The prolonged epidemic of anthroponotic cutaneous leishmaniasis in Kabul, Afghanistan: ‘bringing down the neighbourhood’

TRANSACTIONSOF THE ROYALSOCIETYOF TROPICALMEDICINEAND HYGIENE(2003) 97, 170 176

The prolonged epidemic of anthroponotic cutaneous leishmaniasis in Kabul, Afghanistan: "bringing down the neighbourhood" H u g h Reyburn 1, M a r k R o w l a n d 1,2, M o h a m m m e d M o h s e n 1, B i s m u l l a Khan 3 and Clive D a v i e s 2 1HealthNet International, Kabul, Afghanistan; 2London School of Hygiene and Tropical Medicine, Keppel Street, London WCIE 7HT, UK; Slnstitute of Malaria and Parasitic Disease, Kabul, Afghanistan Abstract In order to investigate the distribution and causes of the spread of anthroponotic cutaneous leishmaniasis (ACL) in Kabul, Afghanistan, a cross-sectional study was conducted during 1997-98 amongst 75 787 residents in the 13 central districts of the city. Using data on active lesions and scars with their times of onset, migration patterns and age of subjects, 2 independent methods were used to estimate, retrospectively, the annual incidence of ACL in recent years. Results indicated a rapid increase in incidence from 1987, peaking in 1996 when an estimated 12% of the population had active disease. Active prevalence was lowest in infants (aged < 2 years), and while risk was gender-independent in children and adolescents, active prevalence in those aged > 20 years was significantly higher amongst females than males (odds ratio [OR] = 1.51, 95% CI 1.34-1.70). About 44% of lesions were located on the head, 38% on upper limbs, 16% on lower limbs and 2% elsewhere. The relative frequency of head lesions dropped with age (P < 0.001), and amongst adults was lowest amongst males (P < 0.001), possibly due to the protective effect of a beard. Within the study population, 32% reported that they had immigrated from outside Kabul, 34% that they had been born in Kabul but had since migrated to another district of the city, and 34% that they had been born in the district of their present residence. Active prevalence amongst those born in their current district of residence was positively associated with the percentage of immigrants in their district (P = 0.027), indicating that a 1% increase in the percentage of immigrants increased the odds of an active lesion by 12% (OR = 1.12, 95% CI 1.01-1.24), but there was no association with the percentage of migrants from other districts in Kabul (P = 0.65) or with war damage (P = 0.33). As active prevalence was not significantly greater in immigrants than local Kabulis, these results support the hypothesis that the epidemic in Kabul has been maintained by a steady influx of susceptible immigrants. It is important that the new opportunities for social development that now exist in Kabul are not hampered by this unpleasant and stigmatizing disease. As population movement is clearly a contributing factor to its transmission, this threat is very real. Keywords: anthroponotic cutaneous leishmaniasis, epidemiology, prevalence, transmission, Afghanistan Introduction Anthroponotic cutaneous leishmaniasis (ACL) has been endemic in parts of Afghanistan for centuries, notably in Herat to the west and Kandahar to the south (Nadim et al., 1979). However, in Kabul, at an altitude of 1850 m, ACL has only been described relatively recently; between 1947 and 1963 only 96 cases were reported from clinics in Kabul, most of which were thought to have been imported from other areas. From 1964 there was a progressive rise in notifications, with 780 cases reported from July 1967 to July 1968 in a newly built housing area on a hillside near the city centre, prompting a round of residual spraying of 6000 houses with D D T (Omar et al., 1968). In 1972, Nadim investigated an outbreak of ACL in new housing projects on the alluvial plains in the northern suburbs of Kabul. A community survey for ACL found a prevalence of 11.3% of active lesions; all ages and both genders were approximately equally affected and only 4.3% of those surveyed had evidence of past infection. Following this outbreak, cases were reported from previously unaffected towns and villages to the north of Kabul (Nadim & Rostami, 1974). Ashford et al. (1992) reviewed records from the 3 clinics specializing in ACL treatment between 1963 and 1990; these showed a slow but progressive rise in the number of cases, although during this period there had also been a large increase in both housing and population. As in 1974, analysis of cases reported in 1990 showed that all ages were at approximately equal risk. The lack of evidence supporting any accumulation of immunity with age, in spite of significant numbers of cases reported over the previous 20 years, led Ashford et al. (1993) to coin the term 'prolonged epidemic', Address for correspondence: Clive Davies, Disease Control and Vector Biology Unit, London School of Hygiene and Tropical Medicine, Keppel Street, London WCIE 7HT, UK; fax +44 (0)20 7927 2164, e-mail [email protected]

