performed monthly on known carriers. Pairedpretreatment microfilarial counts from the same patients at an interval of 4 months suggested a natural fluctuation in microfilarial densities and emphasized the needfor parallel concurrent controls during treatment. During the trial the mean microfilarial densities fell steadily, being reduced by 73 % at 3 months and 90 % after 6 months. Tolerance to the drug-salt mixture was extremely good. Problems ofplanning, salt preparation and assessment of results are discussed and the importance of correct experimental design in future trials is stressed. The authors suggest that a trial ofsalt containing 0.2 % (wlw) diethylcarbamazine may achieve optimum results. They point out that, while the use oJ diethylcarbamazine-medicated salt may prove useful in reducing transmission in closed communities or in endemic areas with no alternative source of salt, its extension to national populations as a sole means ofbancroftian filariasis control would be ofproblematical value. Success in the chemotherapeutic control ofendemic filariasis due to Wuchereria bancrofti has been limited. Diethylcarbamazine, the only effective filaricide in widespread use, gives rise to minor side- effects, and since many microfilarial carriers are symptom-free over long periods, it has been difficult to persuade carriers to undergo treatment (WHO Expert Committee on Filariasis, 1967). It is known that relatively small doses of diethyl- carbamazine, administered over long periods, have a microfilaricidal effect equivalent to the normal therapeutic regime with fewer adverse reactions (WHO Expert Committee on Filariasis, 1967). Experience with iodized salt in goitrogenic areas and with salt medicated with chloroquine for malarial suppression suggested the use of salt medicated with diethylcarbamazine in the control of microfilarial reservoirs. Hawking & Marques (1967) reviewed the theore- tical aspects of the procedure, showed that diethyl- carbamazine remained stable during cooking and * From the Bilharziasis Chemotherapy Centre, (World Health Organization/Medical Research Council of Great Britain/Tanzania), Tanga, Tanzania. presented the results of a pilot trial on 22 microfi- larial carriers given salt medicated with diethyl- carbamazine at varying concentrations up to 0.4% (w/w). Raghavan, Basu and Putatunda I discussed the results of a short-term trial in which 18 microfilarial carriers were given diethylcarba- mazine-medicated salt at a 0.1 % (w/w) concentration of drug. Preliminary conclusions from these two trials, both conducted on nocturnally periodic W. bancrofti, were that the medicated salt was acceptable up to a concentration of drug of 0.4% (w/w) and that, even at a 0.1% concentration, it diminished the microfilarial reservoir markedly. In 1968 this unit was requested to undertake a similar investigation. MATERIAL AND METHODS The microfilarial carriers were male prisoners at a large regional prison near Tanga, Tanzania, nor- 1 Unpublished working document WHO/FIL/68.82. A limited number of copies of this document is available to persons officially or professionally interested on request to Distribution and Sales, World Health Organization, 1211 Geneva, Switzerland. 2361 -195- A. DAVIS & D. R. BAILEY mally holding between 600 and 700 inmates. De- tainees were drawn from a wide area of East Africa and did not necessarily reflect either the incidence or the density of infection in the local population of Tanga. Meetings were held with the prisoners, the plan of the trial explained and their voluntary co- operation was obtained. Microfilarial surveys Microfilarial surveys were performed following the method of Sasa (1963) in which, between 21.00 and 23.00 hours, blood samples were withdrawn from a finger-tip and 3 linear smears each of 10 mm3 were prepared, representing a sampling unit of 30 mm3 of peripheral blood. Individual disposable presterilized lancets were used and the pipettes were cleansed, sterilized and dried with 0.1 % benzalkonium chloride and acetone after each examination. Slides were air-dried, dehaemoglobinized in tap- water, air-dried, fixed in methyl alcohol for one minute, and stained by warming with Mayer's acid haemalum for 5-8 minutes, after which excess stain was washed off. Two pretreatment microfilarial surveys at an interval of 4 months were performed on the prison population in order to familiarize personnel with methodology, to assess observer variation in micro- filarial counting, and to provide within-patient con- trols for the investigation of the stability of micro- filarial densities. After the second survey the total population of prisoners was given diethylcarbama- zine-medicated salt at a concentration of 0.1%I (w/w) for 6 months. Microfilarial surveys were conducted at monthly intervals on known microfilaria-positive cases during this time. In the seventh month a micro- filarial survey was performed on all those prisoners who had been admitted since the start of the trial and thus had received medicated salt for varying periods but whose blood had not previously been examined. Finally a microfilarial survey was conducted on the prison warders who fed outside the prison and there- fore had not received medicated salt. The microfilaria carriers in this group were treated with a therapeutic course of diethylcarbamazine at a total dose of 4.45 g given over 15 days, equivalent to 74 mg per kg for a 60-kg man. Salt supplies The bulk monthly supplies of salt for the prison (600 lb or about 272 kg) were delivered every 4 weeks to a central laboratory where all preparation of the drug-salt mixture took place. Diethylcarbamazine citrate powder was mixed by hand with the salt in a 0.1% concentration (w/w), 10 g of drug in 10 kg of salt, by sieving the drug in layers on to the salt in plastic buckets and by repeatedly transferring this drug-salt mixture from bucket to bucket. Finally the medicated salt was packed as units of 10 kg into plastic bags which were sealed to prevent leaching, which was known to occur with chloroquine-salt mixtures subjected to unsuitable storage