Human Herpesvirus 8 Seroprevalence in Blood Donors and Lymphoma Patients From Different Regions of Italy

Independent epidemiologic studies have shown a strong association between Kaposi's sarcoma (KS) and infection with human herpesvirus 8 (HHV-8 or KSHV). HHV-8 genomic sequences have been found in all forms (1–5) and all histologic stages (6) of KS, with viral DNA being localized to tumor endothelial cells and spindle cells (7,8). Detection of HHV-8 DNA sequences in the peripheral blood of patients who are seropositive for human immunodeficiency virus 1 (HIV) is a strong predictor of progression to KS (9,10). The vast majority of patients with KS have antibodies against HHV-8 (11–14), with seroconversion before clinical disease (15). HHV-8 infection is apparently not common in the general population of the U.K. and the United States, but it is common in groups at high risk for KS (11–14). In addition, HHV-8 seroprevalence is higher in HIV-positive homosexual men than in HIV-positive patients with hemophilia (13), just as acquired immunodeficiency syndrome (AIDS)-related KS occurs far more commonly in homosexual men than in patients with hemophilia. HHV-8 infection is common in the general population of Uganda, which has a high incidence of endemic KS (11,13,14). The classic form of KS affects mainly elderly men in Mediterranean and Eastern Europe (16). If HHV-8 is the major etiologic agent in KS, a higher prevalence of antibodies against HHV-8 would be expected in regions with an elevated incidence of KS. KS is prevalent in Italy, especially in the South, Sardinia, and the lower Po valley (Northern Italy) (16–18). From 1976 through 1984, before the era of AIDS, elevated incidence rates of KS were found in nine Italian cancer registries; these incidence rates were twofold to threefold higher than those in the United States and Sweden and 10-fold higher than those in the U.K. (17–20) (Fig. 1, A, and data not shown). Incidence of Kaposi's sarcoma (KS) and prevalence of serum antibodies against human herpesvirus 8 (HHV-8) latent nuclear antigen (LNA-1). A) Incidence of KS in both sexes per 100 000 per year in Italy, the U.K., and the United States [from (17,19,20)]. B) Prevalence of antibodies against HHV-8 in blood donors from Italy, the U.K., and the United States. C) Geometric mean titers of antibodies against HHV-8 LNA-1 in Italy, the U.K., and the United States. We investigated the prevalence of antibodies against the HHV-8 latent nuclear antigen (LNA-1) by use of an immunofluorescence assay (13) in 747 blood donors from different regions of Italy. Since HHV-8 is also associated with primary effusion lymphoma (21) and multicentric Castleman's disease (22), we additionally investigated the seroprevalence of this lymphotropic herpesvirus in 163 lymphoma patients from the same Italian regions; 78 of these patients had Hodgkin's disease, and 85 had non-Hodgkin's lymphomas (64 of B-cell origin and 21 of T-cell origin). The immunofluorescence assay is the most sensitive and specific assay currently available to detect HHV-8 infection (23). Sera were tested in a blinded manner at a dilution of 1 : 100, and 38 positive sera were also titrated to 1 : 400 000 to enable calculation of a geometric mean titer of antibodies against HHV-8. The results of this serologic survey are summarized in Table 1. The prevalence of antibodies against HHV-8 LNA-1 in the healthy population from Italy was 13.8% (male-to-female ratio = 1.7 : 1; median age = 39 years, range = 21–62 years). This prevalence was significantly higher than that previously reported for blood donors in the U.K. (2.7%) (13) and the United States (1.4%) (12) by use of similar assays (03C72 = 14.71; two-sided P<.001; difference between Italy and the U.K. = 11.1%; 95% confidence interval [CI] = 7.5%–14.7%) (11–13) (Fig. 1, B). HHV-8 seroprevalence was higher in Southern Italy (24.6%) than in North/Central Italy (7.3%) (χ2 = 44.39; two-sided P<.001; difference = 17.3%; 95% CI = 11.8%–22.9%), with striking differences within these regions. In Northern Italy, HHV-8 seroprevalence was more than threefold higher in the Po valley (Modena and Piacenza) than in Milan and Trieste (χ2 = 10.5; two-sided P = .001; difference = 9.1%; 95% CI = 3.1%–15.2%). The geometric mean titer of antibodies against HHV-8 (Fig. 1, C) in seropositive Italian blood donors (n = 38) was 1819 (range = 100–400 000) compared with a titer of 617 (range = 200–6400) in U.S. blood donors (n = 8) (Whitby D, Rabkin CS: unpublished findings) and a titer of 119 (range = 100–200) in U.K. blood donors (n = 4) (Whitby D: unpublished findings). The geometric mean titer for Northern Italy was 455 (range = 100–12 800) (n = 16); for Southern Italy (including Sicily), it was 5204 (range = 100–400 000) (n = 20). In Sicily alone, the geometric mean titer of antibodies was 11 631 (range = 100–400 000) (n = 7). The geometric mean titer for Southern Italy, excluding Sicily, was 3375 (range = 100–102 400) (n = 13). The high prevalence of antibodies against HHV-8 LNA-1 in the healthy population from Italy supports the strong association between infection with HHV-8 and the development of KS. The differences in HHV-8 seroprevalence between the Italian regions mirror the incidence rates of KS, being the highest in Southern Italy and in the Po valley region of Northern Italy. These serologic findings may, in part, also explain the relatively frequent detection of HHV-8 genomic sequences in the semen (24) as well as in the peripheral blood (25) of healthy subjects from certain regions of Italy but not from the U.K. or the United States (26–29). The higher geometric mean titer of antibodies against HHV-8 LNA-1 observed in blood donors in Southern Italy, especially in Sicily, also parallels the higher incidence of KS in these regions. The male-to-female ratio of HHV-8 seroprevalence is similar to the observed ratio in the incidence of classic KS in Italy (approximately 3 : 1) (16). The higher prevalence of anti-HHV-8 antibodies in patients with Hodgkin's disease (Table 1) could be related to the T-cell immune impairment and herpesvirus reactivation that are characteristic of this disease. However, based on this relatively small sample, HHV-8 seroprevalence in patients with Hodgkin's disease and non-Hodgkin's lymphoma is not significantly higher than that found in the healthy population from the same regions (χ2 = 2.29; two-sided P>.1). Our serologic findings are consistent with the infrequent occurrence of HHV-8 genomic sequences in lymphoproliferative disorder biopsy specimens obtained from patients in Italy (30,31). In summary, there is a striking association between the high detection rate of serum anti-HHV-8 LNA-1 antibodies in regions of Italy and the higher incidence rate of KS in Italy than in Northern Europe and the United States.