TRANSACT~NS OF THE ROYAL SOCIETY OF TROPICAL MEDICINE AND HYGIENE (1988) 82, 117-121

A quantitative analysis of the diagnostic value of diethylcarbamazine provocation in endemic Wuchereria bancfofii infection*

117

Mohamed Sabry United States Naval Medical Research Unit No. 3, Cairo, Egvpt

Abstract The efficiency (A) of the day-time provocative

effect of diethylcarbamazine (DEC) on the circulation of microfilariae (mf) of nocturnally periodic Wuchereriu bancrofti compared with the night mf count was evaluated in terms of sensitivity and specificity. In one test, the mean mf count for finger-blood samples from the same person on 3 successive nights was compared with the day-time post-DEC count; another test compared the mf count in one night-tune finger-blood sample with the following day-time post-DEC count. Both tests were efficient (A = O-74 and 0.77, S.E. = 0.07 and 0.07, respectively). Night and day-time post-DEC mf counts were significantly correlated (r = 0.81, P<O-001, and r = O-79, P<O*OOl for the 2 tests respectively), but mf densities tended to be relatively lower after DEC provocation than in the correspond- ing night-blood samples. In both tests the regression coefficients differed significantly from zero (PtO*OOl). The simple 24h test was almost as efficient as the 3-nights test and based on its regression relationship, a table is presented predicting the expected night mf counts from the observed day-time post-DEC counts. The provocative dose of DEC, 2 mglkg body weight, evoked no side-reaction in either microfilaraemic or uninfected persons.

Introduction Night surveys for detecting microfilariae of the

nocturnally periodic Wuchereria bancrofti in the circu- lating blood are difficult to perform and little co- operation from the community members may be expected.

FUJIMAKI (1956) indicated that a minimum of O-8 Kg DEC/ml blood is necessary to kill W. bancrofti microfilariae. SASA (1976) and other workers have shown that administration of DEC at night causes an immediate reduction in numbers of nocturnally periodic W. bancrofti in the blood, whereas giving DEC during the day-time stimulates W. bancrofti microfilariae to circulate in the blood (e.g., WIJ- EYARATNE et al., 1982). Different doses of DEC have been tried, but giving 2 mg/kg body weight, as

*This work was supported by the Naval Medical Research and Development Command, Naval Medical Command, National Capital Region, Bethesda, MD 20814, Work Unit No. 3M161102BSlO.AA.421.

The ouinions and assertions contained herein are the private ones of the author and are not to be construed as official or reflecting the views of the U.S. Navy Department or the naval service at large. Requests for reprints:

Research Publication Division, U.S. NAMRU-3, FPO, NY 09527, USA.

recommended by MANSON-BAHR & WIJERS (1973), and examiuing blood 45 to 60 min after DEC administration, were adopted in this study, the objectives of which were to determine whether (a) th&e is a statistically significant relationship between dav-time mf counts after DEC nrovocation and mf de&y in the standard finger-p&k night blood; (b) there is a statistically sign&ant difference between the relationship of the day-time post-DEC mf count and the mf count during the previous night (“last- night test”), on the one hand, and between the day-time post-DEC mf count and the mean mf count for the previous 3 successive nights (“3-nights test”), on the other hand; and to construct a table, using the linear regression relationship? to enable field workers to obtain, without mathemaucal or statistical calcula- tions, the expected night mf counts predicted from the observed day-time post-DEC counts.

Materials and Methods Lvm~hatic filariasis due to nocturnallv neriodic W.

ban&ofii is endemic in Qalyubia Gove&&ate of the Arab Republic of Egypt. Ezbet Zeinuldin, Ezbet Al-Mekhammes, Ezbet Al-Hakeem, and Kafr Hamam, 4 small villages around Marsafa, one of the main villages of Qalvubia Governorate, were chosen for the study. A- $lot survey had Geen made to determine the prevalence rate of microfilaraemia in these villages.

136 volunteers of 10 years and older of both sexes, excluding pregnant women, were chosen at random from the 4 villages. Each volunteer signed a consent form indicating his or her willingness to participate in the study.