whereby local transmission was being maintained by the importation of both a reservoir of active cases and non-immunes from other areas. Following Ashford's work in 1990 no more data were routinely recorded. Intense warfare in Kabul between 1992 and 1996 resulted in severe damage to more than half of the city (including all records of ACL), with subsequent large-scale population movement and a breakdown of health and sanitary services (Marsden, 1998). Anecdotal reports in 1996 indicated that ACL was seen with increasing frequency in most areas of the city; concern was expressed by a number of agencies including WHO, resulting in the funding of a nongovernmental organization-led programme of which the present study was part, to identify the extent of the problem attributable to ACL and to identify effective means for its control. The present study was undertaken in 1997-98 to describe the distribution of ACL in Kabul over space and time and to suggest causes for any recent increase. An understanding of these questions was intended to be used to plan a programme for prevention and treatment although this subsequently became impossible as a result of the deteriorating political and security situation in 1998-99. It is hoped that now that the outlook for social reconstruction in Afghanistan is improving, an opportunity will exist in the near future to readdress the problem.

Methods Between November 1997 and Jan 1998, a household survey was conducted by a team of survey staff from HealthNet International and the Institute of Malaria and Parasitic Disease of the Islamic Emirate of Afghanistan. Survey areas were chosen to be representative of the 13 administrative districts of urban Kabul (2 remaining districts being rural areas). Within each district, survey areas were selected by the survey team leader to represent different types of house construction

ANTHROPONOTIC CUTANEOUSLEISHMANIASISIN AFGHANISTAN (e.g. brick, mud or cement) and different habitats (e.g. hillsides, semi-rural areas and high-rise blocks) in proportion to their contribution to the district population. Selection of survey areas within districts was made according to the judgement of the survey supervisor who was an experienced cartographer. The target sample size for each district aimed to include a minimum of 5 % of the district population. Survey workers worked in pairs visiting all occupied houses in the streets or areas selected for survey. Houses which were normally occupied but where the inhabitants were absent at the time were not included; refusal to participate was rare and not recorded. The head of household or a responsible adult, nearly always the mother or father, was interviewed. A census of all household residents was made including age, gender, length of time resident in the district and in Kabul, past or active A C L and the estimated month or year of onset. Wherever possible, scars and active lesions were examined. From March to June 1997, a more detailed household survey of A C L was undertaken in district 8 as part of another study; data on the affected site of 1021 active cases amongst a total of 9145 subjects were recorded from clinical examination. These data were used in the analysis of differences in the site affected by age and gender. One area (in district 5) was used to validate the identification of cases. This area was re-visited a few days following the standard survey described above and with no prior warning to the survey staff. The survey data were reviewed to identify consecutive households from which 200 active cases had been identified; these houses were re-visited and cases which had been recorded were identified, informed consent was obtained for a doctor to examine and take a slit-skin smear from each case; slides were fixed in methanol, Giemsastained and examined by an experienced microscopist. Positive slides were checked by a second microscopist before being allocated as 'slide-positives'. A cartographer on the survey team estimated the extent of damage in each district by visual inspection and ranked districts from the most war damaged (rank 1) to the least war damaged (rank 13). N o formal Ethical Review Committee was operating in Afghanistan at the time, but written approval to conduct the study was obtained from the Institute of Malaria and Parasitic Disease, Islamic Emirate of Afghanistan. Informed consent was obtained from all participants in the study, and all active cases found in the study were offered free treatment at the HealthNet International clinic in central Kabul.

