Do fewer cases of Kaposi’s sarcoma in HIV-infected patients reflect a decrease in HHV8 seroprevalence?