Human herpesvirus 8 epidemiology
2003; Lippincott Williams & Wilkins; Volume: 17; Issue: 12 Linguagem: Inglês
10.1097/00002030-200308150-00001
ISSN1473-5571
AutoresNicole H. T. M. Dukers–Muijrers, Giovanni Rezza,
Tópico(s)Parvovirus B19 Infection Studies
ResumoIntroduction Human herpesvirus 8 (HHV-8), also known as 'Kaposi's sarcoma-associated herpesvirus' (KSHV), was first identified in 1994 [1]. Its discovery is a good example of how epidemiological and molecular-biological data can be combined to evaluate causality for a putative agent [2]. In fact, before the discovery of HHV-8, the analysis of surveillance data had suggested that, in HIV-infected individuals, there existed an infectious co-factor other than HIV that was critical to the development of Kaposi's sarcoma (KS), leading to the identification of HHV-8 in KS tissues [3]. HHV-8 is now considered to be the necessary cause of all variants of KS, including AIDS KS, classic KS, endemic KS and iatrogenic KS, and is a determinant of other rare diseases, such as primary effusion lymphoma, or body cavity-based lymphoma and multicentric Castleman's disease [4,5]. However, various aspects of the epidemiology of HHV-8 infection still need to be clearly defined: in particular, the evolution of the HHV-8 epidemic, the appropriate serological assays and their application, the distinct distribution features of HHV-8 in relation to KS, transmission modalities, and the natural history of infection. In the present text, we provide a synopsis of what we do and do not know about the epidemiology of HHV-8 infection, especially with regard to Africa. It is therefore not intended to provide an exhaustive review but to highlight the current thinking about HHV-8, as guided by the pioneering research performed to date. Molecular epidemiology HHV-8 is an ancient virus. The HHV-8 genome shows variation almost exclusively in the K1 gene. Phylogenetic analysis has revealed several distinct subtypes, which are believed to have diverged at least 10 000 years ago. Based on variation in open reading frame (ORF) K1, the first subtypes to be identified were A, B, C, and D, followed by subtypes E and N [6,7]. Subtypes A and C are genetically less distant from each other than they are from subtype B. Within subtype C, at least two distinct subgroups have been identified [8]. In addition, variations were identified at the right end of the genome in ORF K15 that reveal evidence for recombination and distinction between two diverged types with variability that is not linked to ORF K1 [9]. The geographical distribution of predominant subtypes is shown in Fig. 1. Subtypes A and C predominate in Europe, whereas subtype B predominates in Africa [8,10]. Subtype D is rare and has been found in individuals of Polynesian and Australian aboriginal origin with classic KS. Subtype E has been reported to be hyperendemic in Brazilian Amerindians [11]. Subtype N, a novel subtype, has been identified only in South Africa [12].Fig. 1.: Epidemiological patterns of human herpesvirus 8 infection. Countries were allocated to specific patterns on the basis of human herpesvirus 8 (HHV-8) prevalence data. For Mediterranean countries such as Greece, which lacks seroprevalence data, or Egypt, where data are available only for the city of Alexandria, allocation was based on patterns of Kaposi's sarcoma (KS) occurrence. Seroprevalence data for Albania are still unpublished (G.R. personal observation). The letters in the figure refer to HHV-8 genotype.The human herpesvirus 8 epidemic versus the AIDS epidemic In Europe, HHV-8 subtypes do not cluster within specific geographical areas, and types A and C were present in Europe long before the advent of the AIDS epidemic [8]. The various subtypes may have co-evolved with certain human populations [13]. KS was endemic in South Africa even before the advent of the AIDS epidemic [14]. In this and other African regions of KS endemicity, the onset of AIDS has led to an increased prevalence of KS. Similarly, among the subgroup of homosexual men in industrialized countries, HHV-8 has caused numerous cases of KS, once the HIV epidemic has led to widespread immunosuppression. A virus needs a route of transmission to sustain its presence after introduction, and such a route was apparently present, for example among these men. Some studies conducted among homosexual men in north America and Europe [15,16] suggested that there existed concurrent epidemics of HHV-8 and HIV infections and that the two infections have modalities of transmission in common. However, other studies suggested that the prevalence of HHV-8 infection was already high among homosexual men when the HIV epidemic began. Although some initial increases in the prevalence of HHV-8 infection were reported among homosexual men at the start of