Geographic variation in human herpesvirus 8 seroprevalence and possible association with exposure to bites from blood-sucking arthropods.

To the Editor—We read with interest the recent brief report by Kazanji et al. reporting a high seroprevalence for Kaposi sarcoma (KS)–associated herpesvirus, also known as “human herpesvirus 8” (HHV8), in Amerindians of various tribes in French Guiana as well as evidence of broad within-tribe and within-country variations [1]. Seroprevalence rates were higher in the Wayampi and Wayana tribes living in the remote and isolated internal villages of the Haut Maroni River area (Antekoumepata) and the Haut Oyapock River area (Camopi) than in the Palikour and Emerillons tribes living in the coastal region (overall, 30% vs. ∼14%). The reasons for such geographic/ethnic differences among Amerindian communities remain uncertain; however, the higher rates in the Wayampi and Wayana tribes have been attributed to the high rate of intrafamilial transmission via saliva in populations living in the most-isolated and less-developed villages. The prevalence of HHV-8 infection in the general populations of different countries shows substantial and unexplained variation according to geographic location, even across small areas [2]. This has led many investigators to propose the hypothesis that 1 or more as-yet-unknown environmental cofactors associated with the territory are related to HHV-8 infection and KS, including birth/residence in proximity to rivers [2], limited access to clean water [3], and exposure to bites from blood-sucking insects [4]. On the basis of these studies—and on the basis of the premise that HHV-8 is transmitted by saliva, mainly during childhood—we have suggested that certain species of bloodsucking arthropods would promote HHV8 transmission at the bite site when HHV8–infected relatives rub their own saliva on the sites of the bite to relieve a child’s scratching [5]. Further support for the potential role played by biting arthropods has been provided by evidence of a reduction in HHV-8 seroprevalence related to the larvicidal campaign against Anopheles labranchiae in Sardinia [6] and of an uneven distribution of mosquito species among Po River valley areas with high and those with low incidences of classic KS and HHV-8 seroprevalence [7]. The promoterarthropod hypothesis rules out a role for biological and/or mechanical vectors. More-aggressive species—those that bite mainly outdoors and that are not primarily anthropophilic (and, therefore, are less adapted to humans)—could have the greatest potential to be promoter insects, because they elicit strong skin reactions. They include, among others, Culicinae mosquitoes (Aedes vexans and Ochlerotatus caspius), sand flies (Phlebotomus species), black flies (Simulium species), and biting midges (Culicoides and Leptoconops species). In French Guiana, an uneven geographic distribution of a mosquito-borne disease (Mayaro virus fever) has been reported between internal and coastal areas [8] that is similar to the variation reported for HHV-8 seroprevalence [1]. The Haut Maroni and Haut Oyapock areas in which Kazanji et al. found the highest HHV-8 seroprevalence rates were also the places with the highest Mayaro virus seroprevalence rates. The main vectors of Mayaro virus fever are Haemagogus species (Diptera: Culicidae), especially Hg. janthinomys, a canopy mosquito of the rain forest that is more abundant in the internal areas than in the coastal areas of French Guyana. Being capable of feeding on various wild and domestic animals, Hg. janthinomys is not primarily anthropophilic [9]; therefore, it could be a candidate as a promoter arthropod in this geographic milieu. The rain forest habitat is also shared by several other mosquitoes with similar eclectic biting behavior, such as Sabethes chloropterus, an aggressive human biter. Considering that some of the samples for Kazanji et al.’s HHV-8 survey were from the arbovirus laboratory collection of the Institut Pasteur de Guyane [1], it would be of interest to know whether there is any association between the seroprevalence for HHV-8 and that for arboviruses. Because the geographic distribution of arbovirus seroprevalence can be considered a marker of the distribution of vectors, a putative association could support the promoter-arthropod hypothesis. It is possible that other human viruses could use the same mechanism of promoter-arthropod transmission; in this respect, it is interesting to note that a correlation between hepatitis B virus seroprevalence and arbovirus seroprevalence has been documented in the valley of Ribeira in Brazil [10].