From each volunteer the following samples were taken: a 20 mm3 night finger-blood &ple; between 2100 and 2300 h on 3 successive nights (for the “3-nights” test) or on one night only (for the “last-ninht” test): a samole at about 0930 to 1100 h on the folliwing m&ing j&t before giving the provoca- tive dose of DEC, and the post-DEC sample, 45 to 60 min after the DEC dose. 50 mg tablets of DEC citrate (FilararP, Nile Company for Drugs and Chemicals Industries. Cairo. Eevntj were used. The orovocative DEC do&, 2 n&&i-bbdy weight, was’ swallowed immediately by the volunteer while under observa- tion.

Each blood sample was taken from the finger by sterile lancet and a 20 mm3 volumetric micropipette, spread on a clean slide into a rectangle of 1.5 x 4 cm, left to dry overnight at room temperature, stained without lysis for half an hour in 1: 10 Giemsa solution, washed quickly in a solution of distilled water and physiological saline (2:1), and then left to dry. (Smears washed in this solution became clearer than

118

when washed in distilled water only.) Blood smears were examined at 100X magnification.

Results The nilot survey made 2 weeks before the main

study revealed n&rof&raemic prevalence rates of 39% in Ezbet Zeinuldin. 28% in Ezbet Al-Mekham- mes, 245% in Ezbet Ai-Hakeem, and 19% in Kafr Hamam. Of the 136 randomly chosen volunteers, 88 completed the 3-nights test, while 90 completed the last-night test.

The 34ghts test Microfilaraemic counts of 41 volunteers are shown

in Table 1. 39 volunteers had microflaraemia in one

or more of the night samples; 29 (74%) showed positive post-DEC counts, while the remaining 10 had post-DEC counts of zero (Tables 1 and 2), and were considered false negatives. 49 volunteers with all night and day mf counts of zero were omitted.

The last-night test 39 volu&ers performed the last night test, and 38

had microfilariae in their night blood samoles. Volun- teer No. 4, who had a mf count of zero in bight blood, became microfilaraemic after the DEC provocative dose, and was considered a false positive. As shown in Tables 1 and 2, of the 38 night positives, 31 (81.6%) gave positive post-DEC counts, while 8 had post- DEC counts of zero and were considered false

Table 1-W. bancr&i microlilariae counts per 20 mm3 linger-blood sample

Day-time

Volunteer First No. night

Second night

Third night

D2iOE DEC

45 mill after DEC

: 7 4 4

:z 1 X :: 1: 0

i 0 3

: 0 1 : : 1:: 121 16 22 1: 16 8

7 t

E 1:; 1: 1 3: 22 8 5

:: ;s :f 19 :: 25 0 1 1 1

12 ::

12 1:‘: 34 134 17

1; 2; 27

:: 35

0

:;: z’: ;; 1 1: 5 19

i: :t 0 1:

4 x 1

ii 8 A

:i 5: 28 : 5: 0 1 0 3

25 2 15 fS

8 1 s

8 i

29 6; 11: 155 s 3s

8 X X 40 5 37 10 f 6 5

;: X X

x 0 :: 10

: : 33

;; 20 8 0 8 1

72 3;

3: 6z 7; 8 120 2 0 1 0

102 5 :i 2: 5 E X X 14 30 6

8 1

116 X z

117 X 18 1;

i t

120 X 1 2; h :

133 X X 0 4

0 = mf count zero; x = blood sample not obtained

negatives. 51 volunteers with all night and day mf counts of zero were omitted.

Table 2 shows, for comparative purposes, the efficiency rates of both the 3-nights and the last-night tests, and estimates of the mean, the 50% cumulative frequency (5 probits), the median, and the median microfharial density (MFD-50).

Discussion Ejjihncy of the DEC provocative test

Although W. bancrofti is nocturnally periodic the presence of some microfilariae in a few pre-DEC day blood samples could be attributed to taking the blood samples while volunteers were still in the quiescent state after waking.

