Association of Preterm Birth and Low Birth Weight With Romantic Partnership, Sexual Intercourse, and Parenthood in Adulthood

Key Points Question Are adults who were born preterm or with low birth weight less likely to experience social transitions normative of adulthood, such as romantic partnerships, sexual intercourse, or parenthood? Findings In this systematic review and meta-analysis of 21 studies describing up to 4.4 million participants, adults who were born preterm or with low birth weight were less likely to experience a romantic partnership, sexual intercourse, or parenthood than their peers who were born full-term. The likelihood of experiencing these social transitions decreased with lower gestational age and birth weight, and was similar in both young and middle adulthood. Meaning The findings suggest that adults who were born preterm or with low birth weight are less likely to have sexual or partner relationships than adults born full-term, which might put them at increased risk of decreased well-being and poorer physical and mental health.


Introduction
Preterm birth or low birth weight (PT/LBW) is associated with an increased risk for disability, 1,2 neurocognitive impairment, [3][4][5][6] learning difficulties, 3,6 and mental health problems, [7][8][9] with the association being stronger for those with lower gestational age. 3,[10][11][12] These functional deficits are associated with adverse impacts on preterm-born adults' socioeconomic outcomes. 13 However, little is known about whether those born preterm master social transitions into adulthood, such as building a supportive peer group, establishing romantic partnerships, having sexual intercourse, or becoming a parent.
Close, intimate, and supportive relationships are associated with increased happiness and well-being, 14,15 good physical health, 16 and good mental health. 17 Studies have shown that social relationships are more challenging for children born PT/LBW. 18 Indeed, prematurity has been associated with a behavioral phenotype [18][19][20] and personality profile [21][22][23][24] that includes being timid, socially withdrawn, overcontrolling, and disinclined toward risk-taking or fun seeking. These differences may predispose PT/LBW individuals to face greater difficulties in establishing romantic and peer relationships.
In contrast, research on social outcomes of adults born preterm is not conclusive. While Scandinavian registry studies have found that adults born PT/LBW were less likely to ever be in a registered partnership 11,12,25 or to be parents, 12,25 prospective studies have reported conflicting findings across [26][27][28] and within 2,29 studies. Regarding the latter, a Canadian cohort study 2,29 of extremely low-birth-weight infants reported different findings for social outcomes at distinct time points: while no differences were found in rates of marriage or cohabitation and parenthood between the extremely low-birth-weight individuals and those born full term at ages 22 to 26 years, 29 adults with extremely low birth weight were less likely to be married or cohabitating and to have had children during the fourth decade of life. 2 Additionally, there is a lack of research that has analyzed the impact of preterm birth on the quality of close relationships, such as with partners 2,30,31 and friends. 2,30,[32][33][34] Hence, there are inconsistent and scarce findings about the social lives of PT/LBW adults. This systematic review and meta-analysis systematically investigates the association between being born PT/LBW and social outcomes in adulthood, such as ever being in a romantic partnership, ever having had sexual intercourse, parenthood, quality of romantic relationship, and peer social support. Furthermore, we investigate whether there is a dose-response association according to degree of prematurity and whether outcomes are moderated by type of study (ie, cohort or registry), age, or sex.

Methods
This meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Metaanalyses (PRISMA) reporting guideline 35 and was registered with PROSPERO International prospective register of systematic reviews (PROSPERO identifier: CRD42017078286).

Study Selection Criteria
Studies were eligible for review according to the following criteria: (1) the sample included individuals who were born PT (<37 weeks' gestation) or LBW (<2500 g at birth); (2) term control group; (3) adult participants (ie, mean sample age Ն18 years); (4) measured at least 1 of the following social outcomes in adulthood: romantic partnership (eg, dating, cohabitation, marriage), quality of romantic relationship (eg, satisfaction, intimacy), sexual intercourse (ie, if ever experienced sexual intercourse), parenthood (ie, if any live biological child), or social support (ie, positive and supportive relationships with friends); and (5) the study was published in a peer-reviewed journal. If data from the same sample were published in multiple works for the same social outcome, we retained (1) the study with the longest follow-up interval (ie, oldest age at assessment); and (2) the study with the largest sample size and the broadest concept coverage.