conditions. Deliveries to the prison were made once weekly. The prison store received only sealed supplies of medicated salt and no other type of salt was held in stock. Fortnightly stores checks, both in the labora- tory and in the prison, ensured that ledger and actual physical stocks agreed. As a result of the strict indenting and stock-taking system adopted, it was extremely improbable that alternative supplies of salt could have been used. Medicated salt in the prison store was issued daily to the central kitchen, where all cooking was done, and each prisoner received his daily ration of salt from the kitchen at a dining parade. Before the trial the mean daily issue of salt from the prison store was 20 lb and hence a 10-kg bag (ca 22 lb) of medicated salt was adopted as a standard, daily unit for a popu- lation varying between 600 and 700 detainees whose daily ration of salt was 1/2 oz (ca 14 g) per head plus an indeterminate amount used in central cooking procedures. In practice a 10-kg bag was almost completely used up each day. Any remaining salt was carried forward for use on the next day. CHECKS ON MEDICATED SALT Replicated random samples of the drug-salt mixture were sent at intervals to the United King- dom where the diethylcarbamazine content was analysed. The range of the mean diethylcarbamazine content of each batch during the first 5 months was 0.09 %-0.15% and the over-all mean drug content of 0.117% indicated the efficiency of the mixing procedure. Assay failed to reveal the drug in samples taken in the last week of the trial, yet this only diminished the over-all mean diethylcarbamazine content to 0.098%. SALT CONSUMPTION AND DRUG INTAKE Diethylcarbamazine intake per head per day at a 0.1 % concentration of medicated salt and a ration of 14 g of salt would theoretically be 14 mg, which, over the 6-month period of the trial, gave a total intake of about 2.52 g or about 42 mg per kg of body-weight 196 DIETHYLCARBAMAZINE-MEDICATED SALT IN BANCROFTIAN FILARIASIS for a 60-kg individual. There must, however, be wide variation around this theoretical intake as many prisoners were engaged in hard physical labour such as sisal-cutting and presumably took as much salt as could be obtained, while for others engaged in less strenuous pursuits the theoretical intake figure could be regarded as a maximum. SOME DIFFICULTIES The population It was not originally appreciated that in the closed population of a prison there would still be consider- able population fluctuation due to short-term prisoners. For example, the first microfilarial survey of 672 prisoners yielded 148 positive cases (22 %), of whom only 33 remained in prison 4 months later at the time of the second pretreatment survey. Mean- while there had been 203 new admissions among whom the second survey showed 71 microfilaria car- riers (35 %), but the majority were short-term prison- ers and only 31 of this group and the 33 from the first survey remained in prison at the start of the trial. This factor disappointingly reduced the numbers of continuous total observations throughout the trial but provided a sample of paired control observations for the assessment of the stability of microfilarial densities. The salt The bulk salt supplies, as delivered to the central laboratory, consisted of large crystal aggregates, up to 1 cm in diameter, with an appreciable admixture of moisture and foreign particles, totally unsuitable for mixing with the drug. Four random samples, each of 50 g, from the first bulk consignment were weighed, and then reweighed at various times during drying in an oven. Water evaporation increased steadily up to 2 hours at a temperature of 60°C when the mean water content of the replicated samples was 2.295 g±SE 0.038. Expressed as a percentage, the mean water content of the crude salt was 4.59% with 95% confidence limits of 4.35% and 4.83 %. In view of this, all bulk salt supplies were dried, on trays in ovens, for 2 hours at a minimum temperature of 120°C, after which the grosser foreign particles were removed and the salt ground down to finely particulate size in a large hotel-type coffee-grinder. Diethylcarbamazine was then, mixed by hand as described. Drying, grinding, mixing, bagging and sealing the drug-salt mixture were arduous and time-consuming. When 4 men were available no more than 6 bags per day could be completed. For larger trials some mechanical procedure would be essential were the locally available salt not in a form suitable for direct mixing with the drug. MICROFILARIAL SURVEYS Tests of experimental technique Variation in microfilarial recovery in each line ofa 3-line blood smear. To be certain that microfilarial recoveries in each line of the 3 linear smears were comparable, the microfilarial yield in each line of the 148 positive slides from the first blood survey were ranked horizontally and subjected to a distribution- free analysis of variance (Friedman, 1937, 1940). The computed statistic Xr2 for rank totals was 1.47 on 2 degrees of freedom, 0.5 >P >0.3, indicating the sampling variation associated with homogeneous material. The null hypothesis that the microfilarial yield in each line of the 3 linear smears was similar was thus accepted. Observer variation in microfilarial counting. It is essential, before undertaking experiments in which replicated or repeated observations or both are made, to determine whether results are reproducible; in other words, whether adequate control is exercised over the experimental conditions. In experiments involving repetitive counting we have found from previous experience in schistoso- miasis that the first essential step is to determine whether any counter differs significantly in his assessments from others. If such variation exists it must be rectified before commencing the experiment. Four observers were responsible for counting microfilariae and their performance was assessed as follows. From the first blood survey 78 microfilaria- positive slides were randomly selected plus 2 known negative slides. The 80 slides were further randomized into 4 groups of 20. The 4 groups