Familial Occurrence of Preeclampsia

Abstract

The authors reply: The aim of our retrospective cohort study was to investigate the occurrence of preeclampsia in mothers as a risk factor for preeclampsia in daughters and to estimate the magnitude of the relative risk and the population attributable proportion (PAP). 1 Since the 1960s, several hypotheses concerning possible inheritance have been proposed; however, none sufficiently describes the underlying mechanisms. The etiology of preeclampsia is still unknown and the pathophysiology remains poorly understood. Rockhill has commented on our presentation and interpretation of our findings, primarily with respect to the population attributable proportion. We agree with Rockhill that “not all familial history indicates genetic predisposition, and not all genetic predisposition is necessarily reflected in a one-generation first-degree family history.” Besides primiparity and familial occurrence of preeclampsia, however, few other exposures are known to affect preeclampsia occurrence. To our knowledge, other familial factors, such as socioeconomic status, do not influence preeclampsia occurrence. Smoking, however, has been found to be protective for the development of preeclampsia. Throughout the study we have calculated the PAP from the relative risk and the proportion exposed in the population for every specified comparison. Rockhill states that “the estimates of ‘proportion exposed’ are not clearly identifiable anywhere in this article.” They are, however, presented indirectly in the tables. In Table 2 and 3 the proportion of exposed is easily calculated through combination of data given in the table(s). For example, the number of exposed elder daughters are presented as the denominator (N = 1,313) in the calculation of the prevalence of preeclampsia at elder daughter’s first delivery (67/1,313) (Table 2). The proportion exposed among elder daughters at their first delivery can be calculated by dividing the number of exposed subjects by the total number of subjects (eg, 1,313/22,768 = 0.058 estimated as 0.055 by Rockhill). Furthermore, Rockhill questions why the proportion exposed at elder daughter’s first and second delivery differs. The reason is that at the second delivery a subpopulation of elder daughters is “selected” because parity evidently differs between the two groups. Rockhill comments that the proportions exposed “differ widely.” The proportions exposed in the second part of Table 2 represent an “aggregated risk” of preeclampsia, ie, preeclampsia during any of the mother’s pregnancies other than the targeted pregnancy. In Table 3 only women registered with a firstborn daughter or a firstborn son are included. As preeclampsia is more common in a woman’s first pregnancy resulting in a delivery, we do not find it surprising that the proportions exposed in the population were increased in comparison with the women in Table 2. The relative risks for elder daughters, however, were close to the relative risks presented in Table 2. Since a younger daughter’s deliveries were a limited part of the observations, this finding is less precise than the outcomes for an elder daughter’s deliveries (Tables 2 and 3). Ingrid Mogren Ulf Högberg Anna Winkvist Hans Stenlund