Tropical Medicine and International Health doi: 10.1111/j.1365-3156.2006.01619.x volume 11 no pp 738–745 may 2006 Seroepidemiology of leptospirosis in southern Vietnamese children Khoa T D Thai1, Tran Quang Binh2, Phan Trong Giao1,2, Hoang Lan Phuong1,2, Le Quoc Hung1,2, Nguyen Van Nam3, Tran Thanh Nga1,4, Marga G A Goris5 and Peter J de Vries1 Division of Infectious Diseases, Tropical Medicine & AIDS, Academic Medical Center, Amsterdam, the Netherlands Tropical Diseases Clinical Research Center, Cho Ray Hospital, Ho Chi Minh City, Vietnam Binh Thuan Provincial Malaria Station, Phan Thiet City, Vietnam Department of Microbiology, Cho Ray Hospital, Ho Chi Minh City, Vietnam Department of Biomedical Research, Leptospirosis Reference Centre, Koninklijk Instituut voor de Tropen, Amsterdam, the Netherlands Summary objective To estimate the seroprevalence of human leptospirosis in southern Vietnam methods All pupils (n ¼ 961) of two primary schools in two communes in southern Vietnam were screened for the presence of serum Leptospira immunoglobulin (Ig)G Leptospira IgM was tested in 92 randomly selected samples IgM and IgG were tested with a commercially available enzyme-linked immunosorbent assay (ELISA) results Leptospira IgG was found in 123 (12.8%, 95% CI: 10.8–15.1) children with a male:female ratio of 1.5:1 At the age of years the prevalence was 11% Log-linear binary regression of the IgG seroprevalence by age showed a mean seroconversion rate of 1.5% (95% CI: 1.23–1.75) per year Five (5.4%, 95% CI: 2.3–12.1) samples were definitely positive for IgM Presence of serum Leptospira IgG antibodies was significantly associated with swimming in rivers (RR: 1.745, 95% CI: 1.117–2.724) and wading through water (RR 1.793, 95% CI: 1.181–2.722) conclusion Our data emphasize the importance of leptospirosis among children in Vietnam, despite the absence of severe disease, and stress the need for adequate and cheap diagnostics keywords leptospirosis, seroepidemiological studies, incidence, risk factors, regression analysis, Vietnam Introduction Leptospirosis is an ubiquitous zoonosis with preference for warmer climates, caused by infection by pathogenic Leptospira species (Vinetz 2001) The taxonomic classification of the genus Leptospira is very complex Traditionally two species were recognized, the saprophytic L biflexa and the pathogenic strains of L interrogans Of the latter more than 200 serovars have been identified, grouped into more than 20 serogroups (Bharti et al 2003) A recently developed molecular classification system recognizes 16 species (Levett 2001) The spectrum of clinical presentation of leptospirosis in humans is wide, ranging from asymptomatic infection to potentially fatal manifestations The majority of infections are subclinical or mild self-limiting systemic illnesses (Cumberland et al 2001; Levett 2001) It is often difficult to differentiate leptospirosis from other febrile diseases such as dengue, typhoid, malaria and 738 influenza (Sanders et al 1999; Levett et al 2000) Primary healthcare workers usually base their diagnosis on clinical signs and symptoms, without the aid of confirmatory tests Due to a lack of diagnostic awareness and confirmatory tests, leptospirosis is generally an underdiagnosed disease and the true incidence of leptospirosis remains unclear (Levett 1999) Since the 1930s Vietnam is known to be endemic for leptospirosis (de Lajudie & Brygoo 1953; Berman et al 1973a) In a previous study in Binh Thuan province, Vietnam, we showed that IgM-ELISA and two rapid tests for Leptospira antibodies were frequently positive in healthy individuals as well as in febrile patients (Wagenaar et al 2004) Despite the high seroprevalence, leptospirosis is rarely if ever diagnosed in Binh Thuan Apparently, the true incidence of Leptospira infections is much higher than the number of diagnosed cases, but not further quantified We recently showed that seroprevalence studies in children can be used to estimate the incidence of first ª 2006 Blackwell Publishing Ltd Tropical Medicine and International Health volume 11 no pp 738–745 may 2006 K T D Thai et al Seroepidemiology of leptospirosis in southern Vietnamese children dengue virus infections (Thai et al 2005) Seroprevalence studies for leptospirosis in Vietnam have been reported before, but not in children (Van et al 1998; Laras et al 2002) In this study we determined the seroprevalence of IgG Leptospira-specific antibodies in primary school children in Binh