were randomly allotted to the 4 observers, who counted and recorded the number of microfilariae on each slide. Over 4 days each observer counted each of the 20 slides in each group. The exercise was blind as the slides were known by a reference number only and there was no communication between counters. All results were tabulated separately by each observer and handed to the experiment controller for analysis. As the true microfilarial count in any slide was unknown, the best estimate was themean microfilarial count of 4 experienced observers. Any significant deviation from this mean would indicate significant observer variation. The null hypothesis was that 197 198 A. DAVIS & D. R. BAILEY significant deviation from the mean did not exist for any of the 4 observers, and the alternative hypothesis was that significant deviation indicated significant observer variation in microfilarial counts. As the X2 distribution is inappropriate for small expected numbers, 19 of the 80 slides, including the 2 negative slides, were analysed separately because their mean expectation, i.e., the mean expected number of microfilariae, was less than 5. These counts were subjected to an analysis of variance by ranks (Friedman, 1937, 1940), and the computed test statistic Xr2 for this subgroup was 4.453 on 3 degrees of freedom, 0.4 >P >0.3, indicating acceptance of the null hypothesis and insignificant variation between observers' assessments of low microfilarial counts. As a point of interest, no observer found a microfi- laria in either of the negative slides. For the remaining 61 slides where the mean expectation was greater than 5, X72 was 5.316, on 3 degrees of freedom, 0.2 >P >0.1, indicating again non-significance of the variation between observers. Combining the groups yielded a X72 of 5.606 on 3 degrees of freedom, 0.2 >P >0.1, an over-all non- significant finding. Excellent agreement between observers' assess- ments of a wide range of microfilarial counts was further confirmed by a more penetrating partitioned X2 analysis of slides in which the mean expectation was greater than 5.1 It was concluded that the methods and observers used in the trial would carry no risk of significant variation. Nomenclature of microfilarial surveys Survey A was the first pretreatment survey, per- formed in February 1968 on 672 prisoners, yielding 148 microfilarial carriers. Survey B was the second pretreatment survey, per- formed 4 months later in June 1968 on the new admissions to the prison since Survey A. Of 203 new prisoners, 71 were microfilaria-positive but only 31 of these remained in prison at the start of the trial. Survey C (C = controls) referred to the 33 micro- filaria-positive cases of the original 148 in Survey A who were still detained 4 months later at the time of the second pretreatment survey, who were re- examined at that time, and who remained in prison at the start of the trial. (Thus 64 microfilaria-positive cases entered the trial, 33 from Surveys A and C, and 31 from Survey B.) 1 The details of the individual observers' counts and the analysis are available to interested readers from the authors. Survey D was the survey of those microfilaria- positive cases remaining in prison after 1 month of medicated-salt administration (62 patients). Surveys E, F, G, H, and I were surveys performed on the microfilaria-positive cases remaining in prison after 2, 3, 4, 5 and 6 months of medicated-salt administration and represented 52, 47, 32, 28 and 22 patients respectively. Survey J was a terminal survey of 193 prisoners admitted to *prison after the commencement of medicated-salt feeding to the prison population. Survey- W was a survey of 123 prison warders who did not receive medicated salt. Controls As this trial was to be conducted over a 6-month period, it was necessary to determine whether any variability occurred in microfilarial densities in the same subjects over a similar time period in the absence of treatment, i.e., the natural history of microfilaraemia. The two pretreatment blood surveys afforded 33 patients who had remained in prison in the inter- vening 4 months. These subjects were subclassified as Survey C (C = controls) and were re-examined at the same time as Survey B. Thus, 33 paired microfilarial counts, performed by the same method at the same time of night, by technicians in whom there was evidence of excellent agreement in microfi- larial assessments, were available for analysis. It was known that no treatment for filariasis had been given to any prisoner in this interim period. For each pair of microfilarial counts, the second count was subtracted from the first, the null hypo- thesis being that the mean of the sample of differ- ences did not differ from zero. Analysed by Wil- coxon's signed ranks test z= 2.5991, giving a 2-tail probability of 0.0096. Mean microfilarial count, Survey A = 50.76 per 30 mm3 blood 51 (rounded) Mean microfilarial count, Survey C = 37.21 per 30 mm3 blood 37 (rounded) Difference A-C = 13.55 per 30 mm3 blood 14 (rounded) This difference of 14 microfilariae per 30 mm3 blood between the means of the 2 surveys is statis- tically significant at a 2-tail probability level of 9 in 1000 and it represents a 27.5% reduction in mean microfilarial count over a 4-month period without treatment, suggesting a natural fluctuation in the stability of microfilarial densities. DIETHYLCARBAMAZINE-MEDICATED SALT IN BANCROFTIAN FILARIASIS Hawking & Marques (1967) stated that "the relative stability of microfilarial parasitaemias is well known ", but the literature is scanty on this aspect of bancroftian microfilaraemia. Edgar, Beye & Mille (1952) in Tahiti demonstrated a significant seasonal variation in microfilarial counts repeated on the same subjects during one year. Friedheim (1962), in the course of a clinical trial in southern Tanzania, followed microfilarial densities over 6 weeks in 5 patients with W. bancrofti infection who were given an inactive placebo, and noted that in 4 patients the microfilarial count after 6 weeks exceeded the initial count. Analysis of his data by a X2 dispersion test