The relationship of sperm counts to birth rates: a population based study.

Abstract

We determined if a statistical relationship exists between changes in sperm counts and birth rates by comparing data from a single geographic location for a 24-year period. We retrospectively analyzed data from 660 men who banked 1,972 semen samples before vasectomy in Minnesota from 1971 to 1994. Using general linear models, annual variations in sperm count were determined after adjusting for age, duration of abstinence and seasonal (monthly) effects. Adjusted annual mean sperm count was then correlated with regional birth rate data obtained from The National Center for Health Statistics. Multiple regression analysis revealed a significant linear increase in mean annual sperm count at an estimated rate of 1.03 x 10(6) sperm per ml. per year (b = 0.14, t = 5.641, p < 0.0001). There was no effect of age (t = -0.814, p = 0.4156) but there were significant effects of abstinence (b = 0.14, t = 8.808, p < 0.0001) and month of sperm banking (b = 0.025, t = 5.00, p < 0.0001) on sperm counts. Using analysis of covariance there was a significant, nonlinear (year-to-year) fluctuation in mean sperm counts (F = 8.63, p < 0.001). For the study period mean birth rates in Minnesota (live births per 1,000 population) fluctuated yearly from 13.8 in 1973 to 16.7 in 1981. There was a strong correlation between adjusted mean yearly sperm count and annual birth rates (r = 0.63, p = 0.001). We found a statistically significant correlation between yearly variations in mean sperm counts and birth rates. Our data suggest that variations in male reproductive function may affect population based birth rates and, therefore, may be more important than previously understood.