High blood pressure has been estimated to affect 45% of the US adult population and has been identified as one of the most important preventable factors linked to illness and mortality (James et al., Reference James, Oparil, Carter, Cushman, Dennison-Himmelfarb, Handler and Ortiz2014; Muntner et al., Reference Muntner, Carey, Gidding, Jones, Taler, Wright and Whelton2018). This multi-method study examined the overarching hypothesis that a pattern of enmeshment and intensity in key adolescent social relationships will forecast the presence of future problematic romantic and nonromantic relationships, ultimately culminating in higher blood pressure later in adulthood. This hypothesis addresses potential adolescent relational roots of the robust links identified in adulthood between problematic social relationships, poor physical health and, ultimately, heightened risk of early mortality (Holt-Lunstad, Smith, & Layton, Reference Holt-Lunstad, Smith and Layton2010; Uchino, Reference Uchino2009). Although these links are increasingly recognized within adulthood, the pathways by which they play out are still far from clear (Uchino, Reference Uchino2009). Preventive intervention efforts have also been constrained by the degree to which problematic social relationship patterns have already become entrenched by adulthood, rendering efforts to alter these patterns difficult at best (Davila et al., Reference Davila, Mattanah, Bhatia, Latack, Feinstein, Eaton and Zhou2017). Ideally, such efforts should begin well before adult relationship patterns have become ingrained. This, however, requires a developmentally-informed understanding of the precursors of relationship difficulties in adulthood that are linked to problematic health outcomes.
How do adolescent romantic relationships fit into this picture? Developmental task theory suggests that romantic involvement is most likely to be problematic when the intensity of this involvement is out of sync with the capacities of the adolescent at a given developmental stage (McCormick, Kuo, & Masten, Reference McCormick, Kuo, Masten, Fingerman, Berg, Smith and Antonucci2011; Roisman, Masten, Coatsworth, & Tellegen, Reference Roisman, Masten, Coatsworth and Tellegen2004). Early in adolescence, precocious involvement in romantic relationships is linked to long-term social difficulties ranging from depressive symptoms to delinquent behavior (Allen, Schad, Oudekerk, & Chango, Reference Allen, Schad, Oudekerk and Chango2014; Loeb, Kansky, Narr, Fowler, & Allen, Reference Loeb, Kansky, Narr, Fowler and Allen2020). In mid-adolescence, the presence of a significant romantic relationship (i.e., lasting 3 months or longer) has been linked to relatively higher levels of depressive symptoms (Davila, Reference Davila2008; Szwedo, Chango, & Allen, Reference Szwedo, Chango and Allen2015). Later on, across adolescence and early adulthood, high cumulative levels of involvement in romantic relationships have been linked to greater levels of deviant behavior (Cui, Ueno, Fincham, Donnellan, & Wickrama, Reference Cui, Ueno, Fincham, Donnellan and Wickrama2012).
Intense romantic relationships may be particularly problematic for adolescents as their relative immaturity and inexperience may make it harder for them to cope with the psychological demands of such relationships (Davila, Steinberg, Kachadourian, Cobb, & Fincham, Reference Davila, Steinberg, Kachadourian, Cobb and Fincham2004; Szwedo et al., Reference Szwedo, Chango and Allen2015). Moreover, relationships that are intense by virtue of being highly time-consuming – in terms of absorbing time each day and over a period of weeks and months – are likely to crowd out other relationships (Joyner & Udry, Reference Joyner and Udry2000; Roth & Parker, Reference Roth and Parker2001) and thus increase the adolescent's sense of the stakes involved in maintaining the relationship. The result is a formative experience of relationships as emotionally overwhelming, potentially autonomy-threatening, and even, ironically, isolating – an experience likely to influence expectations and behaviors in future relationships going forward.
By taking time away from the development of relationships with nonromantic peers, intense romantic relationships also deprive adolescents of key social learning opportunities. Adolescents who become so wrapped up in a romantic relationship that they stop spending time with other friends are likely to see those other relationships either wither or become more conflictual as their friends come to feel written off. In turn, being cut off from opportunities for the normal development of close peer relationships – at a time when such relationships have been identified as critical to longer-term development – is likely to make these peer relationships more challenging going forward (Allen, Narr, Kansky, & Szwedo, Reference Allen, Narr, Kansky and Szwedo2020; Szwedo et al., Reference Szwedo, Chango and Allen2015). Of note, a lack of positive nonromantic relationships in adulthood has also been reliably linked to higher blood pressure (Grant, Hamer, & Steptoe, Reference Grant, Hamer and Steptoe2009; Shankar, McMunn, Banks, & Steptoe, Reference Shankar, McMunn, Banks and Steptoe2011).