indicates a highly significant departure from homogeneity. Hairston & Jachowski (1968) showed a wave-like progression of microfi- larial counts when repeated on the same subjects in American Samoa with a peak "preceded by an increase in microfilarial density lasting from 3 to 24 months. . . followed by a decrease of somewhat shorter duration ". These observations have an important bearing upon the assessment of the results of medicated-salt trials which are, of necessity, conducted over long periods. There will always be some variability asso- ciated with microfilarial counts. Hairston & Jachow- ski (1968) discussed the sampling error associated with such counts and drew attention to the many variables which may influence the number of micro- filariae per unit volume of blood, e.g., site and depth of skin incision, amount of blood taken, pressure exerted, exact source of blood in the capillary bed, etc. To these must be added variation in observer's assessment of a count and natural fluctuations in microfilarial densities. Precise assessment of the effect of medicated salt on microfilaraemia is possible only if non-treated concurrent controls are incorporated in the design of a trial in order to demonstrate any natural fluctua- tions in microfilarial densities which may exist. RESULTS Results of 2 pretreatment microfilarial surveys Tables 1 and 2 indicate the frequency distribution of microfilarial densities in the prison population at the 2 pretreatment Surveys A and B. Point-pre- valence estimates of microfilaraemia were 148/672 (22%) and 71/203 (35 %). Microfilariae of Acantho- cheilonema perstans were seen on only 2 occasions in 1224 pretreatment blood films (Surveys A, B, C, J, W), one being a lone infection and the other mixed with W. bancrofti. TABLE I CUMULATIVE FREQUENCIES OF MICROFILARIA-POSITIVE CASES CLASSIFIED BY MICROFILARIAL DENSITY PER 30-mm3 BLOOD SAMPLE AT PRETREATMENT SURVEY A Microfilarial Cumulative frequency count Frqec ~N 1 12 12 8.1 2 8 20 13.5 3 4 24 16.2 4 3 27 18.2 5 6 33 22.3 6 1 34 23.0 7 3 37 25.0 8 2 39 26.4 9 2 41 27.7 10 1 42 28.4 11-20 12 54 36.5 21-30 19 73 39.5 31-40 11 84 56.8 41-50 6 90 60.8 51-0 6 96 64.9 61-70 5 101 68.2 71-80 5 106 71.6 81-90 1 107 72.3 91-100 3 110 74.3 101-200 28 138 93.2 201-300 7 145 97.9 >300 3 148 100.0 Total screened = 672. Total microfilaria-positive = 148 (22.0 %) Positive I line only = 16 (10.8_%) Positive 2 lines only = 12 (8.1 %) Positive 3 lines = 120 (81.1 %) No. with positive Correction Medi grade a % positive factor microfilarial N3 N2 Ni 20mm3 |10MM density 120 12 16 22 0.964 0.901 32.5 a See WHO Expert Committee on Filariasis (1967), pp. 45-46. 199 A. DAVIS & D. R. BAILEY TABLE 2 CUMULATIVE FREQUENCIES OF MICROFILARIA-POSITIVE CASES CLASSIFIED BY MICROFILARIAL DENSITY PER 30-mm3 BLOOD SAMPLE AT PRETREATMENT SURVEY B Microfilarial F eCumulative frequency count F No. |_% 1 2 2 2.8 2 2 4 5.6 3 2 6 8.4 4 1 7 9.8 5 3 10 14.1 6 1 11 15.5 7 0 11 15.5 8 2 13 18.3 9 0 13 18.3 10 1 14 19.7 11-20 15 29 40.8 21-30 5 34 47.9 31-40 5 39 54.9 41-50 2 41 57.7 51-60 3 44 62.0 61-70 5 49 69.0 71-80 3 52 73.2 81-90 2 54 76.0 91-100 2 56 78.9 101-200 8 64 90.1 201-300 4 68 95.8 >300 3 71 100.0 Total screened = 203. Total microfilaria-positive = 71 (35 %) Positive I line only = 4 (5.6 %) Positive 2 lines only = 7 (9.8 %) Positive 3 lines = 60 (84.5 %) No. with positive Correction grade a % positive factor for 30 mm3 N3 N, Ni 20 MM3I|10 mm3 Median Results of microfilarial survey on new admissions during the trial Table 3 shows the frequency distribution of microfilarial densities in those prisoners admitted after the start of the trial who had received medicated salt for varying lengths of time, i.e., Survey J. In TABLE 3 CUMULATIVE FREQUENCIES OF MICROFILARIA-POSITIVE CASES CLASSIFIED BY MICROFILARIAL DENSITY PER 30-mm3 BLOOD SAMPLE AT SURVEY J Microfilarial Frequency Cumulative frequency count No.unc 1 2 2 7.4 2 4 6 22.2 3 2 8 29.6 4 1 9 33.3 5 1 10 37.0 6 2 12 44.4 7 0 12 44.4 8 0 12 44.4 9 0 12 44.4 10 0 12 44.4 11-20 5 17 63.0 21-30 2 19 70.4 31-40 2 21 77.8 41-50 2 23 85.2 51-60 1 24 88.9 61-70 1 25 92.6 71-80 0 25 92.6 81-90 1 26 96.3 91-100 0 26 96.3 101-200 1 27 100.0 Total screened = 193. Total microfilaria-positive = 27 (14 %) Positive 1 line only = 5 (18.5 %) Positive 2 lines only = 3 (11.1 %) Positive 3 lines = 19 (70.4 %) microriiariai No. with positive Correctiondensity grade % positive factor Midiar ___|__ _|_N. for 30m M3 micro3lars al N3 N2 Ni 20 mm3 110 mm' density 32.5 a See WHO Expert Committee on Filariasis (1967), pp. 45-46. 19 3 5 14 0.937 0.841 12 a See WHO Expert Committee on Filariasis (1967), pp. 45-46. 60 7 4 35 0.982 0.93 200 DIETHYLCARBAMAZINE-MEDICATED SALT IN BANCROFTIAN FILARIASIS TABLE 4 ARITHMETIC MEANS OF MONTHLY MICROFILARIAL COUNTS FOLLOWING ADMINISTRATION OF SALT MEDICATED WITH DIETHYLCARBAMAZINE AT 0.1 % (w/w) CONCENTRATION (COMBINED FIGURES FOR SERIES I AND 2 OF TABLE 5) Pre- Medicated-salt administration treatm ent (Surveys After After After After After After B & C) I month 2 months 3 months 4 months 5 months 6 months(Survey D) (Survey E) (Survey F) (Survey G) (Survey H) (Survey I) No. of microfilarial counts (N) 64 62 52 47 32 28 22 Mean microfilarial count of total N per 30 mm' blood (x) 44.3 26.18 22.5 11.96 5.5 6.96 4.64 Percentage reduction a of mean microfilarial count 40.9 49.2 73 87.6 84.3 89.5(total units In N used) Mean microfilarial count of paired observations In succeeding column (x,)b 44.18 28.69 20.89 9.16 5.68 5.96 i.e., 62 D at B & C (62 D) (52 E) (47 F) (32 G) (28 H) (221)Q~i2EatD47 F at E, etc. Percentage reduction c of mean microfilarial count on paired observa- tions only b 35.1 52.7 79.3 87.2 86.5 i.e., 52 E at D (52 E) (47 F) (32 G) (28 H) (221) 47 F at E, etc. No. (and proportion) of negative microfilarial counts 6/64 d 13/62 12/52 21/47 10/32 12/28 9/22 (0.094) (0.21) (0.231) (0.447) (0.312) (0.428) (0.409) a Calculated as 100 - ( x 100 where Z = post-treatment mean microfilarial count at Surveys D, E, F, G, H and 1, and Y= pretreatment mean microfilarial count at Survey B and C. Total units