the HIV epidemic in the early 1980s, the prevalence has since remained at a nearly constant level [17–19]. The discrepancy in prevalence trends over time between HIV and HHV-8 indicate that there must be different or at least additional transmission routes for HHV-8 when compared with HIV. Serological assays Currently used serological assays The currently used serological assays are: Western blot, immunofluorescence assays or enzyme immunoassays. In general, these assays are based on latent antigens expressed by ORF 73, lytic antigens expressed by ORF 65, or both, but their accuracy (i.e. sensitivity and specificity) has not been established. There is no gold standard for HHV-8 because the virus cannot be reliably cultured. Currently, negative reference sera are obtained from individuals at a low risk of KS, such as women without KS in northern Europe, whereas positive reference sera are obtained from individuals with KS. However, as individuals with KS generally have higher HHV-8 titres than individuals without KS, the estimated sensitivity may likewise represent an overestimate [20], and as one can never be sure that individuals at low risk of KS are not infected with HHV-8, the estimated specificity may also represent an overestimate. As immunofluorescence assays that detect lytic antigens are generally more sensitive than other assays, their use results in higher estimates of prevalence [21]. In addition, assays based on antigens expressed by ORF K8.1 have relatively high specificity and sensitivity [22]. The existing serological assays nonetheless need to be further refined, and new assays need to be developed. The combining of assays in screening algorithms, as well as the development of new assays, is an area of continuing progress [23–25]. The application of serological assays in epidemiological studies The development of serological assays capable of detecting antibodies against HHV-8 has paved the way for large-scale epidemiological studies, which have provided important information on HHV-8 infection. For example, several cohort studies on HIV and AIDS, in which stored sera were screened for HHV-8 antibodies, have allowed the temporal relationship to be established between HHV-8 infections and KS occurrence in HIV-infected individuals [26–28]. However, the various serological assays are still imperfect, resulting in the misclassification of HHV-8 test results, which influences the results of epidemiological studies. This shortcoming probably accounts for a substantial part of the differences in prevalence rates among studies. However, although absolute prevalence estimates cannot be generated, the distribution of HHV-8 infection among HIV risk groups and geographical areas has been quite comparable in most studies (see the following sections on the distribution of HHV-8 infection). In high-prevalence settings, suboptimal specificity and sensitivity will most likely cause the magnitude of the associations to be underestimated, at least when the misclassification of the outcome is unrelated to the risk factors of interest. In low-prevalence settings, particular caution is recommended, because the number of false positives can exceed the number of true cases, leading not only to the underestimation of risks but also to spurious associations when comparing seronegative with seropositive individuals. The application of serological assays with an optimal level of specificity is thus crucial in such low-prevalence settings as the African countries of Gambia and Botswana [24,29]. Table 1 illustrates the effect of suboptimal serological assays in terms of the positive predictive value and the rate of false positivity.Table 1: Examples of the variation in positive predictive value in generation of false positive cases caused by different sub-optimal serological assays in low prevalence settings (2.5%) and intermediate/high prevalence settings (25%).The distribution of human herpesvirus 8 infection in relation to Kaposi's sarcoma The distribution of human herpesvirus 8 infection by HIV exposure category Among HIV exposure categories at risk of AIDS KS in north America and Europe, the prevalence of HHV-8 shows a distinct spread. Within a defined geographical area, this prevalence is highest among HIV-infected homosexual men, lower among HIV-infected injecting drug users (IDU) [21,30–35] or HIV-infected non-IDU heterosexual individuals, and among children the prevalence of HHV-8 is the lowest [21,36,37]. Interestingly, in one US region, the HHV-8 prevalence rates among young homosexual men (i.e. 15–22 years of age) were comparable with the rates among young heterosexual men, and both were noticeably lower than rates among older homosexual men in that region [38,39]. Overall, the distribution of HHV-8 infection between HIV risk groups mirrors