Assortative marriages by body mass index have increased simultaneously with the obesity epidemic

Abstract

Background: The genetic predisposition to obesity may have contributed to the obesity epidemic through assortative mating. We investigated whether spouses were positively assorted by body mass index (BMI; = kg/m(2)) in late childhood, and whether changes in assorted marriage by upper BMI-percentiles occurred during the obesity epidemic.

Methods: In the Copenhagen School Health Records Register (CSHRR) boys and girls with measures of BMI at age 13 years later became 37,792 spousal-pairs who married between 1945 and 2010. Trends in the spousal BMI correlations using sex-, age-, and birth cohort-specific BMI z-scores across time were investigated. Odds ratios (ORs) of marriage among spouses both with BMI z-scores >90th or >95th percentile compared with marriage among spouses ≤90th percentile were analyzed for marriages entered during the years prior to (1945-1970), and during the obesity epidemic (1971-2010).

Findings: Spousal BMI correlations were around 0.05 and stayed similar across time. ORs of marriage among spouses with BMIs >90th percentile at age 13 were 1.21, 1.05-1.39, in 1945-1970, and increased to 1.63, 1.40-1.91, in 1971-2010 (p = 0.006). ORs of marriage among spouses both >95th BMI percentile were higher and increased more; from 1.39, 1.10-1.81, to 2.39, 1.85-3.09 (p = 0.004).

Interpretation: Spousal correlations by pre-marital BMIs were small and stable during the past 65 years. Yet, there were assorted marriages between spouses with high BMI at age 13 years and the tendency increased alongside with the obesity epidemic which may increase the offsprings' predisposition to obesity.

Keywords: assortative mating; body mass index; childhood; human genetics; obesity; overweight; phenotype.

Figure 1

Flowchart of the study population.

Figure 2

Spousal correlations for standardized BMI z-scores from age 13 years with 95% CI estimates, by 5-year intervals of marriage cohort.

Figure 3

Observed compared to expected ratios of marriage (log scale) according to group-based BMI z-scores at age 13 years, shown for two marriage periods, prior to (A) and during (B) the obesity epidemic. Ratio = 1 corresponds to random marriage. Pearson's Chi squared test, p-values; p = 0.005 (1945–1970) and p < 0.0001 (1971–1989).

Figure 4

Odds Ratios of boys with BMIs above the 90th (●) or 95th (■) percentile marrying girls similarly above the 90th or 95th BMI percentile at age 13 years compared to marriages among pairs under or equal to the 90th percentile. Differences by marriage periods p = 0.006, and p = 0.005, in pairs above the 90th and 95th BMI percentiles, respectively. ORs are adjusted for height at age 13 years as well as marriage age.