in N used. b Figures in parentheses in rows 4 and 5 denotethe number of paired observations available atthe succeeding month's survey. c As In a but calculated on paired observations in succeeding column. d Six microfilarial counts in Survey C had become negative Immediately before treatment but had been positive 4 months previously at Survey A comparison with Tables 1 and 2, both the point- prevalence of microfilaraemia of 27/193 (14%) and the median microfilarial density of 12 were reduced. Analysis of the incidence of microfilaraemia in Survey J in relation to the length of time spent in prison and hence the duration of medicated-salt administration revealed no linear trend of reduction although some sample sizes were small, thus giving a high degree of variability. Results of repeated microfilarial surveys during medicated salt administration The results of monthly microfilarial counts for 6 months during the administration of 0.1 % medi- cated salt are shown in Tables 4 and 5. Discharge of prisoners led to diminishing numbers throughout. Table 4 shows the arithmetic means of the combined Series 1 and 2 of Table 5. Series 1 and 2 were tabu- lated separately to note whether there were any striking differences in the findings at each month. The only difference between the series was that the 33 sub- jects in Series 1 had 2 microfilarial counts before commencing medicated salt, at Surveys A and C, while those in Series 2 had only 1 pretreatment count, at Survey B. As can be seen from Table 5 the results for each series were very similar, giving added con- fidence to the validity of the combined figures. Results ofmicrofilarial survey and diethylcarbamazine- treatment ofprison warders Of 123 warders surveyed, 14 were microfilaria- positive (9%). The mean density was 53.7 micro- 201 A. DAVIS & D. R. BAILEY TABLE 5 ARITHMETIC MEANS OF MONTHLY MICROFILARIAL COUNTS FOLLOWING ADMINISTRATION OF SALT MEDICATED WITH DIETHYLCARBAMAZINE AT 0.1 % (w/w) CONCENTRATION Series I Medicated-salt administration Pretreatment Series 2 and control Pretreat- After After After After After After surveys ment I month 2 months 3 months 4 months 5 months 6 months survey B (Survey (Survey (Survey (Survey (Survey (Survey A C D) E) F) G) H) I) Series I No. of microfilarial counts (N) 33 33 31 26 25 20 18 14 Mean microfilarial count of total N per 30 mm3 blood (x) 50.76 37.21 21.42 20.15 11.72 7.15 8.83 6.07 Percentage reduction a of mean microfilarial count(Total units N in used) 26.7 42.4 45.8 68.5 80.8 76.3 83.7 Mean microfilarial count of paired ob- servations in succeeding column (xb)b 36.52 22.77 20.68 12.95 7.28 8.07 - .e., 31 D at C (31 D) (26 E) (25 F) (20 G) (18 H) (141)26 E at D, etc. Percentage reduction c of mean microfilarial count on paired observations only b 37.7 43.3 65.5 80.8 77.9 i.e., 26 E at D (26 E) (25 F) (20 G) (18 H) (141)25 F at E, etc. No. (and proportion) of negative 0/33 6/33 d 6/31 5/26 12/25 5/20 6/18 5/14 microfilarial counts (0) (0.18) (0.19) (0.19) (0.48) (0.25) (0.33) (0.36) Series 2 No. of microfilarial counts (N) 31 31 26 22 12 10 8 Mean microfilarial count of total N (per 30 mm3 of blood (x) 51.84 30.94 24.85 12.23 2.75 3.6 2.12 Percentage reduction a of mean microfilarial count(Total units in N used) 40.3 52.1 76.4 94.7 93.1 95.9 Mean microfilarial count of paired observations in succeeding 51.84 34.62 21.14 2.83 2.8 2.25 - column (xl) (31 D) (26 E) (22 F) (12 G) (10 H) (81) Percentage reduction c of mean microfilarial count on paired observations only b 33.2 59.2 94.5 94.6 95.7 - (26 E) (22 F) (12 G) (10 H) (8 1) No. (and proportion) of negative microfilarial counts 0 7/31 7/26 9/22 5/12 6/10 4/8(0) (0.13) (0.27) (0.41) (0.42) (0.60) (0.50) a Calculated as 100 - ( x 100) where Z = post-treatment mean microfilarial count at Surveys D, E, F, G, H and 1, and Y = pretreatment mean microfilarial count at Survey B or C. Total units in N used. b Figures in parentheses in rows 4 and 5 in each series denote the number of paired observations available atthe succeeding month's survey. c As in a but calculated on paired observations in succeeding column. d Six microfilarial counts in Survey C had become negative immediately before treatment but had been positive 4 monthspreviously at Survey A. 202 DIETHYLCARBAMAZINE-MEDICATED SALT IN BANCROFTIAN FILARIASIS filariae per 30 mm3 of blood and the median density was 14.5. After treatment with a standard course of diethylcarbamazine, 9 of the 14 warders were micro- filaria-negative on follow-up and the mean count for the series dropped from 53.7 microfilariae to 4.4 per 30 mm3, a reduction of 91.8%. This reduction was artificially low as 1 patient, in whom a pretreatment count of 97 per 30 mm3 was reduced to 53 after treatment, was strongly suspected of defaulting on dosage. The 4 remaining microfilaria-positive patients had pretreatment counts of 322, 102, 96 and 53 microfilariae per 30 mm3 reduced to 3, 4, 1 and 3 microfilariae per 30 mm3 of blood respectively. Clinical Clinical manifestations of filariasis in the prison were known to be uncommon. It was not possible to screen physically all the microfilaria-positive cases revealed in the 2 pretreatment surveys, but of 47 microfilaria carriers who were randomly selected for a parallel investigation of a filarial skin antigen and clinically examined, 41 (87%) had no signs or suggestive history. In 6/47 (13 %) there were 4 hydro- coeles and 3 gave a reasonably definite history of chyluria. Elephantiasis was not seen. Tolerance in the prison population of 600-700 persons to the medicated salt given at a 0.1 % (w/w) concentration for 6 months was good. No complaint of any side-effects was received. This aspect was investigated particularly closely during the first 2 weeks of medicated-salt administration with the co-operation of the prison medical officer. In con- trast, among 14 prison warders treated with a total dose of 4.45 g of diethylcarbamazine in 15 days, equivalent to 74 mg per kg for a 60-kg man, plus an antihistamine compound for the first 4 days, there were 7 complaints in the first week. Five of these patients