Problematic romantic relationships in adolescence have been found to predict similar problems in romantic relationships in young adulthood, suggesting long-term continuities (Madsen & Collins, Reference Madsen and Collins2011). At the extreme, long-lasting adolescent relationships with a high frequency of contact between partners have even been found to predict an increased incidence of romantic relationship violence (Giordano, Soto, Manning, & Longmore, Reference Giordano, Soto, Manning and Longmore2010). These links are important because, ultimately, negativity in adult close relationships has been directly linked both to higher blood pressure and to related problematic physiological indicators such as cardiovascular reactivity and artery calcification (Holt-Lunstad, Uchino, Smith, Olson-Cerny, & Nealey-Moore, Reference Holt-Lunstad, Uchino, Smith, Olson-Cerny and Nealey-Moore2003; Robles & Kiecolt-Glaser, Reference Robles and Kiecolt-Glaser2003; Uchino et al., Reference Uchino, Bosch, Smith, Carlisle, Birmingham, Bowen and O'Hartaigh2013; Uchino, Smith, & Berg, Reference Uchino, Smith and Berg2014).
Consideration of the potential effects of intense romantic relationships naturally leads to the corollary question of why adolescents would be initially drawn to them. The most plausible explanation suggests that an inclination toward intensity would be most likely for adolescents who had grown accustomed to intense and autonomy-impinging relationships in their family of origin (Oudekerk, Allen, Hessel, & Molloy, Reference Oudekerk, Allen, Hessel and Molloy2015). Psychologically controlling parenting behavior, in which adolescent autonomy is undermined and efforts to achieve independence from parents are discouraged, fits this profile. Such behavior has also been repeatedly identified as having long-term ramifications for social interaction patterns (Loeb, Kansky, Tan, Costello, & Allen, Reference Loeb, Kansky, Tan, Costello and Allen2020) and has even been implicated in problematic long-term physiological patterns of stress response (Loeb, Davis, et al., Reference Loeb, Davis, Narr, Uchino, Kent de Grey and Allen2020). Whether the experience of emotionally intense and conflicted relationships within the family actually carries over to predict involvement in intense romantic relationships has not, however, been previously examined.
A mediational “chains of risk” perspective suggests that early stressors, such as those that arise from maladaptive levels of parental or romantic involvement, may predict future health outcomes by cascading forward to influence future relationship struggles, which then mediate links to health outcomes (Ben-Shlomo & Kuh, Reference Ben-Shlomo and Kuh2002; Oudekerk et al., Reference Oudekerk, Allen, Hessel and Molloy2015). Although such mediated chains of risk – for example from adolescent intensity to later relationship problems to higher blood pressure – appear as the most plausible means by which adolescent romantic intensity may be linked to adult blood pressure, a second possibility is that more direct, unmediated links also exist. In adolescence, the stress response system is particularly plastic (Romeo, Reference Romeo2010) and, given growing brain functional connectivity, adolescence has been posited as a “switch point” for the long-term calibration of stress responsivity both in metabolic and anatomic terms (Ben-Shlomo & Kuh, Reference Ben-Shlomo and Kuh2002; Gee et al., Reference Gee, Humphreys, Flannery, Goff, Telzer, Shapiro and Tottenham2013; Goff et al., Reference Goff, Gee, Telzer, Humphreys, Gabard-Durnam, Flannery and Tottenham2013; McEwen, Reference McEwen2007). Adolescence has been suggested to be a second sensitive period for the development of the social brain (Blakemore & Mills, Reference Blakemore and Mills2014). In adulthood, an altered stress response is considered a key part of the hypertensive process (Balint et al., Reference Balint, Gander, Pokorny, Funk, Waller and Buchheim2016).