171

Data were single-entered in Epi Info 5 (CDC, Atlanta, CA, USA) and analysed in Microsoft Excel (Microsoft Corp., Seattle, WA, USA) and Stata Statistical Software 7 (Stata Corp., College Station, TX, USA).

Results Study population parameters A total of 75 787 residents was clinically surveyed, reflecting roughly 6% of the estimated city population of 1.2 million (Table 1). Th e mean age of the population surveyed was 20 years, with 57% aged < 18 years, and 49% male. About 21% (15585) had a history of CL, as demonstrated by characteristic scars, and 4.7% (3550) had active ACL. Assuming that the survey was fully representative of the population of Kabul, this indicated that about 56 000 of the population had active A C L in 1987. Clinical diagnosis of active A C L lesions was validated in one area of Kabul; the findings are summarized in Table 2. Amongst active (ulcerated) lesions, the mean duration of slidepositive lesions was 4 months compared to 5.9 months for slide-negative lesions. Variation in leishmaniasis incidence over time Retrospective estimation of annual A C L incidence in Kabul since the 1970s was carried out by 2 independent methods. In the first analysis (Fig. 1), following the method of Davies et al. (1994), the annual incidence of new cases amongst the non-immune population during each of the previous 20 years was estimated by taking into account the year of onset of previous or active ACL. Th e denominators to estimate incidence in each year were calculated by taking into account age and length of residence in Kabul. Cases where the onset of A C L was earlier than the years of residence in Kabul were excluded. Cases attributed to 1997 include only cases whose onset was reported from the beginning of 1997 to the date of the survey (November 1997 to January 1998). In the second analysis (Fig. 2), past A C L incidence rates were estimated retrospectively from the relationship between length of residence in Kabul and cumulative prevalence (i.e. not taking into account year of onset). Annual incidence rates during each year since 1985 were calculated by comparison of the cumulative prevalence of consecutive age groups (at 1-year intervals). As before, cases where the onset of A C L was earlier than the years of residence in Kabul were excluded. Both analyses demonstrated the sudden emergence of the A C L epidemic in Kabul during the mid-1990s,

Table 1. S t u d y population parameters by district of Kabul, A f g h a n i s t a n , 1997-98

District 1 2 3 4 5 6 7 8 9 10 11 15 16 Total

War damage rank ~ 1 8 2 10 3 4 5 6 9 11 12 13 7

N 3396 6928 6611 9675 7673 5400 5928 2805 5305 8569 6495 5079 1923 75787

Active ACL n 133 443 247 637 404 202 323 59 71 262 385 271 113 3550

ACL, anthroponotic cutaneous leishmaniasis. a1, most war damaged; 13, least war damaged.

Past A C L n (%) 752 2591 1590 1240 1580 709 1801 320 389 1146 1518 1603 346 15585

(22) (37) (24) (13) (21) (13) (30) (11) (7) (13) (23) (32) (18) (21)

Active prevalence in non-immunes %

Cumulative prevalence %

5 10 5 8 7 4 8 2 1 4 8 8 7 6

26 44 28 19 26 17 36 14 9 16 29 37 24 25

Percentage of population Migrant

Immigrant

29 30 43 32 40 35 27 25 36 31 34 38 27 32

30 34 29 33 34 22 27 20 34 30 44 42 35 34

172

H, REYBURN E T A L .

T a b l e 2. C l i n i c a l a n d slide test results o f 217 cases w i t h active c u t a n e o u s l e i s h m a n i a s i s l e s i o n s i d e n t i f i e d b y s t a n d a r d survey, Kabul, A f g h a n i s t a n , 1997-98 Ulcerated lesion Slide t a k e n P e r c e n t a g e slide-positive M e a n age (years) M e a n no. lesions Mean duration (months) Refused slide Subject n o t f o u n d at re-visit Total

138 73% 16.6 3.2 4.6

Chronic lesion 4 25% 7.6 2.6 24

Clinically not CL

Healing lesion

10 10% 27.3 1.6 9.8

37 30% 21.6 2.9 8.1

Total 189 56% 18.7 2.9 6.2 16 12 217

CL, cutaneous leishmaniasis.