Thuan, southern Vietnam, estimated the annual incidence of infection, and searched for risk factors associated with past Leptospira infections Materials and methods Population and samples This study was carried out in two communes, Ham Kiem and Ham Hiep, in the coastal region of Binh Thuan province Binh Thuan is a province in southern Viet Nam ranging from the Truong Son forested mountains in the west to the south Chinese sea 150 km north-east of Ho Chi Minh city Similar to other southern provinces, the coastal range is an agricultural area with paddy fields, fruit orchards and small-scale cattle farming, optimal conditions for transmission of leptospirosis to humans In each commune, there were two primary schools of approximately equal size In both communes, all children, from to 14 years of age, of one primary school participated in a serological survey in March 2003 (Thai et al 2005) Of 977 school children, 16 were absent at the time of the study, leaving 961 children, 496 in Ham Kiem and 465 in Ham Hiep, in the study One millilitre of blood was obtained by finger puncture and collected in plain vials (Greiner, MinicollectÒ) It was left to clot at ambient temperature, centrifuged at approximately 90 g for 15 and transferred to a sterile vial for storage at )20 °C until testing Commercially available monowell ELISA test kits for IgM and IgG were used (Virion/Serion GmbH, Wuˆrzburg, Germany) These tests apply crude antigens from an isolated, concentrated and partially purified extract of Leptospira biflexa, containing genus-specific epitopes for all Leptospira subtypes Antigen-coated microwells were incubated with 100 ll, 1/100 diluted serum for 60 at 37 °C in a moist chamber incubator After washing with sodium chloride solution 0.1% and Tween 20TM, bound antibodies were detected after 30-min incubation with goat antihuman-IgM or -IgG (100 ll per well) and p-nitrophenylphosphate substrate (100 ll per well) For the IgM tests, the samples were pre-treated with monoclonal antibodies against rheumatoid factor, to adsorb potentially crossreacting non-specific antibodies Antibody concentrations were measured as optical density (OD) values, measured at 405 nm, with 630 nm as a reference, with a conventional absorbance reader (TECAN Sunrise Absorbance Microplate Reader, Tecan Austria GmbH, Groăndig/Salzburg, Austria) Results were expressed as the ratio between the sample OD value and the OD value of a calibration sample, enclosed in the ELISA kit (ODR), both after subtraction of the OD value of an enclosed blank specimen Test results were interpreted as follows: ODR IgG values >5 and 20 U/ml and ODR IgG < U/ml; IgM ODR between 15 and 20 U/ml and IgG > U/ml was interpreted as a possible acute secondary infection; IgM ODR between 15 and 20 U/ml and IgG ODR < was considered as a possible acute primary infection Statistical analysis Questionnaire All children were interviewed using a questionnaire concerning recent episodes of fever and certain risk factors related to water, soil and vector-borne transmission, environment around the house, domestic animals and nutrition Younger children sometimes needed help of their teacher to fill in the questionnaire Serology All serum samples were transported to the Department of Biomedical Research of the Royal Tropical Institute in Amsterdam, the Netherlands, on dry ice IgG antibodies against Leptospira were measured in all samples Ninetytwo randomly selected sera, 37 from Ham Kiem and 55 from Ham Hiep were also tested for IgM antibodies ª 2006 Blackwell Publishing Ltd Statistical analysis was performed using the software package s-plus 2000 Professional (Release 2, Mathsoft Inc., Seattle, WA, USA) Overall, age- and gender-specific seroprevalence was calculated with 95% confidence intervals A binary regression model was used as basic model We calculated the annual incidence using binary regression with a complementary log–log link function The association between risk factors and seroprevalence was studied using univariate and multivariate binary regression models a Z 0 zaị ẳ exp@ ka ịda A 1ị Equation describes the relationship between k, the incidence, and the prevalence of antibodies to Leptospira, 739 Tropical Medicine and International Health volume 11 no pp 738–745 may 2006 K T D Thai et al Seroepidemiology of leptospirosis in southern Vietnamese children z, at age a Given the age and the prevalence of antibodies to leptospirosis, the incidence can be estimated assuming a binomial distribution