the pattern of KS occurrence in HIV-infected individuals. Most cases of KS occur among HIV-infected homosexual men, for whom KS has been one of the most common AIDS-defining illnesses, representing 22–29% of such illnesses before the introduction of highly active antiretroviral therapy (HAART) [40,41]. Among HIV-infected IDU and non-IDU heterosexual individuals KS is very rare. The geographical distribution of human herpesvirus 8 infection Although the geographical distribution of HHV-8 infection in the general population has not been completely defined, it is similar to that of KS [42], with a high prevalence in sub-Saharan Africa, relatively high or intermediate prevalence in southern Italy and other Mediterranean areas, and very low prevalence in northern Europe and the United States (Fig. 1). However, these broad areas show variations within that need to be addressed. For example, in Italy, the prevalence varies between less than 10% in the North to more than 20% in the South [36], with the highest rates observed on the islands of Sicily and Sardinia [43]. Wide variations have been observed even within smaller areas, such as northern Italy, where the prevalence ranges from less than 5% in the city of Milan to more than 20% at the lower end of the Po Valley [36,44]. The highest prevalence in the Mediterranean area has been reported in Alexandria, Egypt [45]. A surprisingly low prevalence has been found in Spain and in various Israeli Jewish subpopulations [46,47], although these subpopulations are considered to have a relatively high risk of KS. Prevalence rates are similarly low in Saudi Arabia [48]. The lowest HHV-8 prevalence has been reported in northern Europe and the United States [49], where KS accounts for only 0.02% of all malignancies [50]. In these areas, HHV-8 infection is mostly concentrated among homosexual men at risk of HIV infection [41]. The HHV-8 prevalence is low or even very low in many countries of southeast Asia, in Japan, and in Central and Latin America, where KS is very uncommon [49,51–54]. A high prevalence has been observed among indigenous populations living in remote tribes of Amazonia and Papua New Guinea [11,55]. Geographical areas endemic for Kaposi's sarcoma: the case of Africa The highest prevalence of KS has been reported in sub-Saharan Africa, where endemic KS has been reported to occur. In this area, the HHV-8 seroprevalence among children ranges from 13% in Yaounde, Cameroon, to 37% in Uganda [56,57], and from 36% in Ghana to 47% in Zambia [58,59]. Prevalence among adults ranges from 22% in central Africa to 48% in Zambia [60,61]. Very high rates, between 76 and 87%, have been reported in the Congo and in Botswana (87%) [24], and among older adults in Ghana (43%) and Zambia (71%) [58,59]. The view that most geographical differences in the occurrence of KS are caused by variations in the prevalence of HHV-8 infection is supported by data shown in Fig. 2. It compares HHV-8 prevalence rates reported by several studies [24,45,56–65] and the level of KS occurrence, as defined by the cumulative incidence [66] or the proportional frequencies of KS [67,68].Fig. 2.: Comparison between human herpesvirus 8 prevalence and the occurrence of Kaposi's sarcoma. Kaposi's sarcoma incidence rates per 1000 men and the proportional frequency of all cancers (%) are presented (i.e. numbers in parentheses). In one case (i.e. Cameroon) proportional frequencies in parentheses refer only to men.Ecological fallacia: spurious associations and the search for co-factors in the development of Kaposi's sarcoma The epidemiological pattern of HHV-8 infection appears to be the main determinant of the geographical variations in KS occurrence; in fact, a strong direct correlation between HHV-8 prevalence and classic KS occurrence has been shown by detailed studies in small areas and seems to be confirmed by rough comparisons of data at the global level [41,69]. However, HHV-8 prevalence may not completely explain the geographical variations in KS occurrence, given several exceptions to the pattern. In Alexandria, Egypt [45], in remote areas of the old and new worlds [11,55], and in areas of China [70], the prevalence of HHV-8 infection is high, yet the occurrence of KS is low. Discrepancies between HHV-8 and KS distribution may reflect biases. In some areas, the incidence of KS may be underestimated because of underdiagnosis or underreporting; for example, a study of pediatric KS in Papua, New Guinea [71] seemed to support this hypothesis. Moreover, geographical comparisons of HHV-8 prevalence may be affected by differences in the characteristics of the locally used assays (i.e. sensitivity and specificity), low comparability of the study populations (i.e. differences in age classes), and local variations (i.e. within-country