1412 rx

10 8 "~

6

'~

4 2¸

Year Fig. 1. Estimated annual incidence of anthroponotic cutaneous leishmaniasis in Kabul, Afghanistan, 1976-97.

12

r~

1i 4

1 Year Fig. 2. Estimated annual incidence of anthroponotic cutaneous leishmaniasis in Kabul, Afghanistan, 1985-97.

ANTHROPONOTIC CUTANEOUS LEISHMANIASIS IN AFGHANISTAN with a strildng peak annual incidence rate of 12% in 1996.

Effects of age and gender on leishmaniasis risk T h e most direct test of the effect of age and gender on A C L risk was the analysis of active C L prevalence in the n o n - i m m u n e population, i.e. those with no detectable history of A C L (Fig. 3). T h e 2 most striking patterns observed from Fig. 3 were that active prevalence was very low in infants (aged < 2 years), and that while risk was gender-independent in those aged < 20 years (odds ratio [OR] = 1.03, 95% CI 0 . 9 5 - 1 . 1 2 , )~2 = 0.5, 1 d.f., P = 0.48), amongst adults aged /> 20 years, active prevalence was significantly higher amongst females than males ( O R = 1.51, 95% CI 1.34-1.70, Z2 = 48.1, 1 d.f., P < 0.001). Age and gender associations with risk were further investigated by comparison of the n u m b e r and site of lesions on the body. Data on the site affected of 1021 active cases amongst 9154 people were analysed. T o simplify the analysis 204 cases with lesions on multiple sites were excluded. O f those with a single site affected, 44% had a head lesion, 38% had an upper limb lesion, 16% had a lesion on the lower limb and 2% a lesion on another site. T h e n u m b e r of lesions per case was higher amongst those aged ~> 20 years (mean = 2.06, median = 1.0) than those aged < 2 0 years ( m e a n = 1.5, reed-

173

ian = 1.0) (I<-mskal-Wallis, Z2 = 17.1, P < 0.001), with no significant difference between males and females (Kruskal-Wallis, ?(2 = 0 . 0 8 , P = 0 . 7 8 ) . The m e a n duration of lesions was 5.4 months, again with no significant difference between males and females (unpaired two-tailed t test, P = 0.8) or between those aged ~> 20 and < 20 years (unpaired two-tailed t test, P = 0.85). T h e proportion of head lesions was higher in those aged < 20 compared to those aged > / 2 0 years ( O R = 3.32, 95% CI 2 . 4 4 - 4 . 5 4 , ;42= 63.9, P < 0.001) and the proportion of upper limb lesions was higher amongst those aged >~ 20years ( O R = 2 . 5 1 , 95% CI 1.86-3.4, Z2 = 39.5, P < 0.001) (Figs 4 and 5). T h e site of lesions was significantly different between males and females of different ages. H e a d lesions as a proportion of all lesions amongst those aged > / 2 0 years was higher for females than males (OR=2.38, 95% CI 1.29-2.80, Z 2 = 1 1 . 4 , P < 0.001) and amongst those aged < 20 years was slightly lower amongst females than males ( O R = 0.6, 95% CI 0 . 4 1 - 0 . 8 8 , Z 2 = 7.6, P = 0.005). U p p e r limb lesions as a proportion of all lesions in those aged < 20 was higher amongst females than males ( O R = 1.88, 95% CI 1.24-2.86, ?(2_ 9.7, 1 d.f., P = 0.002), but amongst those aged 1> 20 years there was no significant difference ( O R = 0.70, 95% CI 0 . 4 4 - 1 . 1 , ;(2 = 2.6, 1 d.f., P = 0.1).