Data Collection Process
Two of us (M.M. and A.B.) reviewed titles and abstracts of traced articles. The title and abstract screening was followed by the analyses of full texts to check inclusion criteria. Discordances were resolved by discussion among all authors. When reported information was unclear or numerical data were not obtainable, relevant corresponding authors were contacted for clarification.

Data Extraction
Studies reporting on PT or LBW were grouped into the same category because infants with low birth weight are mostly born preterm. 36 When information was available, we used 4 different gestational age subgroups: extremely preterm (EPT; <28 weeks or <1000 g), very preterm (VPT; 28-31 weeks or 1000-1500 g), moderate-to-late preterm (MLPT; 32-36 weeks or 1500-2500 g), and full-term (FT; >36 weeks or >2500 g). When studies referred to preterm birth without mentioning gestational weeks, data were included in the MLPT subgroup.
A standardized form was used to extract data from each study that included publication details, country, characteristics of participants (year of birth, sample size, gestational age or birth weight, percentages of men, and age), type of study (ie, cohort or registry), type of social outcome, and outcome data (ie, means and standard deviations or numbers and frequencies)   (Table 1). The extraction was conducted independently by 2 of us (M.M. and A.B.) and information was crosschecked for consistency. When inconsistencies emerged information was checked in the original study.

Quality Assessment
Study quality was assessed independently by 2 of us (M.M. and A.B.) using the Newcastle-Ottawa Scale 37 (eTable 1 in the Supplement). Scores could range from 0 to 9. The mean (range) of ratings for study quality was 7.3 (4-9), indicating overall good quality.

Statistical Analysis
Meta-analysis of the overall comparison between adults born PT/LBW and their FT peers was carried out with Comprehensive Meta-analysis version 2 software (Biostat) 38 for each social outcome. We used pooled odds ratios (ORs) with 95% confidence intervals for studies presenting dichotomous outcomes (eg, frequencies) and Hedges g for studies presenting continuous outcomes (eg, means and standard deviations) with random effects. Heterogeneity among studies was assessed with Cochran Q (P value), Higgins I 2 , and τ 2 . Low heterogeneity was defined as an I 2 value of 0% to 25%, moderate heterogeneity as an I 2 of 25% to 75%, and high heterogeneity as an I 2 of 75% to 100%. To explore heterogeneity, we conducted subgroup analyses (dependent on data availability) for degree of prematurity (ie, EPT, VPT, MLPT), type of study (ie, cohort or registry), age groups (ie, young adulthood [18-25 years] or middle adulthood [Ն26 years]), and sex.
Publication bias analysis was assessed through (1) the trim and fill procedure to examine the symmetry of effect sizes plotted by the inverse of the standard error 39 (ideally, effect sizes should Abbreviations: ELBW, extremely low birth weight (<1000 g); EPT, extremely preterm (<28 weeks' gestation), FT, full term; LBW, low birth weight (<2500 g); MLPT, moderate-to-late preterm (32-36 weeks' gestation); NA, not available; PT, preterm; VLBW, very low birth weight (1000-1500 g); VPT, very preterm (28-31 weeks' gestation). a This study reported on an early preterm (<34 weeks' gestation) subgroup overlapping with MLPT subgroup. We excluded this subsample of less than 34 weeks' gestation from the analysis. mirror one another on either side of the mean); (2) the Begg-Mazumdar rank correlation test to examine the likelihood of bias in favor of small sample size studies, 40 in which nonsignificance of correlation indicates no publication bias; and (3) Egger test to examine whether publication bias was related to the direction of study findings. 41 The intercept value provided by this test shows the level of funnel plot asymmetry from the standard precision.

JAMA Network Open | Pediatrics
Because PT and LBW were combined into 1 group, it is essential to prove that the findings of the meta-analysis are not dependent on this decision. Therefore, a sensitivity analysis was undertaken in which we repeated the analysis excluding the studies that reported on LBW only.

