“Marriage trebles the risk of obesity,” the Daily Express has warned. It says that new research shows that once couples marry they are three times as likely to become obese compared with people who live separately.
The study underpinning this story analysed data on several thousand heterosexual people in the US from adolescence onwards to establish the links between relationship status and obesity. It found an association between becoming married and new cases of obesity. It also found that obesity-related behaviour, such as physical activity and television watching, were more similar between couples that had lived together for longer.
Some newspapers have implied that marriage itself causes weight gain, but these conclusions fail to highlight the study’s limitations and the complexities that the researchers discuss. For example, there may be a range of other factors associated with both marriage and the risk of becoming obese. The researchers discuss the effects of shared environmental factors and 'assortative mating', the phenomenon whereby people select their partner on the basis of similarities or dissimilarities to themselves. The benefits that have been associated with marital status through other research, including a trend towards living longer, should also be highlighted.
Where did the story come from?
Dr Natalie The and Penny Gordon-Larsen from the University of North Carolina in the US carried out this study. It was funded by the National Institute of Child Health and Human Development, part of the US National Institutes of Health, and was published in the peer-reviewed medical journal Obesity.
What kind of scientific study was this?
Previous studies have shown that that there are benefits associated with marital status, including reduced mortality. There is also a link between a person’s body mass index and their spouse’s, which is thought to be due to a shared household environment or ‘assortative mating’ (or an interaction between the two), where individuals select partners with similar behaviours and body types.
To date, research in this area has not clearly shown whether being married is associated with body weight and obesity. This retrospective cohort study assessed whether there was a link between ‘romantic relationships’ and obesity or obesity-related behaviours.
Researchers used data from the National Longitudinal Study of Adolescent Health, a prospective cohort study, which began in US schools in 1994. In the first wave researchers interviewed 20,745 children in grades 7-12, and their parents. From this group there were a further two waves of research, with 14,438 subjects being re-interviewed in 1994 and a third wave of interviews for 15,197 people in 2001 and 2002. During this third wave , participants were aged between 18 and 27 years.
The third wave also included a ‘couples sample’ in which respondents recruited their ‘romantic partners’ to take the same interviews. Interviews at each time point included an assessment of height and weight, physical activity, time spent watching TV or playing computer games (<14 hours screen time per week or more), romantic relationship status and other factors, including education, ethnicity and age.
There were two aspects to the study. In the first, researchers investigated whether entering into a relationship or being in a longer relationship, compared with a shorter one, was more likely to be associated with new cases of obesity.
To do this, they compared whether a change from a single to a cohabiting or married status between waves II and III was more closely linked with weight gain than with other status changes. They also looked at what effect the length of a relationship had on obesity. They excluded pregnant women, Native Americans, people obese at baseline, and those who were missing important data, which left them with 6,949 people to analyse in this portion of the study. People with relationship status other than single or dating during wave II were included in the study which assessed the effects of a change in relationship status between waves II and III.
In the second cohort part of their study, the researchers assessed how closely obesity-related behaviours were linked within married couples, cohabiting couples and long-term cohabiting couples when compared to single people or those in shorter relationships.
They selected a random sample of couples of three months or more, comprising a participant from wave III and their opposite sex partner, aged 18 or over. A total of 1,293 pairs of partners were available for analysis after excluding pregnant women and those who had missing data.
The researchers looked at whether the type of relationship (single, dating, cohabiting, or married) and duration of living together (not living together, 0.01 to 0.99 years, 1-1.99 years and 2 or more years) were linked to concordance (similarities) in levels of physical activity (neither partner moderately-to-vigorously active, one partner active or both partners active), obesity (neither obese, one obese, both obese) or screen time (one, neither or both watching 14 hours or less a week of television).
In both analyses, researchers adjusted their calculations to account for ethnicity, education and the age of parent or partner. Obesity was defined as a BMI of 30 or more.
What were the results of the study?
The study has several findings, including:
- Men who went from single/dating at wave II to married at III were 2.07 (95% CI 1.33 to 3.25) times more likely to become obese than those who went from single/dating to dating.
- Women making this same transition were 2.27 (95% CI 1.54 to 3.34) times more likely to become obese.
- Women who went from being single/dating at wave II to single at follow-up or single/dating to cohabiting were more likely to become obese. This association was not significant in men.
- The researchers report, but do not provide results, that women who went from single/dating at wave II to single at wave III were more likely to be persistently obese (i.e. obese at both time points).
In the couples sample, the study found that married and cohabiting partners were less healthy than dating romantic pairs in terms of physical activity, obesity and television/gaming time.
In the results about concordance, i.e. which characteristics couples shared, married couples were 3.3 times more likely than dating partners to share a similar obesity status to their partner than they were to have similar non-obesity status.
Married couples were also twice as likely to contain either one or two less physically active people than those who were dating. Both partners being sedentary was twice as common among cohabiting people than among those who were dating but not living together.
The researchers also found that men were more likely to have two or more weekly bouts of moderate-to-vigorous physical activity, but were also more likely to be the obese partner. Females were more likely than males to have less than 14 hours a week of screen time.
Compared to those living separately, women who cohabited two or more years with their partner were twice as likely to be obese, while this association was not significant for men.
What interpretations did the researchers draw from these results?
The researchers conclude that duration of living with a romantic partner is associated with obesity and obesity-related behaviours, and that the transition from single/dating to cohabitation or marriage was generally associated with an increased risk of obesity.
The authors conclude that a link between negative, obesity-related behaviours was strongest for married couples and couples who had lived together for two or more years. They say that this observation could increase the chance of partners passing on high-risk behaviours to their offspring, and that targeting the shared household environment may be the best way to establish health behaviours and reduce obesity in young adulthood.
What does the NHS Knowledge Service make of this study?
This retrospective cohort study has relied on data from a large number of individuals to establish the links between relationship status, obesity and behaviours related to weight gain. However, the benefits that other studies have associated with marital status should also be highlighted. Similar research suggests that romantic relationships reduce mortality rates and lower cigarette smoking.
There are also some limitations within this study, which may undermine the suggestion that marriage itself is responsible for weight gain:
- The researchers acknowledge that their study sample is a unique, young, heterosexual population, so their findings are not generalisable outside of these characteristics.
- The effects of a relationship lasting more than five years were not assessed in this study because too few of their sample met this criteria.
- There are factors that may change when a person enters a relationship, including diet, presence of children, weight control, etc. The role that these factors may play are not captured in this research.
- One of the greatest limitations of cohort studies in establishing causation is the failure or inability to control for the many possible confounding factors that could be affecting the link between an exposure and an outcome. The researchers adjusted for pregnancy history, which slightly weakened the associations they saw and also suggested that other unadjusted factors may play a part.
- There is no way to fully account for 'assortative mating’, the phenomenon whereby people choose partners similar or dissimilar to themselves. This in itself may explain why married couples (i.e. which may be viewed as a type of relationship with more serious intentions than just dating) are more likely to be similar to each other in behaviour and BMI than those who are just dating. However, the researchers say that the fact that there is a link between a longer duration of sharing a household and higher obesity and obesity-promoting behaviours suggests that common environment in a partnership may play a "more significant role" than assortative mating.
Notwithstanding the limitations outlined above, the study has found a link that deserves further investigation. There are likely to be many behaviours that change after marriage, and understanding these and the links between them could help in tackling obesity.
Analysis by Bazian
Edited by NHS Website
Links to the headlines
The Daily Telegraph, 11 June 2009
Daily Express, 7 July 2009
Links to the science
Obesity (2009) 17 7, 1441–1447