14 12 10-

e~

e

6-

42/ I 0

10

1'5

2'0

2'5

30 3'5 40 4'5 5'0 5'5 Mid-point of age band (years)

60

6'5

70

7'5

8'0

Fig. 3. Active cutaneous leishmaniasis prevalence amongst non-immunes by age, Kabul, Afghanistan, 1997-98. Error bars indicate 95% CI. 100~,

[] Males [_11Fema!es_ i

©

} i

0-4

t

5-9

t

10-19 20-29 Age band (years)

i

i

30-39

i

940

Fig. 4. Head lesions as a proportion of all cutaneous leishmaniasis lesions at each age band, Kabul, Afghanistan, 1997-98. Error bars indicate 95% CI.

174

H. REYBURNETAL. 80

o

60-

©

'~ 40 0 .,~

20

0-4

5-9

10-19 20-29 Age band (years)

30-39

I>40

Fig. 5. Upper limb lesions as a proportion of all cutaneous leishmaniasis lesions at each age band, Kabul, Afghanistan, 1997-98. Error bars indicate 95% CI. Effects of migratory history and war damage on leishmaniasis risk Migration and immigration rates in Kabul since 1976 were estimated from the analysis of reported durations of residence in Kabul and in the district of present residence, and year of birth. Estimated immigration to Kabul prior to 1990 appears to show a bias to rounded numbers, but indicated high immigration to Kabul during 1970s, with smaller peaks in 1983 and between 1996 and 1997. Migration within Kabul has been high throughout the last 30 years with increasing rates from 1992 to 1997 (Fig. 6). At the time of this survey in 1997-98, 24 528 (32%) reported that they had been born outside Kabul, 25 487 (34%) that they had been born in Kabul but had since moved to another district of the city, and 25 747 (34%) that they had been born in the district of their present residence. Cumulative prevalence of ACL amongst non-immunes by years resident in Kabul for these groups is shown in Fig. 7. For each subpopulation the current force of infection was estimated from these data by quadratic regression of -In(proportion susceptible) vs. years in Kabul, i.e. to allow for a steadily increasing trend in incidence (Ashford et al., 1993). Data up to 10 years were used to fit the model, i.e. before the fitted graphs pass the point of infection. The estimated current forces of infection,

taken from the linear term in the models were: 0.070/ year (95% CI 0.050-0.091), 0.082/year (95% CI 0.064-0.101) and 0.085/year (95% 0.032-0.138) for stayers (those born in their district of present residence), movers (born in Kabul but who had since moved district) and immigrants (born outside Kabul), respectively. It is notable that active prevalence amongst the susceptible (i.e. unscarred) population, 0.059 (3550/60 202), was rather less than any of these measurements of force of infection indicating that active ACL lesions tend to persist for < 1 year, and/or that incidence in 1987-98 was less than in previous years Comparisons of incidence estimated from changes in cumulative prevalence with time exposed in Kabul need to be interpreted with caution; as risk is not ageindependent (Fig. 3), age differences between the 3 migration groups were likely to affect these estimates of incidence. The mean age of those who were resident in their place of birth was 10.2 years, compared to 16.2 years amongst those who had moved district within Kabul since birth, and 33.6 years amongst those who had migrated into Kabul since birth. In order to control for the likely confounding effect of age in estimating risk of infection amongst different migration groups, we investigated the effect of migration group on active prevalence, after controlling for age and gender.

16[] Migrant

14-

• Immigrant

12.o ©

8 6

E

42 0 ~rm

Year Fig. 6. Percentage of the population migrating within and immigrating into Kabul, Afghanistan, 1976-97.

ANTHROPONOTIC CUTANEOUS LEISHMANIASIS IN AFGHANISTAN

175

40°x o o []

30-

x

o°o~o Xx

oo~ x

× x x

o~eo°

•

x

oo

Xmx~

,

•

• Born Kabul, never moved] a Born Kabul, moved xNot born in Kabul

x

~

o

x

o

x

o •

×

e~

.~o

20

a

"

x: i"

1 0 x"°

0

1~0

20

3'0

40

5'o

6;