variability). For example, the surprisingly high prevalence of HHV-8 infection found in some western African countries with no KS endemicity, such as Gambia or the Ivory Coast [21,72] might be the result of limited sample sizes or a low specificity of first-generation assays. It should be noted, however, that the development of KS was not observed in a large group of Ethiopian immigrants to Israel, although a specific Western blot found that HHV-8 was common in a subgroup of these immigrants. This discrepancy might reflect biases, as mentioned above, but they could also imply unidentified co-factors protecting the larger group from the development of KS [73]. A different distribution of unidentified co-factors might increase the risk of KS in areas with similar HHV-8 prevalence, but in that case the factors previously considered to account for KS (on the basis of ecological studies) would at least partly be discharged after controlling for HHV-8. Host factors that affect the risk of developing KS, such as age and sex, do not explain geographical differences. Ethnic-related factors, such as genetic determinants like HLA, may be influential, but no evidence of their effect after accounting for HHV-8 infection has been found. Viral factors (i.e. the circulation of different HHV-8 subtypes) are not likely to play an important role, because there is no clear evidence of differences in subtype virulence, and the co-circulation of different subtypes is common. However, the effect of viral factors needs to be more thoroughly explored. Much attention has always been paid to environmental factors. For example, KS has been associated with exposure to volcanic soils [74], and seems to be more common in highland areas located at altitudes over 2000 feet, although high incidence rates have been found at a broad range of altitudes and temperatures, in contrast to the pattern of Burkitt's lymphoma [66,67]. HHV-8 prevalence, which is high throughout the equatorial belt of Africa, probably accounts for the high incidence of KS in that area, although a residual effect on the risk of KS by other factors cannot be ruled out. HHV-8 infection is also likely to explain why being born in an area with endemic malaria may be a risk factor for KS. This association has been found in Italy [75] but only inconsistently in Africa, where many of the highest estimated KS rates have been reported in high-altitude areas in which malaria is practically non-existent [66]. Existing data do not exclude an association between HHV-8 and malaria, but do not support a role for malaria in increasing the risk of KS among HHV-8-positive individuals. A confounding factor might be the maternal practice of sucking blood from a child's mosquito bite. Whether saliva from an HHV-8-infected mother could thereby pass HHV-8 to the child's blood has yet to be determined [76]. Transmission modalities of human herpesvirus 8: have they been concretely defined? Do human herpesvirus 8 transmission modalities fit with those of other specific herpesviruses? Many studies have been performed to identify the precise mode(s) of HHV-8 transmission, but given the often contradictory results, the specific routes remain in debate. Phylogenetically, HHV-8, as Epstein–Barr virus (EBV), is positioned among the γ-herpesviruses, as summarized in Fig. 3 [77–90]. However, the transmission modalities of HHV-8 seem to be in between those of EBV, mainly transmitted through saliva, and those of herpes-virus 2, an alpha herpes virus that is mainly transmitted through sexual contact. Moreover, HHV-8 transmission modalities appear to differ between highly endemic and subendemic areas. The number of HHV-8 DNA copies is higher in saliva than in peripheral blood mononuclear cells (PBMC) or semen [85–87]. It has been suggested that HHV-8 can replicate in the oropharynx. There is scant virological evidence of genital sites of replication [88].Fig. 3.: Human herpesvirus 8 in relation to other human herpesviruses, epidemiological and phylogenetic links. EBV, Epstein–Barr virus; EHV, equine herpesvirus; HCMV, human cytomegalovirus; HHV-8, human herpesvirus 8; HSV, herpes simplex virus; PRV, pseudorabies virus; VZV, varicella zoster virus. Phylogenetic tree crudely adapted from Moore et al. [77], with permission.Although studies evaluating transmission are often inconsistent, they point to certain directions that will be addressed below. Transmission of human herpesvirus 8 through blood International studies have provided some evidence that HHV-8 can be transmitted by blood or blood products [30,91–94]. Some studies, which were performed in intermediate or high endemic populations, have shown that the prevalence is somewhat higher among IDU compared with the general population, or they have demonstrated an association between HHV-8 