This suggests the possibility that the stress of overly intense romantic experiences in adolescence – and perhaps of the breakups that typically eventually follow (Sbarra & Ferrer, Reference Sbarra and Ferrer2006) – may have a formative effect on long-term stress response patterns and, ultimately, on adult blood pressure. For example, a devastating breakup in adolescence may leave the individual chronically tense and feeling vulnerable in future relationships. This notion is consistent with a weathering perspective, which posits that stress at a vulnerable stage of development can have long-term physical health implications regardless of whether or not the stress continues (Geronimus, Reference Geronimus2013). In line with this perspective, several recent studies have now linked problematic parental, friend, and romantic relationships in adolescence to longer-term health difficulties in ways that appeared relatively unmediated by future relationship qualities (Allen, Loeb, Tan, & Narr, Reference Allen, Loeb, Tan and Narr2017; Loeb, Davis, et al., Reference Loeb, Davis, Narr, Uchino, Kent de Grey and Allen2020). To the best of our knowledge, however, neither chains of risk nor weathering pathways from adolescent relationship qualities to adult blood pressure have been previously examined.
This 14-year, multi-method, prospective study examined intense involvement in romantic relationships in a community sample of mid to late adolescents as a predictor of a cascading pattern of relationship difficulties culminating in higher adult resting blood pressure. A developmental model was assessed in which overinvolvement was viewed as predicting adult blood pressure both directly and as mediated via future difficulties in establishing successful social relationships.
This study tested four primary hypotheses.
-
Hypothesis 1. Intense adolescent romantic involvement will predict higher adult resting blood pressure.
-
Hypothesis 2. Intense involvement will be predicted by psychologically controlling parenting and will, in turn, predict a cascade of future romantic and peer relationship qualities reflecting negativity and conflict in later relationships.
-
Hypothesis 3. The primary long-term effects of early intense involvement will be mediated via future relationship qualities; the possibility of direct (i.e., weathering) predictions, independent of future relationship qualities, is also examined.
-
Hypothesis 4. Predictions will exist over and above baseline family income and trait markers of adult personality characteristics.
Method
Participants
This article was drawn from a larger longitudinal investigation of adolescent peer and romantic partner influences on adult development. The final sample of 146 participants was a subsample of an original sample of 184 participants assessed as part of a broader study at age 13 years, who had both data on romantic relationships at ages 17–19 years and on blood pressure at ages 29 and/or 31 years (an attrition rate of approximately 1% per year across 18 years). The final sample included 62 males and 84 females and was racially/ethnically and socioeconomically diverse and representative of the community from which it was drawn: 85 adolescents (58%) identified themselves as Caucasian, 44 (30%) as African American, 1 (1%) as Hispanic, 2 (2%) as Asian, 11 (8%) as of mixed race/ethnicity, and 3 (2%) as from other or unreported groups. The adolescents’ parents reported a median family income in the range US$40,000–59,999 at the initial assessment. Adolescents were recruited from seventh and eighth grades of a public middle school drawing from suburban and urban populations in the southeastern USA. Information about the study was provided via an initial mailing to parents with follow-up presentations to students at school lunches. Formal recruitment took place via telephone contact with parents. Students who had already served as close peer informants in the study were not eligible to serve as primary participants. Of the students eligible for participation, 63% of adolescents and parents agreed to participation when the parents were contacted. Adolescents provided informed assent before each interview session, and parents and adult participants provided informed consent. Interviews took place in private offices within a university academic building.
The assessments in this study were obtained from adolescents, their mothers, and their close friends (as described later) at mean ages (in years) of 13.3 (SD = 0.64), 14.3 (SD = 0.76), 15.2, (SD = 0.81), 16.4 (SD = 0.87), 17.3 (SD = 0.88), 18.2 (SD = 1.28), 19.7 (SD = 1.07), 25.7 (SD = 0.99), 26.6 (SD = 1.00), 27.6 (SD = 1.00), 28.8 (SD = 1.21), 30.2 (SD = 0.96), and 30.8 (SD = 1.01). At 26 years of age, assessment of romantic partners of participants who were in a relationship of 3-months duration or greater also provided data on their satisfaction with that relationship.
Attrition analyses
Attrition analyses compared the subsample of participants who had data available for adolescent-era romantic intensity and adult blood pressure (N = 146) to those who only had data for adolescent-era romantic intensity (N = 36). Females were more likely than males to have both types of data available (86% continuation rate for females vs. 74% for males; p = .04). Attrition analyses also compared the subsample of participants in the final study who did have romantic partner data versus those who did not. No differences were found.