Study Characteristics
Of 1829 articles screened, 21 studies were eligible for quantitative analysis (Figure 1). According to our selection criteria, it was possible to identify 14 studies for romantic partnership, 9 for sexual intercourse, 11 for parenthood, 3 for quality of romantic relationship, and 5 for peer social support.
We also identified 5 studies for number of friends, 32

Differences in Social Outcomes Between Adults Born PT/LBW and FT
Meta-analysis results ( Table 3 and Figure 2) revealed that PT adults were less likely to have ever been involved in a romantic partnership than those born FT (OR, 0.72; 95% CI, 0.64-0.81). Heterogeneity analysis indicated high variation in effects between studies. Subgroup analysis according to the  One study included in the qualitative synthesis was excluded from meta-analysis because it reported only on number of friends. Significant differences between PT and FT adults were found for the quality of romantic relationship (Table 3 and Figure 2). Adults born PT/LBW perceived the relationship with their partner as significantly more satisfying or intimate than those born FT. Heterogeneity was not significant for this variable. Furthermore, we observed no significant differences between PT/LBW and FT adults regarding the peer social support.   34  Odds ratio (OR) or standardized mean difference (SMD) for individual studies are indicated by squares and 95% CIs by horizontal lines. Pooled estimates and their 95% CIs are represented by diamonds. The size of the squares and the diamonds are proportional to the weight assigned to the relative effect sizes. The arrow for the study of Hallin and Stjernqvist 32 indicates that the 95% CI exceeds the limit for the effect size range.

Publication Bias
Under the random-effects model, the point estimate for the combined studies was 0.70 (95% CI, 0.67 to 0.73) for romantic partnership, 0.04 (95% CI, 0.02 to 0.07) for quality of romantic relationship, 0.54 (95% CI, 0.39 to 0.74) for ever having experienced sexual intercourse, 0.78 (95% CI, 0.66 to 0.93) for parenthood, and 0.15 (95% CI, −0.32 to 0.01) for peer social support. With the use of trim and fill, these values remained unchanged for all relational outcomes, indicating no publication bias. The Begg-Mazumdar rank correlation and Egger test were not statistically significant for all outcomes, indicating no evidence of publication bias.

Sensitivity Analysis
Results remained the same after excluding studies that reported only birth weight. Hence, PT adults were less likely to be in a partnership (OR, 0.75; 95% CI, 0.66-0.85), to have ever had sexual intercourse (OR, 0.48; 95% CI, 0.31-0.76), and to be parents (OR, 0.80; 95% CI, 0.67-0.97) in comparison with FT adults.