infection and injecting behaviour [30,33,52]. In countries or populations that are highly endemic for HHV-8, transmission by blood is not believed to be a major modality, but could play a role. In low-prevalence areas, the prevalence of HHV-8 is generally similar between IDU and the general population. Among IDU, although syringe sharing is a common means of transmitting HIV-1, hepatitis B virus and hepatitis C virus, it does result in a similar spread of HHV-8 [90]. It is likely that in low-prevalence areas transmission by blood occurs very rarely. The introduction of HHV-8 in low-prevalence populations might occur occasionally, e.g. by individuals who had had sex with homosexual or bisexual men or by individuals coming from areas that are endemic for HHV-8 infection. Non-sexual person-to-person transmission A study of HHV-8 prevalence within families [95] reported equally high rates among individuals with classic KS and their relatives, and these rates were considerably higher than those among a sample of individuals from the general population who were matched by sex and age. Findings from countries that are endemic for HHV-8 infection suggest that the virus can be transmitted among family members and that transmission is associated with close contact and crowding [95–98]. These correlates are also commonly noted for many other viruses, including hepatitis B, herpes simplex virus, and EBV. Non-sexual person-to-person transmission and childhood transmission are rare in the United States and north-central Europe, but do occur in countries where HHV-8 infection is more widespread, such as African and Mediterranean countries [45,57,64,95,96,99–101]. The high prevalence of infection before adolescence in endemic areas and the higher frequency of the detection of HHV-8 in saliva compared with semen [86] provide indirect evidence of non-sexual transmission. It is thus hypothesized that viral spread through salivary contact might be involved in childhood transmission in endemic countries, although this is not yet evidenced by studies [102]. Additional modes of non-sexual person-to-person transmission are shown by cases of seroconversion to HHV-8 after heart transplantation [103] and kidney transplantation [104–106]. These findings indicate that HHV-8 could be transmitted from the donor organ during transplantation. Mother-to-child transmission There is little evidence of the vertical transmission of HHV-8 during pregnancy. That mother-to-child (MTC) transmission occurs is suggested by the high MTC correlation found in several studies [97,101], and MTC transmission appeared to increase with increasing maternal antibody titres [62]. However, most studies have shown that HHV-8 prevalence in children increases with their age [97,107,108], and that most children born to HHV-8-positive mothers have passively transmitted antibodies and then serorevert by 24 months of age [109]. In summary, MTC vertical transmission is very unlikely to occur, although definitive evidence is still lacking. What sexual practices lead to human herpesvirus 8 infection? Among heterosexuals, correlates with sexual transmission have not been consistently established [110,111]. Heterosexual transmission is not likely to be a major route and seems to require frequent sexual exposure to occur. Prevalence seems to be elevated mainly in heterosexuals who have a fairly high number of sexual partners or a sexually transmitted disease [87,112,113]. Among homosexual men, there is consensus that the transmission of HHV-8 is sexual; thus in this population AIDS KS can be considered to be a sexually transmitted disease [16,26,114]. However, the specific mechanism remains a topic of considerable debate. Some of the early studies on the relationship between KS and sexual practices reported an association between oral–anal sex and AIDS KS [115,116]. Cross-sectional studies have demonstrated an association between the presence of HHV-8 antibodies and anal–genital sex, oral–anal sex, and deep kissing with an HIV-1-positive partner [16,86,117]. Few studies have prospectively examined the risk factors for recent HHV-8 infection, and all found evidence of sexual transmission among homosexual men, e.g. recent sexual contact with an HIV-positive partner being a risk factor for seroconversion [26,78]. Only one study related specific sexual techniques to serococonversion [114], finding recent oral-genital sex to be a strong risk factor. Longitudinal studies of viral shedding have shown that HHV-8 is not substantially shed in semen or rectal tissue, but is persistently found in the oral cavities of some individuals [86]. Oral–genital sex is a common practice among homosexual men, as is deep kissing. Therefore, if saliva is a carrier of HHV-8, it may transmit inefficiently. The risk of HHV-8 transmission