for z In using this Equation we assumed that leptospirosis is endemic; that all children were born and raised in the same region with a constant overall risk of exposure; that antibodies remain detectable lifelong and that case fatality rates not increase over time or by age For univariate comparisons of frequencies chi-square tests were also used All tests were carried out at a significance level of 0.05 (IgM > 20 U/ml) and four had IgM concentrations between and 20 U/ml IgG against Leptospira was not detectable in these nine (9.8%, 95% CI: 5.2–17.6) cases and so they were classified as acute primary leptospirosis Risk factors Leptospira IgG seroprevalence was significantly higher in boys (86, 16.7%) compared with girls (n ¼ 37, 8.3%; P < 0.001) and in children older than years In the univariate binary regression analysis, age, gender, having waded (or regularly wading) through water and swimming in rivers and canals were significant predictors of the presence of Leptospira IgG (Table 1) Children who waded through water were 1.793 times more likely to have Leptospira antibodies Swimming in rivers was associated with a higher probability of having Leptospira IgG (RR ¼ 1.745) Using variables that were significantly associated with Leptospira IgG, a multivariate model was produced Having ‘waded through water’ and ‘swimming in rivers’ were strongly associated, and because these risk factors were probably not distinguished by the respondents, one of these items could freely be omitted from the final model The final model estimated the increase of the seroprevalence with increasing age in which gender and ‘having waded through water’ remained included Boys reported exposure to fresh water bodies much more often than girls [35 of 438 (20.3%) boys vs 105 of 515 (7.9%) girls; P < 0.001] but the results in Table shows that gender and exposure to fresh water were also independent risk factors Results Seroprevalence and incidence Of the 961 children tested for serum Leptospira IgG, 123 (12.8%) had probably (n ¼ 79, 8.2%, 95% CI: 6.7– 10.1%) or certainly (n ¼ 44, 4.6%, 95% CI: 3.4–6.1%) been exposed to Leptospira About 838 (87.2%, 95% CI: 84.9–89.2) children had no detectable IgG Leptospira antibodies Figure shows the seroprevalence of Leptospira IgG, probable and definite, stratified by age The seroprevalence of Leptospira IgG increased from 11% in 7-year-old children to 25% in children of 12 years old The prevalence in age groups of 13 and 14 years was difficult to interpret due to the small number of children in these groups Overall the seroprevalence increased significantly with 1.47% (95% CI: 1.231–1.751) per year Assuming lifelong presence of IgG antibodies, this value corresponds to the annual incidence rate of first infections Of the 92 randomly selected samples and tested for serum IgM Leptospira antibodies, five were definitely positive Girls (pos/total)a 7/56 Boys (pos/total)a 7/76 8/89 6/95 8/103 14/92 8/96 5/79 22/106 25/103 1/14 10/30 0/5 0/3 1/12 1/2 Seroprevalence (%) 50 40 30 25.0 14.9 10 740 20.0 20 10.6 16.5 11.3 7.6 5.9 10 11 Age (years) 12 13 14 Figure Seroprevalence of Leptospira immunoglobulin (Ig)G in primary school children in two villages in southern Vietnam, by age The error bars indicate the 95% confidence intervals The broken line indicates the estimated annual incidence The table indicates the number of examined girls and boys per age group aNumber of positive cases/total number of children tested ª 2006 Blackwell Publishing Ltd Tropical Medicine and International Health volume 11 no pp 738–745 may 2006 K T D Thai et al Seroepidemiology of leptospirosis in southern Vietnamese children Table Risk factors associated with serum Leptospira IgG antibodies in primary school children in southern Vietnam Frequency Variable Age (approximate annual incidence) Gender Peri-domestic animals Chicken Oxen Pigs Water source River water Tap water Well water Water tanks Water and soil contact Living near rice field Bare feet contact with mud Waded through river Swam/washed in river Living near market Peri-domestic hygiene Bromeliads Coconut husk Discarded cans Plant saucers Littering plastics Sanitary Toilet Vector-borne transmission Use of bed net Bed net recently re-impregnated Spraying of residual insecticides Modern convenience Electricity Television Univariate binary regression Multivariate binary regression Total (n) IgG-positive (%) RR 95% CI P-value RR 95% CI P-value 961 12.8 1.470 1.446 1.231–1.751 1.194–1.705