Discussion
Our findings revealed that adults born PT/LBW are less likely to experience romantic partnerships, sexual intercourse, or parenthood. Nevertheless, when they were in a romantic partnership or had friends, the quality of these relationships was similar to those experienced by FT adults.
Using summary data from prospective studies with more than 4 million participants provided evidence for a temporal association between being born PT/LBW and establishing social transitions into adulthood, here defined as romantic partnership, sexual intercourse, and parenthood. The associations were robust across degree of prematurity, age groups, and sex. These findings are consistent with the increasing recognition of the impact that early life influences have on outcomes in adulthood. 13,44,45 Furthermore, our findings are in line with evidence of a preterm behavior phenotype that follows into adulthood, 21,22,24 which might be associated with more difficulty engaging in these transitions for individuals born PT/LBW. We verified that the strength of the associations between PT/LBW and social transitions were in general small for romantic partnership and parenthood and moderate for sexual intercourse. The associations diverged depending on degree of prematurity, type of study, and age group. The subgroup analysis for degree of prematurity revealed that the likelihood of PT/LBW experiencing a romantic partnership, sexual intercourse, or parenthood decreased with lower gestational age.
Indeed, a significant dose-response association was found between degree of prematurity and rates of romantic partnership and parenthood, with adults born EPT 67% less likely to be in a romantic partnership and 69% less likely to be parents than those born FT.
With respect to the type of study, we found that PT/LBW adults were less likely to have experienced romantic partnership or parenthood in cohort studies compared with registry studies.
This difference may be related to the fact that cohort studies included mainly individuals born at less than 32 weeks' gestational age, whereas registry studies included the full range of preterm birth, and the likelihood of occurrence of these transitions decreases with lower gestational ages.
Adults born PT/LBW were overall less likely than those born FT to be parents. However, this difference was not significant in the younger age group. This finding is in line with the findings of Saigal et al. 2 A likely explanation is that, consistent with the general trend for first parenthood to take place in the late 20s or early 30s, 47 few participants in the FT group were parents, yielding no difference between groups. However, once parenthood was assessed in middle adulthood, the differences between PT/LBW and FT groups emerged. At a societal or population level, it suggests that prematurity is associated with a cross-generational fertility loss. Adults born PT/LBW are less likely to become parents and their parents were already less likely to have subsequent children after their preterm child was born. 48 Overall, rather than a delay, our findings suggest persistent difficulties in making these social transitions that have been associated with negative outcomes later in life, 49 With respect to sex, it was only possible to include 4 to 5 studies in these subgroup analyses. We verified that both men and women born PT/LBW were less likely to have experienced romantic partnerships or sexual intercourse than their counterparts born FT. No differences were found for parenthood; however, it is important to note that there were few participants with children in this subgroup analysis. Previous studies have not been consistent when analyzing the role of sex on social outcomes. 2,26,28 Although it was possible to pool data from more than 1200 participants in these analyses, the lack of studies reporting on sex highlight the need for future research to clarify its moderating role.
We found that PT/LBW individuals perceived their romantic relationships slightly more positively than FT individuals, and that there was no difference for perceptions of peer social support between both groups. Although it was not possible to assess the amount of friends in this metaanalysis, most studies have found that PT/LBW adults had fewer friends 42,43,53 than FT adults. In addition, studies on PT/LBW children and adolescents reported poorer-quality relationships with peers 18,54 than those born FT, including being bullied by peers more often. 55 Hence, our findings suggest that despite fewer close relationships, relationship quality was not poorer when PT/LBW adults had friends or a partner, or the quality of relationships in PT/LBW individuals improves into adulthood. Longitudinal studies are required to explore these alternative explanations. Furthermore, the degree of prematurity is associated with physical and mental health and cognitive development, [11][12][13]45 and information on disability was not available for most studies. Thus, it could not be assessed whether functional deficits or disability moderated the association between PT/LBW and social outcomes. In this study, PT and LBW were treated as 1 factor. Although these constructs show high comorbidity and our sensitivity analyses revealed consistent results, it would be important to disentangle the effects of PT and LBW and their possible additive effects on social outcomes. This would involve considering data on birth weights appropriate for gestational age or small for gestational age, which most studies included in the meta-analyses did not report. Future research should address these limitations by conducting individual participant meta-analysis and obtaining data directly from the study authors.

JAMA Network Open | Pediatrics
The heterogeneity of studies was high, indicating considerable variation. This might arise from incorporating cohort and registry studies with various sample sizes. To address this possibility, we used a random-effects model in the analysis and conducted moderator analyses. Nevertheless, our moderator analysis explained only some of the heterogeneity. Thus, the findings from the current study should be interpreted with caution and analysis should be repeated when more adulthood data becomes available from the cohort studies. Also, only English publications were considered in this meta-analysis and, therefore, potential language bias should be taken into account.

Conclusions
This systematic review and meta-analysis provides a qualitative and quantitative overview of the current state of knowledge concerning social outcomes in adults born PT/LBW. Pooling data from multiple cohort and registry studies provided evidence that fewer adults born PT/LBW experience romantic partnerships, sexual intercourse, or parenthood. These associations are stronger the lower the gestational age and were found in young and middle adulthood. However, when PT/LBW individuals were in a romantic partnership or had friends, the quality of these relationships was at