developed tender thickened spermatic cords, testicular pain or both, and headache and joint pains were recorded. Although the intensity of side-effects was only moderate, they caused grumbles in a highly disciplined body of men. Urine samples After medicated salt had been given for 5 months (about 150 days), 10 urine samples were taken 2 hours after the main meal of the day from a random selection of those prisoners known to be microfilaria- positive at some time during the trial. The samples were flown to the United Kingdom, where the diethylcarbamazine content was estimated at the National Institute for Medical Research. TABLE 6 DIETHYLCARBAMAZINE EXCRETION IN URINE AND CURRENT MICROFILARIAL COUNTS IN 10 SUBJECTS AFTER 150 DAYS' ADMINISTRATION OF 0.1 % DIETHYLCARBAMAZINE-MEDICATED SALT a Mg diethyl- TMicrofilarial countSubject carbamazine iloodper 100 ml urine per30 mm3 blood 1 10.2 7 2 4.0 0 3 7.1 0 4 1.9 12 5 1.4 40 6 2.9 15 7 0.8 18 8 2.9 0 9 8.0 25 10 2.1 0 Mean= 4.13 { Mean= 11.7 a Theoretical intake = 14 mg diethylcarbamazine per day = total theoretical intake of 2.1 g drug. The results of these estimations and the current microfilarial count of each of the prisoners from whom urine was collected are shown in Table 6. No correlation between the 2 variables was demons- trable, Spearman's rs giving a probability of over 30%. DISCUSSION Diethylcarbamazine-medicated salt at a concentra- tion of 0.1 % (w/w) of drug was acceptable and non- toxic when given daily to between 600 and 700 inmates of a closed population over a 6-month period. Urinary estimations of drug excretion showed that in fact the prisoners were receiving medicated salt at that time and assays of the drug content .of random samples of medicated salt confirmed that the mixing procedure was accurate in achieving the desired concentration. Repeated microfilarial surveys demonstrated that the compound was undoubtedly effective in reducing the total microfilarial reservoir of this sample although relatively ineffective in complete eradication of parasitaemia. 203 A. DAVIS & D. R. BAILEY There was a steady fall in mean microfflarial counts during administration of the drug-salt com- pound which levelled at 4-6 months. Expressed in another way, the percentage reduction of the mean microfilarial count of each survey rose from 40%- 500% at 1-2 months to around 80 %-90% after 4 months. From 20% to 45% of counts became negative at different times. With the exception of Series 2 at 3 months (Table 5), the agreement of mean counts and percentage reduction of the mean count (Tables 4 and 5), whether calculated on the total number of subjects available or on paired observa- tions of those present in the succeeding month, was striking at all times. This was interpreted as indica- ting that the losses at each month were a true repre- sentative fraction of the total frequency distribution of the available microfilarial counts and that a disproportionate number of outlying high or low values had not been lost, which might have distorted the general pattern of response. Perhaps the low proportion of negative counts was not surprising. One of the peculiarities of diethyl- carbamazine administered in various types of fila- riasis both in animals and in man is that although 90% of microfilariae in the blood disappear even after small doses, a few often persist, for reasons which are obscure (Hawking, 1963). Furthermore, mass treatment campaigns (e.g., Sasa, 1967), although highly successful in reducing the mean or median count of a microfilarial population by lowering the total frequency distribution of micro- filariae, left a small residue of microfilaria carriers with low counts. It might be unrealistic to expect a higher proportion of microfilaria-negative cases in this trial after a total dose of approximately 2.52 g diethylcarbamazine given over 180 days (= 42 mg per kg for a 60-kg individual) when the standard total dose of 72 mg per kg given in 12 separate doses pro- duced only 80%-90% negativity (Sasa, 1967). It can also be surmised that the proportion of microfilaria- negative cases will vary among different populations given the standard treatment, depending on varying microfilarial densities in those populations. The important finding was that a reduction of mean microfilarial density of some 50% had occurred in the sample after 2 months of a very low concentra- tion of drug and this reduction increased over the duration of the trial. The results of treatment of the warders with a standard therapeutic course of the same drug suggested a normal response to diethyl- carbamazine. A basic postulate in attempting the control of endemic filariasis is that the reduction of human microfilarial reservoirs will lead to diminished transmission. It has been theorized mathematically that in principle the critical densities of host and vector below which a parasitic population cannot maintain itself are most important for those diseases in which the parasite sexes are separate, as in the filariases (Hairston, 1962, 1965; Macdonald, 1965). There has recently been a mathematical demonstra- tion that the transmission of W. bancrofti from mosquito to human host is, in Burma at least, surprisingly functionally inefficient (Hairston & de Meillon, 1968). In comparing differences in the transmission of malaria and bancroftian filariasis in West Africa, Brengues et al. (1968) noted that man was the chief agent of spread of bancroftian filariasis in areas where villages were scattered, that the human carriers of W. bancrofti must remain in any new area for some considerable time to create a focus of infection, and that an important factor in restrict- ing spread of filariasis was the low infection rate among vectors. Added to these mutually compatible observations the production of low microfilarial reservoirs by non-toxic means thus assumes some importance in the mechanisms of control. Hence further trials on medicated-salt administration in bancroftian filariasis to give a more precise definition of advantages and limitations are desirable. A broad comparison between the results