through receptive ano-genital or insertive oro-anal sex may be even lower, because little or no virus has been detected in semen or faeces [118]. However, if saliva is involved in transmission, it is not clear why heterosexuals in low endemic countries have such a low prevalence compared with homosexual men. Perhaps the fact that homosexual men generally engage more often in various sexual techniques when compared with heterosexual individuals may disproportionally affect their likelihood of acquiring the virus. Considerations in interpreting epidemiological studies on transmission In spite of the bias inherent in their design, cross-sectional studies uniformly identified the sexual techniques posing the highest risk of HIV acquisition [119]. As was noted before, associations of a large-order magnitude (as observed for HIV transmission) withstand a certain degree of bias (such as that caused by misclassification) and thus can still be detected [120]. Given the discrepant results of HHV-8 studies, the associations we seek in HHV-8 transmission are probably of low magnitude. To detect subtle associations, studies should be designed to reduce bias as much as possible: by the use of optimal serological assays and prospective information and by the detailed measuring of variables. In addition to study design, there are other possible explanations for the differences in the findings on risk factors obtained in various countries and populations. In certain subpopulations, such as homosexual men, there may exist multiple routes or a certain transmission route (e.g. via saliva or blood) that is more efficient because of facilitating factors, such as higher virus titres. Moreover, the public health importance of an association must be evaluated separately for each population or geographical area. A given transmission route may account for more infections in a high-prevalence setting than in a low-prevalence setting, simply because the chance to encounter the virus is greater. What do we know about the natural history of human herpesvirus 8 infection? The natural history of HHV-8 infection is still undefined. KS, the main clinical finding associated with HHV-8 infection, occurs in a small proportion of infected individuals unless certain co-factors are present, especially HIV-related or transplant-related immunosuppression. Clinical series and longitudinal studies have provided important clues for better understanding the natural history of HHV-8 infection, as summarized in Fig. 4. We can now identify the determinant of KS development and, to a lesser extent, the development of rarer conditions such as primary effusion lymphoma and multicentric Castleman's disease. Other diseases in which HHV-8 infection has been hypothesized to play a role, such as multiple myeloma and sarcoidosis, will not be considered here because of the inconsistency of the association [121].Fig. 4.: The natural history of human herpesvirus 8 infection: clinical stages and determinants of progression. HHV-8, Human herpesvirus 8; KS, Kaposi's sarcoma; PBMC, peripheral blood mononuclear cell.Human herpesvirus 8 primary disease There are few reports of cases of acute disease attributable to primary HHV-8 infection in immunosuppressed or immunocompetent individuals. The first report concerned an HIV-infected homosexual man presenting with a severe illness consisting of the sudden onset of fever, arthralgia, cervical lympadenopathy, splenomegaly, and cytopenia [122]. Within 2 months, there was spontaneous resolution of the clinical symptoms and lymph node lesions (i.e. angiolymphoid hyperplasia, intense plasmacytosis, and foci of KS). Seroconversion to HHV-8 was estimated to have occurred from 4 months to 5 weeks before the onset of clinical symptoms [122]. Other cases of primary HHV-8 disease were reported in two kidney recipients. Four months after transplantation, they developed disseminated KS and an acute syndrome characterized by fever, splenomegaly, cytopenia, and marrow failure with plasmacytosis. Serum samples collected at the time of transplantation had been negative, but the first available post-transplantation serum sample, obtained when the symptoms began 4 months later, showed seroconversion to HHV-8 [105]. A fourth case, consisting of the multiorgan dissemination of HHV-8 in a one-month-old girl with DiGeorge anomaly [123], suggested that severe symptoms may occur in patients presenting with primary immunodeficiency. The severity of the illness observed in these patients was related mostly to the associated immunosuppression and was not what might be expected in the HHV-8 primary infection of an immunocompetent host. In a study of 108 persistently HIV-negative homosexual men who were initially negative for HHV-8 [124], five cases of primary HHV-8 infection developed