The efficiency (A) of the DEC provocative test is expressed as the proportion of mf positives or negatives revealed by the DEC test amongst those shown to be, respectively, positive or negative by their night samples. False positives ((w) are those shown to be positive by the DEC test, but who were negatives in the standard night test; false negatives @) are those negative by the DEC test but positive by the night test. Sensitivity of the DEC test is expressed as the proportion of positives obtained by the test compared with the total number positive by the standard method+, the night test, and IS measured as 1 - fl. Specificity IS the proportion of those negative by the DEC test compared with the total number negative by the night test, and is measured as l-or. Thus, the efficiency (A) of the DEC test is evahtated by taking into account false positives and false negatives, A = I-(ar+fl) (WIJEYARATNE et al., 1982). The standard error (SE.) of A is:

S.E. A=

119

where PI = CL, PZ = fi and Pl+ql=P2+q2=1; nl and g are the number of samples on which the estimates of OL and S respectively are based. The value of A varies from 0 to 1: the higher the value, the greater the efficiency.

The values obtained in the present last-night test (Table 3), A = 0.77 and S.E.A = O-07, represent a much higher efficiency than those reported by WIJ- EYARATNE et al. (1982) from the Nigerian savanna - A = 0.58 and S.E.A = O-11.

The relatively lower sensitivity compared with specificity? shown in Table 3, could be attributed to the biologrcal rhythm of the mf, the state of activity or sleep of the volunteer when donating the blood sample and the low night mf counts of 5 or fewer per 20 mm4 of blood, as the range of 5 to 20 mf/20 mm3 night blood is the optimum for transmission of W. bancrofti by C&x pipiens m&stus, the principal vector of filariasis in Egypt (SARRY, 1987). As the post-DEC counts are in general relatively lower than night counts, the number of mf provoked following low night counts (5 mg or fewer) will be about one mf/20 mm3, and the chance of observing mf in the blood sample is low. This increases the numbers of false negatives (p) and consequently reduces the degree of sensitivity (1 - S) of the DEC provocative test. This phenomenon of low night infections, and hence very low post-DEC mf counts, has to be taken into account in considering the purpose of the study. For example, if the study is to evaluate vector control measures, an increase in the number of very low post-DEC counts or of false negatives would indicate a decrease in the transmission rate and that the disease is dying out. On the contrary, if the purpose of the test is to assess the parasitological response, it might be necessary to calculate correction factors for these low grade microfilaraemias, and it is preferable to examine a random sample of the population at night.

Tabk 2-E%icknc~ sates of day-time DEC prov~~athn on lboctlmauy perhdic w. bancroffi showing for comparaive puTpow the estimates of the mean, the %@b mumdative frequency (5 p&its), the mdim, and the median microfihial density (MFD-SO)

hiicrofdsrial count/20 mm3 blood Rstio of micm6lsrial count

50% cumulstive w/b cumulstive No. No. False Fak frequency frrpuency

Test cxamimd positive(%) postives negatives Mean (5 probits) Median MFD-50 km (5 probits) Mcdisn MFD-50

3-nights 88 39w3) 272 17 18 13 100 100 100 100 Post-DEC 29(74.4) “0 10 5.6 3.3 5 4.7 20.8 194 278 36.2

Last-night 90 3N42.2) 28.5 19 19.5 15.4 100 100 100 100 Post-DEC 3H81.6) 1 8 5.8 3.5 4 4.6 204 184 20.5 29.9

Table 3-Evaluation of the day-time DEC provocative test in the diagnosis of W. bancrof?i infections in an endemic area io Qalyubia Governorate, Egypt, in comparison with Wijeyaratne’s data from the Nigerian savanna (WIJEYARATNE et al., 1982)

Present study Wijeyaratne’s 3-nights test Last-night test last-night test

Sensitivity 0.74 0.79 0.69 Specificity 1.00 O-98 O-89 False positives (ar) False negatives (S) x3

0.02 0.11 0.21 0.31

;,=;-‘“+B’ . . 0.74 0.07 O-77 0.07 0.58 o-11

I I r o-o 0.5 1-o 1.5 2.0 2.5

Mean of microfilariae counts, log (count+11

Fii. 1. Microf&riae counts repeated on same volunteers for 2 or 3 successive nights. Numbers beside a point indicate the number of counts with the Same value.

