DECREASE OF HUMAN T-CELL LYMPHOTROPIC VIRUS TYPE I PREVALENCE AND LOW INCIDENCE AMONG PREGNANT WOMEN FROM A HIGH ENDEMIC ETHNIC GROUP IN FRENCH GUIANA Patricia TORTEVOYE 1 , Philippe TUPPIN 1,2 , Christian PENEAU 3 , Gabriel CARLES 3 and Antoine GESSAIN 1,4 * 1 Unite ´ d’Epide ´miologie des Virus Oncoge `nes, Institut Pasteur, Paris, France 2 Etablissement franc ¸ais des Greffes, Paris, France 3 Centre Hospitalier Andre ´ Bouron, Saint Laurent du Maroni, Guyane Franc ¸aise, France 4 Unite ´ d’Oncologie Virale, Institut Pasteur, Paris, France To assess the prevalence and incidence of human T-cell lymphotropic virus type I (HTLV-I), 4,234 pregnant women of different ethnic origins were tested before each delivery between 1991 and 1997 in a high HTLV-I endemic area of French Guiana. HTLV-I was significantly more prevalent among ethnic groups of African descent as the Noir-Marrons (4.8%, 95% confidence interval [CI]: 4.0 –5.6) and Haitians (5%, 95% CI 1.6 – 8.4). An age dependence of HTLV-I sero- prevalence was observed. The mean age of Noir-Marron HTLV-I seronegative women was lower than for HTLV-I seropositive women (24.7 vs. 28.6, p < 0.001). A decline in HTLV-I seroprevalence was observed, particularly in the Noir-Marron younger than 21 years old (p  0.04). For five HTLV-I seroconversions observed, the incidence per 100 women-years in the Noir-Marron group was 0.19 (95% CI 0.02– 0.35) for all women, 0.32 in those 25 years old or younger (95% CI 0 – 0.64), and 0.07 in those older than 25 years (95% CI 0 – 0.2). This observation was inconsistent with HTLV-I seroprevalence observed for those 25 years old or younger (2.8%) and those older than 25 (8.3%). These data demonstrate, for the first time outside Japan, a birth cohort effect for HTLV-I in a highly endemic ethnic group. Int. J. Can- cer 87:534 –538, 2000. © 2000 Wiley-Liss, Inc. Human T-cell lymphotropic virus type I (HTLV-I) is the etio- logical agent of adult T-cell leukemia (ATL) and of a chronic progressive neuro-myelopathy, the tropical spastic paraparesis/ HTLV-I–associated myelopathy (TSP/HAM). It is transmitted by sexual contact, with the most efficient transmission being from male to female, by transfusion with infected cellular lymphoid blood products, and by breast-feeding from mother to child. This human retrovirus is endemic, with foci of high seroprevalence (2% in adults) in Southern Japan, sub-Saharan Africa, the Ca- ribbean basin, and neighboring mainland America (Mueller, 1991). HTLV-I seroprevalence increases with age and is higher in older women than in older men. This has led to several hypotheses: (1) a birth cohort effect linked with lifestyle modifications influencing horizontal and vertical transmissions, (2) more efficient sexual transmission from male to female, (3) more frequent blood trans- fusions for females, (4) delayed seroconversion due to age-asso- ciated immune dysfunction (Stuver et al., 1993; Takesaki et al., 1995). The birth cohort effect has been studied particularly in Japan where a declining tendency of HTLV-I seroprevalence was observed among blood donors and young women (Takesaki et al., 1995; Ueda et al., 1989; Yamaguchi et al., 1992). HTLV-I inci- dence rates have been reported mainly for high-risk groups such as prostitutes and sexually transmitted disease (STD) patients (Dela- porte et al., 1995; Figueroa et al., 1997). A 6-year study among pregnant women attending for deliveries at the hospital of Saint Laurent du Maroni (French Guiana) was conducted to study the trends of prevalence and the incidence rates of HTLV-I. In this area live the Noir-Marrons, an ethnic group of African origin with only recent modifications of lifestyle. They have one of the highest HTLV-I seroprevalence rates in the world and a high incidence of ATL (Gerard et al., 1995; Plancoulaine et al., 1998; Tuppin et al., 1995). MATERIAL AND METHODS Area and study population French Guiana is an overseas French department in the Ama- zonian forest complex, on the northeast of the South America continent between Brazil to the east and south and Surinam to the west. Its population was estimated to be 115,000 in the 1990 census, with a large variety of ethnic groups. The Saint Laurent du Maroni area, with about 20,000 inhabitants, is located close to the mouth of the Maroni river bordering Surinam. The only regional public gynecology and obstetrics unit is located at the only hospital (Andre ´ Bouron) in Saint Laurent, although there are centers for prenatal care in the larger villages around Saint Laurent. From July 1, 1991 to June 30, 1997, all hospital deliveries were enrolled. Each delivery was registered with maternal sociodemographic and gynecological data obtained from the medical files. The results of maternal serological tests performed with the women’s informed consent before delivery were available from the medical files. Laboratory assays Serum samples were screened for HTLV-I at the hospital lab- oratory by ELISA (Cobas Core anti-HTLV-I/II EIA; Roche, Basel Switzerland). All samples giving positive or borderline ELISA results were retested by Western blot (WB; Cambridge-Biotech, Worcester, MA). A WB was considered positive for HTLV-I when virus-specific bands corresponding to the major gag antigens p19, p24, and p53 and to the recombinant rgp21 and gp46 envelope glycoproteins were displayed. Statistical analysis For seroprevalence analysis, women were included at each delivery and the trends of seroprevalence were tested using chi- square test for trends. To examine baseline differences between ethnic groups or between seropositives and seronegatives, t-test, analysis of variance, and chi-square tests were used for univariate analysis. Linear regression analysis was used to test the trend of mean age among HTLV-I–positive and negative women. To esti- mate incidence, women-years of observation were computed indi- vidually for each woman as the time between her first and last delivery or seroconversion during the survey period. Analyses were performed using SAS statistical software (SAS, Cary, NC). RESULTS A total of 5,980 births were registered at the hospital during the 6-year survey period. There were 4,234 women: some had twins or two or more deliveries during the survey. Among them, 2,995 were Grant sponsor: Agence Nationale de Recherches sur le SIDA and COR- DET. *Correspondence to: Antoine Gessain, Unit´ e d’Oncologie Virale, Insti- tut Pasteur, 28 rue du Dr Roux, 75724 Paris cedex 15, France. Fax: (33) 1 40 61 34 65. E-mail: agessain@pasteur.fr Received 5 November 1999; Revised 2 February 2000 Int. J. Cancer: 87, 534 –538 (2000) © 2000 Wiley-Liss, Inc. Publication of the International Union Against Cancer