and were associated with mild, non-specific signs and symptoms of diarrhoea, fatigue, localized rash, and lymphadenopathy. In a study conducted in Egypt [125], primary HHV-8 infection was associated with a febrile maculopapular skin rash among immunocompetent children, who seroconverted within 6 months after the onset of clinical symptoms. Primary symptomatic HHV-8 infection may thus occur in both immunocompetent and immunocompromised hosts, presumably with a different level of severity. The frequency with which symptoms may occur is still unknown. Risks and determinants of progression to Kaposi's sarcoma Transplant recipients The incidence of KS after organ transplantation is 500–1000 times greater than in the general population [126,127]. From 0.5 to 5% of organ transplant recipients develop KS, depending on their geographical origin and the type of organ received [128]. In a cohort of 220 transplant recipients, KS developed within 26 months in two of the 25 patients (8%) who had developed a primary HHV-8 infection after renal transplantation but in none of 14 patients who were HHV-8 positive before transplantation [91]. By contrast, an Italian retrospective study showed that 10 out of 11 organ recipients (91%) who developed KS were HHV-8 seropositive before transplantation, whereas only one had a primary HHV-8 infection [106]. These data are supported by those of another study [129], which showed that KS developed in three of the 21 recipients (14.3%) who were HHV-8 seropositive before kidney transplantation, but in none of those who seroconverted after transplantation. On balance, the few findings suggest that, at least in highly endemic areas, individuals who are already infected before transplantation are more likely to develop KS than those who become infected afterwards; they are somewhat confirmed by studies showing that HHV-8 may reactivate after organ transplantation [130,131]. HIV-infected individuals HIV-infected individuals have a risk of KS that is thousands of times higher than that of the general population. Early retrospective studies of viral DNA in the PBMC of individuals with AIDS KS showed evidence of HHV-8 infection well before the occurrence of KS [132–134]. In a study of 143 HIV-infected homosexual men, who were followed for a median of 30 months, more than 50% of those who initially had HHV-8 sequences detectable by polymerase chain reaction in the PBMC had progressed to KS within 3–5 years [135]. When HHV-8 serology became available, longitudinal studies showed that the 10-year probability of developing KS was 49.6% among HHV-8-seropositive men who were HIV-infected at baseline [26]. Later, studies of HHV-8 co-infected individuals with known dates of HIV seroconversion found an actuarial progression rate of nearly 30% after HIV seroconversion [27]. Two other cohort studies of individuals with known durations of HIV and HHV-8 infection showed that the risk of KS is higher when HHV-8 seroconversion occurs after HIV seroconversion, with risk ratios ranging between approximately 1.6 and 2.5 [28,136]. The hazard of KS from dual infection appears to increase by more than 27% for each year of infection with HIV before HHV-8 seroconversion [136]. These data suggest that KS among HIV-infected individuals is more likely to occur when an individual becomes more immunocompromised. They are supported by findings of an independent association between the progression to KS and low CD4 cell counts [134–136]. Other indicators of the development of KS among HIV/HHV-8 co-infected individuals, such as high HHV-8 titres and high HHV-8 viral loads, have also been identified [134,137]. Finally, KS seems to occur less often with HIV-2 infection than with HIV-1 infection, even when the two HIV types are accompanied by the same level of HHV-8 prevalence. This suggests that other co-factors specifically related to HIV-1, as opposed to HIV-2, are involved in the development of KS [29]. KS responds very well to HAART, and has been shown to regress during such treatment [138–141]. This is epidemiologically evidenced by strong decreases in the incidence of AIDS KS since the use of HAART became widespread [142,143]. These decreases were initially attributed to supposed decreases in HHV-8 incidence [144], but this assumption was refuted when the beneficial role of HAART on KS was explored [17]. The dramatic decline in KS rates after the introduction of HAART led to the hypothesis that protease inhibitors directly affect the skin tumour. This hypothesis has recently been confirmed by the identification of an anti-angiogenic effect of protease inhibitors in animal models [145]. General population There is little information on the risk of KS among immunocompetent HHV-8-positive individuals. This lack of data results, at least partly, from the