Summary In order to elucidate survival rates and risk of second primary cancer, we assessed 204 patients with histologically confirmed classic Kaposi's sarcoma (KS) who were identified in 11 Italian population-based cancer registries. One hundred and thirty-nine were men (median age 70 years) and 65 were women (median age 72). One, 5 and 10 year survival rates were 0.92, 0.69 and 0.46 respectively. Median survival was 9.4 years (i.e. not different from the Italian general population of the same sex and age (Tappero et al., 1993). Followup studies of patients with non-AIDS-associated KS are, however, few (Templeton and Bhana, 1975;Safai et al., 1980;Garcia et al., 1989;Biggar et al., 1994;Brambilla et al., 1994). An excess of second primary malignancies, particularly non-Hodgkin's lymphomas (NHLs), has been reported (Safai et al., 1980), but this seems to be substantially less marked in classic and African-type KS than in HIV-associated KS (Dictor and Attewell, 1988;Garcia et al., 1989;Biggar et al., 1994;Stein et al., 1994).
Materials and methodsIn order to elucidate survival rates and risk of second primary cancer, we assessed a series of 204 patients with classic KS that was identified in the context of a casecontrol study on risk factors for KS onset . Eligible case subjects were aged 50 years or older and had histologically confirmed classic KS in the period 1976-91, according to 1 of 11 population-based Italian cancer registries. To exclude HIV-related KS, several checks were made, including assessment of medical records and death certificates and anonymous linkage of KS cases with the mandatory records of all AIDS cases in Italy . This led to the exclusion of 2 out of 206 initial KS patients. Occurrence of death or second primary cancer(s) was assessed using the same cancer registry data during a total of 1409 person-years of follow-up.We computed both observed Kaplan-Meier and relative survival curves (Parkin and Hakulinen, 1991). Annual expected probability of death and life expectancy for the general Italian population of the same sex and age groups were used for this purpose. Hazard ratios (HRs) of death and corresponding 95% confidence intervals (CIs) according to sex, age group and disease site(s) were determined by means of the Cox method.
ResultsOf a total of 204 KS patients, 139 were men (median age 70 years, range 50-90) and 65 women (median age 72 years, range 50-93) ( (Figure 1). On account of the higher average age of the study population, relative survival rates were close to unity, i.e. 0.97 (s.e. = 0.02), 0.89 (s.e. = 0.04) and 0.83 (s.e. = 0.08).The observed survival curve in women was superimposable on the one in men (death HR= 1.08, 95% CI = 0.70-1.68) KS patients aged 65-74 years and 75 years or more showed significantly lower survival rates than those aged [50][51][52][53][54][55][56][57][58][59][60][61][62][63][64] (death HR=3.37, and death HR=6.51, 95% CI=3.15-13.47, respectively) (data not shown). KS patients whose lesions were restricted to lower limbs had an HR of 1.02 (95% CI=...