= 2.5- + E 2 2-o-

s

$'

9 l-5- c

;

: l*O-

'C co

E

2 o-5- .0 E e

-last-night

L 03 I z I 0.0 0.5 1.0 1.5 2.0 2.5

Day-time post-DEC microfilariae counts, log (count+ I)

Fig. 2. Regression lines comparing night microfilariae counts with DEC provocation is a consistent test, as the results day-time post-DEC micro6lsriae counts for the 3-nights and the of replicate tests do not vary in one direction, but in last-night tests (n=39 for both tests). areas with low mf densities it should be interpreted

Table 4-Expected microfihraemic night counts predicted from the observed day-time post-DEC provocation counts of W. bancmhi per 20 mm3 fioger-blood

Post-DEC count Night count Post-DEC count Night count (observed) (expect4 (observed) (expected)

1 7 2 10 3 14 4 18 5 22 6 25 7 29 8 32 9 36

10 39 11 42 12 46 13 49 14 52 15 56 16 59 17 62 18 66 19 69 20 72 21 75 22 79 23 82 24 85 25 88

26 92 27 95 28 98 29 101 30 104 31 108 32 111 33 114 34 117 35 120 36 123 37 127 38 130 39 133 40 136 41 139 42 142 43 145 44 149 45 152 46 155 47 158 48 161 49 164

The use of different volumes of blood samples, ranging from 10 mm3 to 1 ml, and of different mf counting methods during microfilarial surveys, needs to be considered to Iind factors that may correlate results obtained by the different methods. SASA (1976) applied correction factors to the data reported from Fiji by SOUTHGATE (1974), and found that the differences in mf rates obtained using blood smears made with 20 mm3 and 60 mm3 were the same as would be expected from the differences in the size of the blood samples. However, the differences between the rates observed by the blood smear methods and those obtained by the counting chamber or membrane filtration methods were much larger than expected from the volumes of the blood samples. This may be due to loss of a considerable proportion of microfilar- iae (about 30%) during lysing of blood smears (DENHAM et al., 1971). To overcome this loss I use the method described above for spreading the 20 mm3 blood sample, which does not need lysis before staining. In general, there is no correction factor applicable to a certain mfrate because the efficiency of detection of mf is not dependent upon the mf rate of the population but upon the pattern of distribution of the mf density in the population concerned (SASA, 1967, 1974).

In my experience, the DEC provocative test evoked no side-reactions in either microfilaraemic or unin- fected volunteers in areas endemic for W. bancroffi. It may cause febrile reactions in patients with high densities of Brugia malayi (WHO, 1974), and should be used with caution in endemic areas of Onchocercu volvulus or Loa loa infection, due to the risk of a severe Mazzotti reaction.

121

Expected microjilaraemic night counts predicted from corresponding observed day-time post-DEC counts

Using the regression equation of the last-night test, Table 4 shows observed post-DEC mf counts (C), from 1 to 50 per 20 mm3 finger-blood, and their corresponding expected night counts arranged, for the first time, into a tabular form to avoid the need for calculation and data transformation by field workers. References Denham, D. A., Dennis, D. T., Ponnudurai, T., Nelson,

G. S. & Francis? G. (1971). Comparison of a counting chamber and tick smear methods of counting micro- tilariae. Transactions of tke Royal Society of Tropical Medicine and Hygiene, 65, 521-526.

Diem, K. & Lenmer? C. (1970). S&nt$c Tables (7th edition). Basle, Switzerland: J. R. Geigy, p.61.

Fujimaki, H. (1956). Studies on the chemotherapy of 6lariasis. Nagasaki Medical Journal, 31, 930-947.

Manson-B&r, P. E. & Wijers, D. J. (1973). Ban&de induced appearance of Wuckereria bancrofti micro6lariae in the peripheral blood by day. In: Parasitoses of Man and Animals in Dar es Salam. Kampala. East African Litera- nue Bureau, pp. 353-337. -

with caution. However, the test is almost entirely specific and highly sensitive.