low incidence of KS in the general population, hampering studies on the risk of KS among immunocompetent individuals. A study conducted among individuals over 50 years of age on three islands of the Mediterranean (i.e. Sardinia, Sicily, and Malta) showed yearly rates ranging from 1 : 3108 to 1 : 3574 among HHV-8-positive men and from 1 : 4866 to 1 : 10 970 among HHV-8-positive women. The male-to-female ratio ranged between 1.4 and 3.5, suggesting that sex-related factors may determine differences in the development of KS [43]. The role of sex is further supported by the observation of almost identical rates of HHV-8 infection among men and women in different areas of the world, in contrast with increasing male-to-female ratios of KS along with age, which may be as high as 10 : 1, especially in high-incidence areas. How male sex contributes to the risk of the development of KS remains undefined. Except in HHV-8 endemic areas, KS usually occurs after 50 years of age. Age is thus an established co-factor for the development of KS among HHV-8-positive individuals. Whether the early acquisition of HHV-8 may be a contributing factor to KS is unknown. The possible role of other host-related factors needs to be investigated after controlling for HHV-8 infection. These include malnutrition (especially in highly endemic areas of KS in Africa) or genetic factors that could either increase the risk of KS (i.e. some ethnic backgrounds or HLA DR5) or decrease this risk (i.e. HLA DR3) [67,146]. Finally, with regard to the immunological response, patients with classic KS are predominantly characterized by immune activation (i.e. elevated level of neopterin or β-2 microglobulin), although an element of immunosuppression cannot be excluded (i.e. lower CD4 cell counts partly caused by low lymphocyte counts) [147]. A recent case–control study of HHV-8-positive individuals with and without KS found cigarette smoking to be protective for KS, perhaps because of its effect on cytokines, but this hypothesis needs to be confirmed [148]. Whether the different HHV-8 subtypes vary in infectiousness or virulence is still unclear. It has been suggested that subtype A is more aggressive than subtypes B or C [149], but observations were based on a small number of HIV-infected individuals and require more study. Conclusion Because current HHV-8 serological assays are suboptimal, it is impossible to study HHV-8 adequately and simply in low-risk settings. Also, in high-risk settings, a reliable HHV-8 serological assay is essential, because misclassification can lead to biased results. Assay algorithms are being developed. The distribution of HHV-8 infection clearly overlaps with the distribution of KS. This does not completely exclude the possibility that other co-factors, unequally distributed throughout the different areas of the world, may explain some of the observed discrepancies. However, most of the potential determinants of KS identified before the discovery of HHV-8 are now ruled out. The issue of whether or not HHV-8 can be transmitted by blood is important because of its potential consequences for blood screening. Transmission by blood is probably very rare in countries of low endemicity, where groups at high risk of HIV (and thus of HHV-8) are already being excluded from donation. However, blood-borne transmission is a possibility that must be considered in high endemic countries. Although there is evidence that HHV-8 may be transmitted through saliva, additional studies are needed to investigate sexual transmission routes and the role of deep kissing, as well as non-sexual transmission routes, both in endemic areas and in countries with a concentrated pattern. The improved identification of transmission routes can help to minimize the risk of KS among HIV-infected individuals and among individuals at high risk of iatrogenic KS. The natural history of HHV-8 infection remains to be completely defined. Primary disease has been identified, but its frequency is unknown. Longitudinal studies have provided precise estimates of the risk of KS among HIV and HHV-8-co-infected individuals and, to a lesser extent, among transplant recipients. Little information is available on the risks and determinants of KS among apparently immunocompetent HHV-8-positive individuals. Although many advances have been made since the identification of HHV-8, many questions remain unanswered. The goals of increasing our knowledge of transmission and natural history, developing accurate and reliable serological assays, and performing sustainable intervention continue to present a challenge for various fields of research. Acknowledgements The authors would like to thank Mark Kanieff for his helpful comments on an earlier draft and Lucy Phillips for editing the final manuscript.
Referência(s)