These results suggest that attrition was not likely to have distorted any of the findings reported. Nonetheless, to best address any potential biases due to attrition in longitudinal analyses or missing data within waves, full-information maximum likelihood methods were used, with analyses including all variables that were linked to future missing data (i.e., where data were not missing completely at random). Because these procedures have been found to yield the least biased estimates when all available data are used for longitudinal analyses (vs. listwise deletion of missing data) (Arbuckle, Reference Arbuckle, Marcoulides and Schumaker1996), the entire original sample was utilized for these analyses. This full sample thus provided the best possible estimates of variances and covariances in measures of interest and was least likely to be biased by missing data.
Procedure
In the initial introduction and throughout all sessions, confidentiality was assured to all study participants and adolescents/adults were told that no one would be informed of any of the answers they provided. Participants’ data were protected by a confidentiality certificate issued by the US Department of Health and Human Services, which protected information from subpoena by federal, state, and local courts. Transportation and childcare were provided if necessary. Adolescent/adult participants and participants’ romantic partners were all paid for participation.
Measures
Primary measures
Resting blood pressure (ages 29 and 31 years)
A Dinamap Pro100 monitor (Critikon Corporation, Tampa, FL) was used to measure systolic blood pressure (SBP) and diastolic blood pressure (DBP). The Dinamap uses the occillometric method to calculate blood pressure. Cardiovascular assessments were obtained via a properly-sized occluding cuff positioned on the upper nondominant arm of the participant. Mean SBP and DBP were calculated by averaging three measurements taken during a baseline assessment during which the participant was asked to relax, sit quietly, and watch a soothing 10-min video. Measurements were taken 7, 8.5, and 10 min into the video. Participants were assessed at age 29 and, for a subset, again at age 31 with assessments averaged across the two time points.
Romantic relationship intensity (ages 17–19 and 28–30 years)
The intensity of an adolescent's romantic experience was assessed in terms of the amount of time spent alone with the current romantic partner and the duration over which this relationship extended. Participants were also asked how much time they currently spend alone with their current romantic partner and given the choice to answer either as a daily average or weekly average (answers given in weekly terms were subsequently translated into average daily hours). Across ages 17–19 years, means ranged from 2.2–2.6 h spent alone with the partner per day; across ages 28–30, means ranged from 3.7–4.8 h spent alone with the partner per day. These hourly averages were then multiplied by the duration of the current relationship in months, with means in a given year ranging from 12.3 to 15.8 months for the adolescents and 50.1 to 62.7 months for the adults. This product was used as the romantic relationship intensity score (effectively the month-hours of time spent alone with a romantic partner). Assessments were repeated each year from ages 17–19 years and 28–30 years and the values were averaged to yield the final score used in analyses. Participants who were not in a relationship at a given assessment point received an intensity score of zero for that point.
Maternal psychological control (age 16 years)
Participants’ mothers completed the 10-item psychological control versus psychological autonomy subscale of the Childhood Report of Parenting Behavior (Schaefer, Reference Schaefer1965), a widely used measure of psychological control (Barber, Reference Barber1996). Example items include: “I say if my child really cared for me, she would not do things that cause me to worry;” “I am less friendly with my son/daughter if s/he does not see things my way;” “I am always telling my son/daughter how s/he should behave.” Internal consistency was excellent (Cronbach's α = .79).
Peer conflict (ages 25–29 years)
Participants reported their level of conflict with their closest nonromantic friend at the time of the assessment using three scales from the Network of Relationships Inventory (Furman & Burhmester, Reference Furman and Buhrmester1985); each scale comprised three items. These included scales for conflict, antagonism, and punishing behaviors in the relationship. Scores for these items were summed and averaged to yield a total conflict score for each year. The assessment was repeated annually when participants were 25–29 years of age and the scores across years were averaged to yield the final peer conflict measure. The measure displayed strong reliability within year (Cronbach's α = .90–.94 across year) and a fair degree of consistency from one year to the next (r = .51–62, p < .001).
Romantic partner satisfaction (age 26 years)
Romantic partners of participants were surveyed at participant age 26 years regarding their satisfaction with the relationship with the participant using the three-item satisfaction scale of the Network of Relationships Inventory (Furman & Burhmester, 1985). Internal consistency for the scale was good (Cronbach's α = .90).
