Birth Weight and Blood Pressure in Spouses

Abstract

ArticlePlus Click on the links below to access all the ArticlePlus for this article. Please note that ArticlePlus files may launch a viewer application outside of your web browser. https://links.lww.com/EDE/A215 To the Editor: Low birth weight is correlated with high adult blood pressure according to the programming hypothesis,1–3 but the estimated effect is small4 and may even be confounded by social conditions or lifestyle factors.3,5 We studied the association between birth weight and blood pressure in couples to adjust for the social conditions that couples share. We randomly selected 1000 cohabiting couples, 50 to 65 years of age, from the Danish cohort study Diet, Cancer and Health. Using the place and date of birth, and names as identifiers, we located each person's birth record in the state archive; these midwife records include data on birth weight. If we could not match by one surname (maiden name) and place of birth for both partners, the couple was excluded from further analysis. We extracted data on midlife systolic and diastolic blood pressure (SBP and DBP) from the health examination at recruitment to the cohort study. Data were analyzed in a mixed model of 2 linear regression equations allowing for dependence between measurements from spouses.6 Analyzing husband and wife jointly allows for adjustment for latent lifestyle factors. We located 1670 of the 2000 individuals with birth data (84%). The midwife records of the 330 remaining persons were either missing or too ambiguous for identification. Among the 1670 persons with midwife data, both partners' data were available in 472 couples. Table 1 shows the slope estimates of SBP and DBP on birth weight from several models. The confounder adjustment for both latent and fixed lifestyle covariates did not substantially change the effect of birth weight on SBP and DBP. Inclusion of current body mass index in the analyses also did not affect estimates.7,8TABLE 1: Effect of Birth Weight (per kilogram)Furthermore, we found an unexpected, relatively large positive correlation between the blood pressure of one spouse and the birth weight of the other. According to the data, a man with low birth weight will experience high SBP as seen in Figure 1. At the same time, his high SBP correlates with both high birth weight and high SBP of the wife. That is difficult to explain in causal terms. This unexpected pattern was present but less apparent for DBP.FIGURE 1.: Correlations of birth weight and systolic blood pressure in the couple. The P values refer to statistical significance of the hypothesis of no correlation.These estimated associations could be the result of chance. Otherwise, we speculate whether selection of spouses or effect modification by lifestyle or treatment of high blood pressure is at work. We recommend that other spouse data are used to see whether our findings can be replicated. The proportion of couples with missing birth data either for one or both partners was large but most likely unrelated to the hypothesis under study. Data were missing when midwife records were simply lost in the archive and for persons born on a day when a midwife had more than one delivery of mothers with the same surname. None of these conditions should cause selection bias, except by chance. Birth weight and outcome data were collected independently, which makes differential misclassification unlikely. Our objective was to put the programming hypothesis on birth weight and adult blood pressure to a critical test. We got results that corroborate the association, but the finding of a correlation between the blood pressure of one spouse and the birth weight of the other may also serve as a warning sign. One should be careful when interpreting results from 2 measurements made 50 years apart without much control of what took place in the time between. For more details regarding this study, refer to the supplementary material available with the online version of this letter. ACKNOWLEDGMENTS We thank Hanne Grand for patiently reading the midwife protocols. Thanks are also due to Anne Tjønneland and the staff at the Danish study Diet, Health and Cancer. Anne Vingaard Olesen Institute of Public Health Department of Epidemiology University of Aarhus Aarhus, Denmark [email protected] Erik Thorlund Parner Institute of Public Health Department of Biostatistics University of Aarhus Aarhus, Denmark Kim Overvad Department of Clinical Epidemiology Aarhus University Hospital Aarhus, Denmark Jørn Olsen The Danish Epidemiology Science Centre at the Institute of Public Health University of Aarhus Aarhus, Denmark UCLA School of Public Health Department of Epidemiology Los Angeles, California