of this trial and the pilot trials of Hawking & Marques (1967) and Raghavan, Basu & Putatunda (op. cit.) reveal some discrepancies, and the comparative data are shown in Table 7. Calculations have been made on paired observations only, i.e., on data from those subjects presenting at the succeeding microfilarial survey. The point of similarity in all trials is the marked reduction of the mean microfilarial count in the samples, commencing soon after diethylcarbamazine- medicated salt administration. The dissimilarities, which involve the degree of reduction, are probably due to 3 factors: the small size of sample, with a consequent inherently greater variation; the different pretreatment mean microfilarial counts in the samples, even correcting for the different sampling volumes of blood; and the varying amounts of compound given and taken. For example, although the series from India and Tanzania showed respec- tively a 94.1 % and 52.7% reduction in mean micro- filarial count at 2 months after comparable amounts of compound, the Indian series is composed of a smaller sample with a lower mean pretreatment 204 DIETHYLCARBAMAZINE-MEDICATED SALT IN BANCROFTIAN FILARIASIS 0 0~~~~~~~~~~~~44,U) co c~0~o w~~~~~~~~~~0~ ~~~~~~~~~. 0 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ E 2 ~~~~~~~~~~0 ___ ~~.0 0- p - c C) 0o 0~~~~~~~~~~~~~~~~~~~~~~~~~ co 71Eo Ul)~~~~~~~~~~~~~~ 20~~~~~~~~~~~~~~~~~~~~~~~ C4 E~C 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 T 01C E0 ~ oj >) o)z EI 0 o .0 > ,) IX (D (D N~~01 0 4~~~~~~~~~~~~~~~~~~c0 CCc0 C 0o 0' r- CO c C 0 0 co~~~~~~~~~~~~~~~~~~~~~~~.> 4) U') C; ~~~~~~~~0110 ~2 0 -4) N~~~~~~~~i 0)0 c 2~~~~~~~~~~~~~~~0~ ~ ~ ~~~~~C 0 C- 6 1a -. cooE~~~~~~-E ' 0o 0 - a (31 c o~0 ~~~~~~~~~~~~~~~0 co c CC04 co 0 E E CC4 cm 0> C a.-0-- -oC~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~3..0CO 0 cct0 0 0L( C C 4- -t 0 c~~0o~ CD 0E2 E E cojc z E E E E C, E.E Q E CE L. 0 ~ 02- C--W~ 0 -- 0 205 A. DAVIS & D. R. BAILEY count. Again the high percentage reduction of mean count in the Brazilian Group A is in a small sample with a low pretreatment count and given a much higher dose of active compound. Allowing for variables, it may be concluded tentatively that the dissimilarities in results are not so great as would appear at first sight. In future trials of this form of control, the experi- mental design will prove of great importance. Many variables are involved: time, concentration of drug, frequency distributions of microfilariae and possibly different patterns of the stability of microfilarial densities. In addition, there will always be sampling variations in the physical techniques of blood collec- tions and observer variation in the assessment of numbers of microfilariae present on blood slides. The former can be minimized by constant attention to detail, and a system of incorporating check counts on slides showing high or low outlying values or on those distorting the general pattern of response will guard against undesirable degrees of observer variation. In this trial all monthly microfilarial counts were reviewed and any unusual readings were checked blind by different observers. The mean of these counts was taken as the most accurate estimate of the microfilarial density. Control observations will be essential to allow for variations in microfilarial density in time. There was a reduction of the mean microfilarial count of 27.5% in 4 months in pretreatment counts on paired obser- vations and similar fluctuations have been noted by other investigators in various areas. The influence of this natural fall in mean count on the therapeutic estimates of the efficiency of drug-salt after 1 and 2 months cannot be assessed, but after 3 months there was little doubt of the efficiency of the compound (Table 4). Of course, it may be argued that the con- trol sample itself was small and may have given a biased estimate of variation in stability of microfi- larial density. This only indicates the need for testing of further larger samples. In conclusion, 3 trials have shown, in small samples, that salt medicated with diethylcarbamazine in concentrations ranging from 0.1 % to 0.4% (w/w) was acceptable, non-toxic and effective in reducing a microfilarial reservoir. Further trials on populations with varying distributions of microfilarial densities would more precisely determine effects and limita- tions, provided strict experimental design, control and technique were incorporated. A trial of diethyl- carbamazine-medicated salt at 0.2% (w/w) concen- tration might achieve optimal microfilarial reduction combined with acceptability. The over-all place of diethylcarbamazine-medica- ted salt as a sole or predominant method of chemo- therapeutic control in filariasis is more problematical. It can be criticized on grounds of wide variation in individual salt consumption, the lack of control or knowledge of actual drug intake, the technical and practical difficulties of large-scale mixing and dis- tribution of the drug-salt compound, and, in the absence of national legislation, the availability of alternative non-medicated salt, an important point in human populations not attuned to preventive medicine. There may be doubts about the effects of an active drug, given in less than optimal dosage over relatively prolonged periods, on helminth resistance, although this is as yet undescribed. The long-continued admin- istration of a number of therapeutically effective drugs which are completely safe for short-term use may occasionally have undesirable side-effects in a few individuals. Since the chemotherapeutic control of endemic filariasis using medicated salt would involve large populations, it would be desirable to seek further information on long-term toxicity by continuous administration of diethylcarbamazine in salt to laboratory animals. Similarly, although experi- ments have shown that diethylcarbamazine-medicated saltremained stableduringcertain cookingprocedures this is no guarantee that the variety of cooking methods or the variety of foods, additives or preserv- atives that would be employed in mass use would not affect the stability of the drug. Especial care would be needed in areas where Wuchereria bancrofti and Onchocerca volvulus coexisted in the human population because of the unpleasant side-effects which may occur in patients withonchocerciasiswho are given diethylcarbamazine. It may be that diethylcarbamazine-medicated salt would play a useful role as an adjunct to other, more orthodox means of control and its use can be en- visaged in closed communities in endemic areas or in populations with no alternative sources of salt. Extrapolation to national populations may pose formidable difficulties. 