History of serious illness prior to 18 years of age
During the adult phase of the study, participants reported on the presence versus absence of 43 distinct significant health problems that were first experienced prior to age 18 years and which had led, at some point, to hospitalization. This measure provided a baseline marker of potential pre-existing health difficulties and has been previously found to predict current adult health quality (Allen, Uchino, & Hafen, Reference Allen, Uchino and Hafen2015).
Big five personality traits (age 24 years)
At 24 years of age, the big five personality traits were assessed with the 50-item International Personality Item Pool (Goldberg et al., Reference Goldberg, Johnson, Eber, Hogan, Ashton, Cloninger and Gough2006), using a 5-point Likert-scale, summing across 10 items each tapping constructs of extraversion, agreeableness, conscientiousness, emotional stability, and imagination. This measure has previously demonstrated strong internal consistency, retest reliability, convergence with longer big five personality measures, and self-peer agreement (Goldberg et al., Reference Goldberg, Johnson, Eber, Hogan, Ashton, Cloninger and Gough2006). For this sample, Cronbach's α ranged from.74 to .89.
Covariates and measures used in post-hoc analyses
Body mass index (ages 29 and 31 years)
Body mass index (BMI) was assessed at the assessments at 29 and 31 years of age. Height (in meters) and weight (in kilograms) were assessed with the participant wearing light clothing and BMI was calculated using the standard formula BMI = weight/height2. The BMI was then averaged across the two assessment points and log-transformed.
Household income
Household income was assessed in terms of the total family income as reported by parents in the adolescent's household at participant ages of 13, 16, and 18 years, on a scale ranging from 1 (less than US$5000 annually) to 8 (more than US$60,000 annually) (in 1998–2003 US$). Scores were averaged across these three assessments.
Close friendship quality (ages 17–19 years)
Each year, from age 17 to 19 years, close friends rated participants on their competence at establishing and maintaining a strong close friendship, using the four-item friendship competence scale from the Adolescent Self-Perception Profile (Harter, Reference Harter1988). Although the scale was originally labeled “close friendship competence,” examination of the items suggests that it is better conceptualized as a measure of the quality and intimacy of the friendship. Items focused, for example, on the extent to which the teen had “a close friend they share secrets with,” “a friend close enough to share really personal thoughts with,” and a “really close friend to share things with.” Results were averaged across the 3 years to produce the final scale. Internal consistency was good (Cronbach's α = .65–.74 within years).
State of mind regarding attachment (age 24 years)
The Adult Attachment Interview (AAI) (George, Kaplan, & Main, Reference George, Kaplan and Main1996; Kobak, Cole, Ferenz-Gillies, Fleming, & Gamble, Reference Kobak, Cole, Ferenz-Gillies, Fleming and Gamble1993) was used to probe individuals’ descriptions of their childhood relationships with parents in both abstract terms and with requests for specific supporting memories. For example, subjects were asked to list five words describing their early childhood relationships with each parent and then to describe specific episodes that reflected those words. Other questions focused on specific instances of upset, separation, loss, trauma, and rejection. Finally, the interviewer asked participants to provide more integrative descriptions of changes in their relationships with parents and the current state of those relationships. The interview consisted of 18 questions and lasted an hour on average. The interviews were audiotaped and transcribed for coding.
The AAI Q-set (Kobak et al., Reference Kobak, Cole, Ferenz-Gillies, Fleming and Gamble1993) was designed to closely parallel the AAI classification system (Main, Goldwyn, & Hesse, Reference Main, Goldwyn and Hesse2002) but to yield continuous measures of qualities of attachment organization. Each rater read a transcript and provided a Q-sort description by assigning 100 items into nine categories ranging from most to least characteristic of the interview, using a forced distribution. All interviews were blindly rated by at least two raters with extensive training in both the Q-sort and with formal workshop training and certification for coding using the AAI classification system. Q-sorts were then compared with a dimensional prototype sort for secure versus anxious interview strategies, reflecting the overall degree of coherence of discourse, the integration of episodic and semantic attachment memories, and a clear objective valuing of attachment. The individual correlation of the 100 items of an individual's Q-sort with a prototype sort for a maximally secure transcript was then used as that participant's scale security score (ranging from −1.00 to 1.00). Inter-rater reliability, assessed via the intraclass correlation coefficient, for the final security scale score was .71, which is considered in the good range for this statistic (Cicchetti & Sparrow, Reference Cicchetti and Sparrow1981).