Correlation between day-time post-DEC and night mic- rojilarial counts

After trying various mathematical formulae, it was eventually demonstrated that the frequency distribu- tion of positive cases by mf density is generally logarithmically normal (SASA, 1%7), and this was confirmed by SOUTHGATE (1974) during his studies of filariasis in Fiji. Fig. 1 shows a straight-line rela- tionship between the mean and the variance of mf counts in blood samples taken from the same person on 2 or 3 successive nights, and the gradient of the straight line indicates logarithmic transformation, taking into account that persons with all counts zero were omitted. If negatives are included, the regression line takes the form of a Polya-Eggenberger distribu- tion, and does not fit the logarithmic normal type.

The logarithm of the count plus one, log (C + l), is used (MCMAHON et al., 1979), and from data shown in Table 1 the correlation coefficients (r) of the day-time post-DEC mf density and the night mf density were high: t = O%l, P<O+IOl; r = O-79, P<O-001 (DIEM & LENTNER, 1970) for the S-nights and the last-night tests respectively, where: n = number of volunteers having positive mf counts (n = 39 in both tests); x = log (day-time post-DEC mf count + 1) = log (C + 1); and y = log (night mf count + 1) = log (C + 1).

Quantitative relationship between the day-time post-DEC and the night microfilariul counts

Mathematical relationship. Applying the regression function y = a + b x, where a is a constant, and b is the regression coefficient, for the 3-nights test, where n = 39, y = O-593 + O-982 x, and for the last-night test, where n = 39, y = 0.624 + O-938 x.

The regression coefficients in both tests (b = O-982 and 0.938, respectively) differed significantly from zero, P~O.001 (DIEM & LENI’NER, 1970).

Graphical relationship. Using the regression equa- tion, where x ory = log (C + l), the co-ordinates of 3 points instead of 2, to be safer, were allocated in order to draw the regression line. Three round number values for C (1, 10, and 35) were chosen for convenience and values of x and y were calculated for these numbers for both tests. In Fig. 2 the regression lines of the 3-nights test and the last-night test almost coincide, which means that there is no need to perform the tedious and annoying 3-nights test so long as the last-night test will give almost the same result and efficiency.

McMahon. 1. E.. Marshall. T. F. DeC.. Vaughan. 1. P. & Kolstr+, N: (1979). ‘Tanzania F&r&s P&$ct: a provocauve day test with diethylcarbamazine for the detection of microfilariae of nocturnally periodic Wuckereria bancrofii in the blood. Bulletin of tke World Health Organization, 57, 759-765.

Sabry, M. (1987). A study of the quantitative relationskip between the blood densi@ of Wuchereria bancrofti micro- jikariae and development of infective larvae in Culex pipiens molestus in Egypt. Ph.D. Thesis, University of London.

Sass, M. (1%7). Microtia survey methods and analysis of survey data in filariasis control programmes. Bsdletin of the WorU Healtk Organization, 37, 629-650.

Sass, M. (1974). Methods for estimating the efficiency of detection of microlilariae in various volumes of blood samples. South-East Asian 3oumal of Tropical Medicine and Public Health, 5, 197-210.

Sass, M. (1976). Human FiLuiasis, a Glob@ Ssavey of f;e~aologv and Conzrol. Tokyo: Umverslty of Tokyo

Southgate, B. A. (1974). A quantitative approach to parasitological techniques in Bancroftian filariasis and its effect on epidemiological understanding. Transactions of tke Royal Society of Tropical Medicine and Hygiene, 68, 177-186.

Wijeyarame, P. M., Singha, P., Verma, 0. P. & Motha, B. (1982). Evaluation of the diethylcarbamazine provocative test in the diagnosis of Wuckereria bancrofti infections in the Nigerian savanna and the effects on Diperakmema perstans. Transactions of tke Royal Society of Tropical Medicine and Hygiene, 76, 387-391.

WHO (1974). Third Report of the WHO Expert Committee on Filariasis. Geneva: World Health Organization, Technic- al Report Series, No. 542, pp. 53-54.

Accepted for publicatkm 19 May 1987