Mortality selection in the first three months of life and survival in the following thirty-three months in rural Veneto (North-East Italy) from 1816 to 1835

Background: A number of studies have examined the influence of life conditions in infancy (and pregnancy) on mortality risks in adulthood or old age. For those individuals who survived difficult life conditions, some scholars have found a prevalence of positive selection (relatively low mortality within the population), while others have observed the prevalence of a so-called scar-effect (relatively high mortality within the population). Objective: Using micro-data characterized by broad internal mortality differences before the demographic transition (seven parishes within the region of Veneto, North-East Italy, 1816-35), we aim to understand whether children who survived high mortality risks during the first three months of life (early infant mortality) had a higher or a lower probability of surviving during the following 33 months (late infant mortality). Methods: Using a Cox regression, we model the risk of dying during the period of 3-35 months of age, considering mortality level survived at age 0-2 months of age as the main explanatory variable. Results: We show that positive selection prevailed. For cohorts who survived very severe early mortality selection (q0–2>400‰, , where the subscripts are months of age), mortality hazard of death during the following 33 months was 20%-30% lower compared to the cohorts where early mortality selection was relatively small (q0-2<200‰). Conclusions: This result points to a homeostatic mechanism: mortality variability among the cohorts is, for q0–35, half that of the mortality variability for both q0–2 and q3–35.