Popularity (ages 13–16 years)
Adolescents’ capacity to establish themselves as preferred social companions with a range of their peers was assessed using a limited nomination procedure. Each adolescent, their closest friend, and two other target peers were asked to nominate up to 10 peers in their grade with whom they would “most like to spend time on a Saturday night.” The raw number of “like” nominations each teen received was standardized within grade level as a measure of desirability as a social companion in the broader peer group following the procedure described by Coie, Dodge, and Coppotelli (Reference Coie, Dodge and Coppotelli1982). This approach to assessing social acceptance has been previously found to be relatively stable over time and related to adolescent attachment security, qualities of positive parental and peer interactions, and short-term changes in levels of deviant behavior (Allen, Porter, McFarland, Marsh, & McElhaney, Reference Allen, Porter, McFarland, Marsh and McElhaney2005; Allen, Porter, McFarland, McElhaney, & Marsh, Reference Allen, Porter, McFarland, McElhaney and Marsh2007; McElhaney, Antonishak, & Allen, Reference McElhaney, Antonishak and Allen2008). Ratings were averaged across the 4 years to produce the final rating.
Social skills (ages 13–17 years)
The test-based Adolescent Problem Inventory (Freedman, Rosenthal, Donahoe, Schlundt, & McFall, Reference Freedman, Rosenthal, Donahoe, Schlundt and McFall1978; Gaffney & McFall, Reference Gaffney and McFall1981) was used to assess adolescents’ social problem-solving skill. Adolescents were asked to describe their most likely response to a series of five problematic hypothetical situations. These responses were then rated by coders who were unaware of the hypotheses of the study using a scale of 0 to 10, rating competence in resolving the situation at hand in a way that also makes future problematic situations less likely. Situations included conflicts with peers, romantic partners, and teachers, and situations in which adolescents might be tempted to engage in delinquent behaviors. For example, the participant was asked how they might deal with a situation in which a gym teacher unfairly called the student out and told him/her to do extra push-ups in front of the class. Another asked how the participant might deal with being harshly confronted by a parent about clothing choices just before heading out to a party. This measure was obtained repeatedly at each of the first five waves of data collection using different prompt situations at each wave. Inter-rater reliabilities calculated using the intraclass correlation coefficient ranged from r = .86–.96, which are considered in the “excellent” range for this statistic (Cicchetti & Sparrow, Reference Cicchetti and Sparrow1981). The mean of scores across these five waves was used in analyses and this mean demonstrated good internal consistency (Cronbach's α = .75).
Anxiety (age 18 and 20 years)
The 20-item trait anxiety scale from the State–Trait Anxiety Inventory (Spielberger, Sydeman, Owen, & Marsh, Reference Spielberger, Sydeman, Owen, Marsh and Maruish1999) was used to measure stable individual differences in anxiety proneness. Responses used a 4-point Likert scale to which participants indicated their agreement to statements such as “I worry too much over something that doesn't really matter.” The overall trait anxiety scale has demonstrated strong psychometric properties and external validity (Spielberger et al., Reference Spielberger, Sydeman, Owen, Marsh and Maruish1999), and internal consistency for the scale in this study was high (Cronbach's α = .89 and .90). Scores were averaged across the two assessment periods.
Attachment anxiety in adulthood (age 23–25 years)
Attachment anxiety was measured using the Behavioral Systems Questionnaire (Furman & Wehner, Reference Furman and Buhrmester1994), which assesses attachment styles with a current partner. The five-item preoccupied attachment subscale includes behaviors such as overdependence and neediness. Higher scores indicate greater endorsement of anxious attachment behaviors. The internal consistency for the adult anxious attachment score was good (Cronbach's α = .71).
