In this study, the overall analyses of 2766 pregnant women revealed a prevalence of 1.05%. Previous studies in endemic areas of Brazil have found a similar prevalence [19, 24] demonstrating that Southern Bahia is another region where the virus circulates with a prevalence much higher than in other regions of the country–at least three to ten times higher [6, 21–23]. Additionally, this prevalence was much higher than the prevalence of HIV (0.22%–CI 95%: 0.08-0.48) and Treponema pallidum (0.47%–CI 95%: 0.52-0.80) in the analyzed population. It is noteworthy that prevalence rates for these two microorganisms can still be overestimated, since they were calculated from the results of rapid test for HIV and Venereal Disease Research Laboratory (VDRL) for Treponema pallidum. In the case of HTLV, all samples were subjected to confirmatory tests.
As we are studying a specific group, the results probably do not represent the overall population. However, in a study comparing data from specific populations (injecting drug users, blood donors and pregnant women), Hlela et al., have suggested that blood donors and pregnant women in Southern America and the Caribbean may be more representative of the general population and can therefore be suitable for estimating prevalence in these regions . In fact, even though the prevalence rates in pregnant women do not represent the overall population, they are very important because the virus can be transmitted to children during pregnancy and, most importantly, during the breastfeeding process. Besides that, different clinical manifestations related to HTLV-1, such as: infective dermatitis (IDH), adult T-cell leukemia/lymphoma (ATL), HTLV-1-associated myelopathy/tropical spastic paraparesis (HAM/TSP) occur in individuals who have been vertically infected .
Accordingly, it has been suggested that the detection of HTLV infection through prenatal or neonatal screening can be fundamental in sub-areas with high seropositivity rates, permitting to take preventive measures to reduce vertical transmission . A classical study has demonstrated that the refraining of breast-feeding for HTLV-1 positive mothers has dramatically reduced vertical transmission in Japan . Nonetheless, bottle-fed children can also become vertically infected in much lower frequencies. Then, the prenatal detection is more effective for prevention, whereas additional measures such as an elective cesarean in HTLV-positive pregnants can be taken. In a previous study involving forty-one bottle-fed children from Brazil, no case of vertical transmission was observed. In this case, 81.5% of the children were born by an elective cesarean section and this fact may have contributed to the absence of transmission .
In relation of the route of transmission, it was suggested that, in Salvador, the infections have been acquired via breastfeeding, and, in second place, sexually. In this study, the analysis of HTLV-1 serology in relatives, partners and children of previous pregnancies of the index case (pregnant) has revealed HTLV-1 positive cases in different family members, highlighting partners, mothers and children (1 son and 1 daughter–2.3 and 8 years old, respectively). Therefore, it can be assumed that the virus infection in Southern Bahia can be spread both sexually and vertically. In fact, both routes of transmission have been related to HTLV in endemic areas . In addition, it is noteworthy that the two mentioned HTLV-positive children were breastfed. In this study, all of the HTLV-positive women contacted were advised not to breastfeed and the newborns were followed up to two years. Until that time, no positive case has been detected by PCR. Nonetheless, it has been argued that it is necessary to keep in mind that, in developed countries, the advice of not breastfeeding should be made carefully, because the health risk of early weaning can be higher than the risks of HTLV-1 related diseases .
Still on the familial transmission, it should be emphasized that half the evaluated families had at least a HTLV-seropositive member. Besides that, the analysis of infection rates in family members indicated a seropositive rate of 32.55% (14/43). This number is higher than that recently found in a survey evaluating familial transmission  in Pará state (25.2%), another endemic area for HTLV in Brazil. Without a doubt, the above-mentioned data reinforce the need to establish strategies of active surveillance in household and family contacts as an important epidemiological surveillance action aimed at detecting early the virus infection and preventing the transmission by sexual or parenteral way. In effect, it has been shown that the virus spreads silently within families and that there is a familial aggregation of this infection .
This study has detected an association with marital status, but it was not precise. Besides, there was no association of HTLV-1 infection with age, education, and income according to what was found by Magalhães et al. , in the analysis of pregnant women from a medium sized town in Northern Brazil, unlike the observed in other studies conducted in Salvador, which have found an association between HTLV infection and lower income [5, 17]. In the way, we have not found any association between the self-reported skin color and HTLV infection. However, it has been recently detected a higher HTLV prevalence in donors with black skin color . In fact, the two groups (HTLV-1 positive and soronegative) analyzed in our study are very similar in terms of social, demographic and ethnic characteristics, according to the population treated at public hospitals of medium-sized Brazilian cities, where the majority of people is subjected to low levels of income and education.