7o

Years of residence Fig. 7. Cumulative prevalence of anthroponotic cutaneous leishmaniasis by years of residence in Kabul, Afghanistan and by migration status, 1997-98. Movers had a slight (but significant) higher risk than both stayers ( O R = 1.12, P = 0.005) or immigrants (OR = 1.25, P < 0.001), but there was no difference in risk between immigrants and stayers. Hence, the relatively high force of infection estimated above for immigrants appears to be attributable to the age structure of this subpopulation, rather than to their migratory status per se. Both immigration and intra-Kabul migration are likely to be associated geographically with environmental factors which could directly impact on sandfly populations, and hence transmission. For example, immigrants and migrants could be poorer than stayers and so live in different housing conditions; or high levels of migrants or immigrants in an area could lead to high-density living. Alternatively, migration could impact on transmission (i) by introducing infectious cases, or (ii) more significantly, by introducing susceptibles and so preventing herd immunity from dampening the effective reproduction number, R (Anderson & May, 1992). In order to distinguish these effects, in the final analysis we tested whether the active prevalence amongst stayers in districts was significantly affected by either war damage (graded by inspection), the percentage of the population who were immigrants, and/or the percentage of the population who were migrants (after controlling for age and gender) (Table 1). There was no evidence of any effect due to war damage (P = 0.334) or due to migrant rates (P = 0.649); but active prevalence amongst stayers was significantly positively associated with the percentage of immigrants in a district ( P = 0.027), indicating that a 1% increase in the percentage of immigrants increases the odds of an active lesion by 12% (OR = 1.12, 95% CI 1.01-1.24). Discussion Using 2 independent analytical approaches (Figs 1 and 2), this study has demonstrated a large increase in incidence of ACL in Kabul from 1987-96 (when it peaked) with some suggestion of a fall in 1997, following which no more data were available. Recent observations reported by the W H O indicate that incidence rates in Kabul are again extremely high (WHO, 2002), possibly as the result of the recent war. As in the 1990s, the period analysed in this paper, it is not immediately apparent whether the increase in incidence is due to environmental changes directly impacting on sandfly

populations, or due to the immigration of either infectious cases or a susceptible population. The timing of the 1990s increase in ACL incidence would certainly suggest a connection with the heavy fighting and associated events (population movement, destruction of buildings, breakdown of rubbish collection and sanitary services) that occurred between 1992 and 1995 (Marsden, 1998). Previous studies in other endemic areas have claimed an environmental association between social/environmental disruption and epidemic ACL. For example, Tayeh e~ al. (1997) suggested an association between building rubble and ACL in Aleppo, and Masoom & Marri (1982) reported a severe outbreak of ACL in Quetta, Pakistan following the 1935 earthquake. Whilst there was massive destruction of large areas of Kabul in the 3 - 4 years prior to the peak of ACL incidence, using our rather crude scoring system, we were unable to demonstrate any significant association with war damage. It may be that environmental damage has been associated with an increase in ACL transmission but this has been offset by an inverse relationship between population density and war damage, i.e. that war damaged areas were less densely populated. The one area of Kabul (district 9) with a relatively low prevalence of ACL is almost exclusively composed of concrete housing blocks; in such an area there may be a lack of breeding sites and those living above the second floor may be protected from sandfly bites (Hewitt et al., 1998). It is difficult to separate the contribution of various events that occur following natural disasters and war. Migrants and immigrants are likely to experience poor living conditions (missing windows, crumbling walls, etc.) as well as potentially increasing the susceptible population and the reservoir of infectious cases. All these effects would lead to a relatively rapid effect on the non-migrant population ('bringing down the neighbourhood') and any differential effect between migrants and the settled population is likely to be difficult to measure with a cross-sectional survey. Inspection of the temporal variability in incidence and migration activity in Kabul (Figs 1 and 6) suggests that annual changes in both migration and immigration rates mirror the increasing trend in incidence. However, correlating temporal trends is not a robust test of the hypothesis as there are too many potential confounding variables. Instead, we used cross-sectional analyses to measure