Anxiety in romantic relationships (age 22 years)
Target participants completed the Multi-Item Measure of Adult Romantic Attachment (Bartholomew & Horowitz, Reference Bartholomew and Horowitz1991; Brennan, Clark, & Shaver, Reference Brennan, Clark, Shaver, Simpson and Rholes1998). From this 36-item measure, anxiety in romantic relationships was assessed. The format for this measure asked participants how they generally felt in romantic relationships, not just in the current relationship. Items were scored on a 7-point scale, ranging from “strongly disagree” to “strongly agree,” with lower scores reflecting lower levels of anxiety and avoidance in romantic relationships. Sample items include questions such as “I get uncomfortable when a romantic partner wants to be very close,” and “I do often worry about being abandoned.” The scales demonstrated good reliability (α = .94–.95).
Results
Preliminary analyses
Means and standard deviations and intercorrelations for the primary variables examined are presented in Table 1. Male gender was strongly related to future blood pressure and was thus included in all analyses. We also examined possible moderating effects of gender on the relation of intensity to future blood pressure. Moderating effects were assessed by creating interaction terms based on the product of the centered main effect variables. No moderating effects of gender were found for any of the analyses reported. Three observations for romantic intensity in adolescence and four for intensity in adulthood were more than 3.5 standard deviations above the mean and were trimmed to 3.5 standard deviations from the mean so as not to unduly influence results (the results would not have differed appreciably had they been kept untrimmed). Residuals from the analyses reported in the following were all consistent with a normal distribution (i.e., skewness and kurtosis both less than ±1.0).
Table 1. Univariate statistics and intercorrelations among primary constructs
Note: Target participant age(s) at time of assessment are in parentheses. SD = standard deviation. ***p < .001, **p < .01, *p < .05.
Primary analyses
Hypothesis 1: Intense romantic involvement in adolescence will predict higher adult resting blood pressure
For all primary analyses, SAS PROC CALIS version 9.4 (SAS Institute, 2015) was employed using full-information maximum likelihood handling of missing data. For hypothesis 1, separate linear regression models were used to examine predictions of SBP and DBP. Analyses began by entering effects of gender and history of childhood illness. Adult BMI was entered next. In the final step, adolescent romantic relationship intensity at ages 17–19 years was entered. The results (Table 2) indicated that, after accounting for effects of gender, prior illness, and current BMI, romantic relationship intensity was a robust predictor of future blood pressure, accounting for increments of approximately 9% and 14% of the variation in adult SBP and DBP, respectively.
Table 2. Predicting adult resting blood pressure from intense romantic involvement in late adolescence
Note: SBP = systolic blood pressure; DBP = diastolic blood pressure. ***p < .001, **p < .01, *p ≤ .05.
Hypothesis 2: Intense romantic involvement in adolescence will be predicted by psychologically controlling maternal behavior and will, in turn, predict a cascade of future romantic and peer relationship qualities reflecting negativity, conflict, and intense involvement in later romantic relationships
The link from psychologically controlling maternal behavior to adolescent romantic intensity was first examined in a regression model that also controlled for participant gender. This link was highly significant (β = .25, p < .001). Psychological control was not predictive, however, of future romantic or peer relationship qualities beyond adolescence or of adult blood pressure. Next, the predictions from intense romantic involvement to adult social functioning were examined. Romantic intensity in late adolescence was significantly predictive of higher future conflict and negativity in close friendships at ages 25–29 years and future romantic relationship intensity at ages 29–31 years, but not of future romantic partner satisfaction at age 26 years (results presented in Table 3).
Table 3. Predicting future relationship qualities from adolescent romantic relationship intensity
Note: ***p < .001, **p ≤ .01, *p < .05. β weights are from final models.
Hypothesis 3: The primary long-term effects of early intense involvement on blood pressure will be mediated via future romantic and peer qualities, with an examination of the possibility of direct effects independent of future relationship behaviors
We next considered the extent to which our indicators of relationship difficulties across early adulthood mediated the effects of intense involvement in adolescence. This also allowed us to assess the extent to which predictors from different developmental stages were unique versus redundant in their prediction of future adult blood pressure. Figure 1 presents effects for the prediction of both SBP and DBP. All temporally sensible paths were considered, with all significant paths diagrammed. The final model fit the data well (goodness of fit index = .98, adjusted goodness of fit index = .91, root mean square error of approximation = 0.03, χ2 (13) = 15.3, p = .29). The indirect effects of adolescent romantic intensity on blood pressure were significant for both SBP (indirect effect = .096, 95% confidence interval (CI) [0.036, 0.156]) and DBP (indirect effect = .1032, 95% CI [0.039, 0.168]). This final model accounted for 49.1% of the variance in SBP and 31.0% of the variance in DBP. This represents increments in prediction for the relationship factors modeled over and above BMI, gender, and history of childhood illness of 18.9% and 25.4% of the variance for SBP and DBP, respectively.
