Our objective was to study birthweight among surviving siblings in families with and without a perinatal loss, and to evaluate whether different causes of death were associated with the results. Data were for 1967-98 from the Norwegian Medical Birth Registry. Births were organised with the mother as the observation unit through the personal identification number, providing sibship files. We analysed 550 930 sibships with at least two singletons, 208 586 sibships with at least three singletons and 45 675 sibships with at least four singleton births. We compared mean birthweight and gestational age between infants in sibships with and without a perinatal loss, total losses and the different causes of death. Surviving siblings in families with a perinatal loss had significantly lower mean birthweights than their counterparts in unaffected families, after adjusting for gestational age, interpregnancy interval, time period and marital status. An exception was found when cause of death was a birth defect, when growth retardation among surviving siblings was not found on average. We conclude that families who have lost an infant because of a birth defect do not appear to have an increased risk of adverse birth outcome associated with growth restriction.