176 the effects of migratory activity on individuals and populations (Table 1). T h e s e showed that while migrants appeared to be at the greatest risk (indicating that they may be the poorest group in Kabul), the percentage of migrants within a district had no apparent impact on risk for people who already lived there. Conversely, immigrants appeared to be at no greater risk than local Kabulis, but as immigration rates increased there was a significant increase in risk for locals living in close proximity (i.e. within the same district). Thus, our analyses provide the first quantitative evidence in support of the Ashford er al. (1993) assertion that the epidemic in Kabul was sustained by high levels of immigration, which provided a continuing source of non-immunes. However, as active prevalence is no higher in immigrants than stayers (even when we only focus on recent immigrants, data not shown), there is no evidence that immigration maintains transmission through the introduction of infectious cases. Indeed, the extent to which immigrants to Kabul are coming from A C L - e n d e m i e regions is unclear. T h e W H O continues to demonstrate great concern for the ongoing massive epidemic in Kabul, and is urgently aiming to help introduce an appropriate control programme. Vector control programmes against the local vector, Phlebotomus sergenti (Killick-Kendrick et al., 1995), need to be designed on the basis of empirical evidence of where and when people are being infected. In the absence of extensive entomological data collection, this can sometimes be inferred indirectly by examination of the demographic risk factors for infection, and by examining the site of the lesions on the body. In this study we showed that children aged < 2 years had a significantly lower risk of active prevalence than those aged > 2 years (Fig. 3), possibly because of the long incubation period of Leishmania tropica (Ashford et al., 1993). Alternatively, infants may be bitten less often by sandflies as they generate a relatively weak p l u m e of odour attractants such as carbon dioxide (Pinto et al., 2001). Facial lesions were relatively more c o m m o n in children than adults (Figs 4 and 5), which is consistent with findings from other studies of A C L (Nadim & Tahvildari-Bidruni, 1977; Seyedi-Rashti et al., 1984; Ashford et al., 1992). Active prevalence of A C L was higher amongst adult females than males (as previously found in suburban centres of Kabul: Ashford et al., 1992) and other A C L - e n d e m i c urban foci (e.g. N a d i m & Tahvildari-Bidruni, 1977; Rowland et al., 1999), perhaps because males sleep away from h o m e more often and urban A C L transmission is mainly domestic. Facial lesions were also more c o m m o n amongst adult w o m e n than men, presumably due to the protective effect of wearing a beard which was legally enforced in Kabul from 1996. T h e choices for vector control in Kabul are currently whether to spray houses with residual insecticides, whether to provide insecticide-treated nets, or whether to treat chaddors and bedsheets with insecticide. T h e recent intervention trial in Kabul (Reyburn et al., 2000) indicates that all 3 methods should be effective, but the costs should be least if chaddors and sheets are treated. Recent political developments in Kabul suggest that there will soon be improved opportunities for social development, and it is important that this is not associated with a continuing epidemic of A C L . As population m o v e m e n t is clearly a contributing factor to the transmission of A C L in Kabul, this threat is very real.

Acknowledgements HealthNet International's Leishmaniasis Control Programme was supported by Norwegian Church Aid and the European Commission Humanitarian Office. Neither of these donors accepts responsibility for the information provided or views expressed. We thank Profs Dick Ashford and Anthony Bryceson for their advice and support, and Ilona Carneiro for help and comments on data analysis. We also thank the Institute of Malaria and Parasitic Disease, at the time of the