206 DIETHYLCARBAMAZINE-MEDICATED SALT IN BANCROFTIAN FILARIASIS 207 ACKNOWLEDGEMENTS We thank Dr F. Hawking for arranging assays of the diethylcarbamazine content of salt samples through the kindness of the Wellcome Research Laboratories, for arranging diethylcarbamazine estimations in urine by courtesy of Dr Williamson and Mr Cover of the National Institute for Medical Research, United Kingdom, and for discussions during the trial. The drug was supplied as diethylcarbamazine citrate powder by the Wellcome Research Laboratories, Becken- ham, Kent, United Kingdom. We are grateful to Dr N. B. Akim, Chief Medical Officer, Ministry of Health and Housing, United Republic of Tanzania, to the Commissioner of Prisons, United Republic of Tanzania, and in particular to the Regional Senior Superintendent of Prisons, Mr A. B. Mwaijande, and his staff for helpful co-operation. RANSUM A ACTION DU SEL MEDICAMENTE PAR LA DIETHYLCARBAMAZINE DANS LA FILARIOSE DE BANCROFT L'essai decrit dans le present article, mend pendant 6 mois dans une prison de Tanzanie, visait a dvaluer l'efficacite de la diethylcarbamazine, administree sous forme de sel medicamente 'a la concentration de 0,1 % (p/p), dans la filariose a Wuchereria bancrofti periodique. Deux enquetes preliminaires avaient rdveld dans des groupes de 672 et 203 ddtenus des taux de prevalence de la microfilaremie de 22% et 35% respectivement. En raison des fluctuations de la population de la prison, seuls 64 porteurs de microfilaires etaient presents au moment oil l'essai a debute. Au cours des examens de sang pratiques avant et pendant ce demier, on s'est assure que les variations enregistrees lors de l'examen d'un meme echantillon de sang par plusieurs observateurs restaient dans des limites tres acceptables. Des numera- tions couplees effectuees chez 33 detenus a 4 mois d'in- tervalle, en dehors de tout traitement, ont par ailleurs mis en evidence une reduction spontanee de 27,5 % de la densite microfilarienne moyenne et indique la necessite de cons- tituer des groupes tdmoins pour des enquetes de ce genre. Le sel medicamente contenant 0,1 % de citrate de diethylcarbamazine a ete prepare dans un laboratoire central. Chaque dMtenu a requ quotidiennement 14 g de sel, correspondant a la prise theorique de 14 mg par jour et d'une dose totale de 2,52 g de diethylcarbamazine en 6 mois, soit 42 mg/kg pour un homme de 60 kg. La tolerance a ete tres bonne et aucune reaction secondaire n'a ete observee. Par contre, des plaintes ont ete for- mulees par 7 gardiens (sur 14) traites classiquement par la diethylcarbamazine a la dose totale de 4,45 g en 15 jours, soit 74 mg/kg pour un homme de 60 kg. Les numerations de microfilaires pratiquees mensuel- lement sur les 64 porteurs presents pendant la duree de l'essai ont montre une reduction des densites microfila- riennes moyennes de 40 %, 50 %, 73 %, 88 %, 84% et 90% respectivement apres 1, 2, 3, 4, 5 et 6 mois d'adminis- tration du sel medicamente. Les examens de sang sont devenus negatifs chez 9 des 14 gardiens traites par la methode classique. Apres 5 mois de traitement par le sel medicamente, on a determine la teneur en diethylcarbamazine de 10 echan- tillons d'urine preleves au hasard chez des prisonniers. Dans chaque cas, l'examen a ete positif, mais on n'a note aucune correlation entre les taux d'excretion et les densites microfilariennes. La recherche des microfilaires effectuee a la fin de 1'essai sur des detenus soumis au regime du sel medicamente pendant une duree variable a montre une reduction de la prevalence des porteurs (27 sur 193, soit 14%) et une chute des densites microfilariennes moyennes par rapport aux resultats fournis par les deux enquetes preliminaires. Les auteurs insistent sur la necessite de definir avec precision les conditions experimentales de semblables essais et de disposer de groupes temoins. I1 semble que I'administration de sel medicamente renfermant 0,2% de diethylcarbamazine permettrait d'obtenir une reduction optimale des densitds microfilariennes sans apparition de reactions secondaires notables. L'emploi du sel medica- mente semble une mesure d'appoint pleine de promesses, surtout pour le traitement de collectivit6s fermees, mais sa generalisation a 1'echelle d'un pays ne peut etre envi- sagee qu'avec une extreme circonspection. REFERENCES Brengues, J., Subra, R., Mouchet, J. & Nelson, G. S. (1968) Bull. Wld Hlth Org., 38, 595 Edgar, S. A., Beye, H. K. & Mille, R. (1952) Amer. J. trop. Med. Hyg., 1, 1009 Friedheim, E. A. H. (1962) Ann. trop. Med. Parasit., 56, 387 Friedman, M. (1937) J. Amer. statist. Ass., 32, 675 Friedman, M. (1940) Ann. Math. Statist., 11, 86 Hairston, N. G. (1962) Population ecology and epidemio- logical problems. In: Wolstenholme, G. E. W. & O'Connor, M., ed., Bilharziasis, London, Churchill, pp. 36-62 208 A. DAVIS & D. R. BAILEY Hairston, N. G. (1965) Bull. Wld Hlth Org., 33, 45 Hairston, N. G. & Jachowski, L. A. (1968) Bull. Wid Hlth Org., 38, 29 Hairston, N. G. & Meillon, B. de (1968) Bull. Wld Hlth Org., 38, 935 Hawking, F (1963) In: Schnitzer, R. J. & Hawking, F., ed., Experimental chemotherapy, New York, Academic Press, vol. 1, p. 901 Hawking, F. & Marques, R. J. (1967) Bull. Wld Hlth Org., 37, 405 Macdonald, G. (1965) Trans. roy. Soc. trop. Med. Hyg., 59, 489 Sasa, M. (1963) Bull. Wid Hlth Org., 28, 437 Sasa, M. (1967) Bull. Wld Hlth Org., 37, 629 WHO Expert Committee on Filariasis (Wuchereria and Brugia Infections) (1967) Wld Hlth Org. techn. Rep. Ser., 359, 5, 24