The aim of this study was to evaluate the long-term relationship between the self-perception of health and survival data in an Italian population. The study was a cross-sectional survey performed in 1978-1982; survival data were obtained from the municipal administration register 10 years later. Subjects were selected in the city of Genoa (NW of Italy) through a preventive bio-medical survey. Baseline data involved 433 Italian subjects 60+ years old (34.4% males, age range 60-92), living at home, with adequate/ good physical and psychological autonomy. Self-perception of health and symptomatology were recorded using standardized items. Self-perception of health, scored from 0 (healthy) to 4 (very sick), was closely related to the survival status (UML-LOGIST: beta = 0.72 +/- 0.25, aOR (exp beta) = 2.06, CI(95%) = 1.27-3.30, chi 2 = 8.70, p = 0.0032), controlling for the effect of age, gender, years of schooling, income, doctor's appraisal. Symptomatology was evaluated through 30 close items; muscular strength and motility (beta = 0.51 +/- 0.31, p = 0.1), psychic troubles (beta = 0.47 +/- 0.24, p = 0.05), articular pain and rachialgias (beta = -0.62 +/- 0.25, p = 0.01), neuralgia (beta = -0.76 +/- 0.37, p = 0.04), and dyspeptic disorders (beta = -0.74 +/- 0.32, p = 0.02), were the symptomatologic conditions linked to survival levels. These findings confirm the importance of the self-perception of health as a predictor of survival status also in an Italian sample. To explain the predictive covariance of the baseline self-perception of health versus mortality, the decreased variability in the adaptive processes during the transition from psychophysiological to physiopathological states was hypothesized.