H. REYBURN ETAL. Islamic Emirate of Afghanistan, for cooperation and implementation of field surveys.

References Anderson, R. M. & May, R. M. (1992). Infectious Diseases of Humans. Dynamics and Control. Oxford: Oxford University Press. Ashford, R., Kohestany, K. & Karimzad, M. (1992). Cutaneous leishmaniasis in Kabul: observations on a prolonged epidemic. Annals of Tropical Medicine and Parasitology, 86, 361-371. Ashford, R. W., Rioux, J.-A., ffalouk, L , Khiami, A. & Dye, C. (1993). Evidence for a long-term increase in the incidence of Leishmania tropica in Aleppo, Syria. Transactions of the Royal Society of Tropical Medicine and Hygiene, 87,247-249. Davies, C. R., Llanos-Cuentas, A., Canales, J., Leon, E., Alvarez, E., Monge, J., Tolentino, E., Gomero, Q., Pyke, S. & Dye, C. (1994). The fall and rise of Andean cutaneous leishmaniasis: transient impact of the D D T campaign in Peru. Transactions of the Royal Society of Tropical Medicine and Hygiene, 88, 389-393. Hewitt, S., Reyburn, H., Ashford, R. & Rowland, M. (1998). Anthroponotic cutaneous leishmaniasis in Kabul, Afghanistan: vertical distribution of cases in apartment blocks. Transactions of the Royal Society of Tropical Medicine and Hygiene, 92, 273-274. Killick-Kendrick, R., Killick-Kendrick, M. & Tang, Y. (1995). Anthroponotic cutaneous leishmaniasis in Kabul, Afghanistan: the high susceptibility of Phlebotomus sergenti to Leishmania tropica. Transactions of the Royal Society of Tropical Medicine and Hygiene, 89, 477. Marsden, P. (1998). The Taliban: War, Religion and the New Order in Afghanistan. Karachi: Oxford University Press, pp. 88-102. Masoom, M. & Marri, S. M. (1982). Current status of leishmaniasis in Pakistan. In: Current Trends in Leishmania Research. Bhaduri, A. N., Basu, M. K., Sen, A. K., Kumar, S. (editors). India: Council for Scientific and Industrial Research, pp. 231-236. Nadim, A. & Rostami, K. (1974). Epidemiology of cutaneous leishmaniasis in Kabul, Afghanistan. Bulletin of the World Health Organization, 51, 45-49. Nadim, A. & Tahvildari-Bidruni, Gh. (1977). Epidemiology of cutaneous leishmaniasis in Iran: B. Khorassan. Part VI. Cutaneous leishmaniasis in Neishabur, Iran. Bulletin de la Societe de la Pathologie Exotique et de ses Filiales, 70, 171-177. Nadim, A., Javadian, E., Noushin, M. K. & Nayh, A, A. K. (1979). Epidemiology of cutaneous leishmaniasis in Afghanistan. Part 2. Anthroponotic cutaneous leishmaniasis. Bulletin de la Societe de Pathologic Exotique et de ses Filiales, 72, 461-466. Omar, A., Saboor, A., Saidi, S. M. & Serf, V. (1968). Cutaneous leishmaniasis in Afghanistan. Afghan Journal of Public Health, 1, 1-6. Pinto, M. C., Campbell-Lendrum, D. H., Lozovei, A. L., Teodoro, U. & Davies, C. R. (2001). Phlebotomine sandfly responses to carbon dioxide and human odour in the field. Medical and Veterinary Entomology, 15, 132-139. Reyburn, H., Ashford, R., Mohsen, M., Hewitt, S. & Rowland, M. (2000). A randomized controlled trial of insecticide-treated bednets and ehaddars or top sheets, and residual spraying of interior rooms for the prevention of cutaneous leishmaniasis in Kabul, Afghanistan. Transactions of the Royal Society of Tropical Medicine and Hygiene, 94, 361-366. Rowland, M., Munir, A., Durrani, N., Noyes, H. & Reyburn, H. (1999). An outbreak of cutaneous leishmaniasis in an Afghan refugee settlement in north-west Pakistan. Transactions of the Royal Society of Tropical Medicine and Hygiene, 93, 133-136. Seyedi-Rashti, M. A., Keighobadi, K. & Nadim, A. (1984). Urban cutaneous leishmaniasis in Kerman, Southeast Iran. Bulletin de la Societe de Pathologie Exotique et de ses Filiales, 77, 312-319. Tayeh, A., Jalouk, L. & Cairncross, S. (1997). Twenty years of cutaneous leishmaniasis in Aleppo, Syria. Transactions of the Royal Society of Tropical Medicine and Hygiene, 91, 657-659.

WHO (2002). Cutaneous leishmaniasis, Afghanistan. Weekly Epidemiological Record, 77, 246. Received 30July 2002; revised 5 November 2002; accepted for publication 8 November 2002