Figure 1. Relationship Predictors of Adult Blood Pressure.
Hypothesis 4: Predictions will exist over and above baseline family income and trait markers of personality characteristics
We examined the big five personality measures as a block predicting both SBP and DBP after accounting for gender and BMI using the procedure described above. In neither case did the block approach significance (all p values > .30), nor were any of the measures individually linked to either measure of blood pressure after accounting for effects of gender and BMI. Hence, these measures were not considered further.
We also examined family income as reported by parents repeatedly across the adolescent years and also found that it was not predictive of either measure of blood pressure; hence this was also not considered further.
Post-hoc analyses
Given the findings above, post-hoc analyses addressed several follow-up questions.
Do predictions from early romantic intensity change if adolescents do or do not have a strong close friendship in this period?
Given the observed links of blood pressure to later close friend relationships, we examined whether the predictions from early romantic intensity would be altered by the quality of close friendships in adolescence (i.e., exacerbated for those with poor-quality friendships; ameliorated for those with good-quality friendships). We examined friendship quality aggregated across ages 17–19 years as a moderator of predictions from intensity to later blood pressure. For both SBP and DBP, close friendship quality did serve as a moderator in the expected direction. Simple slopes analysis using the PROCESS macro version 3.3 (Hayes, Reference Hayes2019) in SAS (SAS Institute, 2015) indicated that, for DBP, the effects of intensity remained significant in a range from +1SD to −1SD from the mean of friendship quality, with the effect becoming smaller as friendship quality increased. For SBP, early romantic intensity remained predictive across low and mean levels of friendship quality but was nonsignificant as a predictor when friendship quality was at 1SD or more above the mean. This suggests that romantic intensity was more likely to predict higher blood pressure when adolescents were lacking in good-quality close friendships. Friendship quality was not directly correlated with either intensity or future blood pressure, however (all p values > .25).
Might the results have been driven by the inclusion of adolescents who were not in a romantic relationship (i.e., who had scores of 0 on intensity)?
To test this possibility, these adolescents were excluded from analyses. When the analyses were rerun with the reduced sample, the effects of intensity were consistently larger than they were in the full sample, indicating that – if anything – including these adolescents reduced the magnitude of the intensity effects observed.
Might blood pressure predictions have simply reflected the underlying effects of attachment insecurity, attachment anxiety, general anxiety, or low popularity and poor social skills?
The analyses considered the degree of security/insecurity in young adults’ state of mind at age 24 years, reported attachment anxiety in romantic relationships at age 23–25 years, reported anxiety at ages 18–20 years, and overall level of popularity and social problem-solving skill earlier in adolescence as predictors of either measure of adult blood pressure. No significant predictions were found to adult blood pressure, nor were these variables linked to adolescent romantic intensity – with the exception of popularity, which inversely related to adolescent romantic intensity (r = −.16, p = .03).
Discussion
This prospective study identified a robust link from intense adolescent romantic involvement to higher adult resting blood pressure at 29–31 years of age. Predictions were observed to measures of both SBP and DBP, and remained strong even after accounting for participant gender, history of childhood illness, and current BMI. Intense early romantic involvement was found to be predictable from a prior experience of autonomy-undermining maternal behavior. The links from intense romantic involvement to adult blood pressure appeared to be at least partially mediated via problematic peer and romantic relationships in adulthood. Even after accounting for these mediators, however, adolescent romantic intensity continued to account for a substantial variance in adult blood pressure.
Predictions from adolescent romantic intensity were slightly stronger to DBP than SBP, which is important given that it is DBP in young adults that is most strongly predictive of future cardiovascular disease (Franklin et al., Reference Franklin, Larson, Khan, Wong, Leip, Kannel and Levy2001). Of note, although medical intervention is only considered helpful, and therefore indicated, at higher levels of blood pressure, research has consistently found that the link between blood pressure and cardiovascular disease is linear from the low–normal to high levels of blood pressure (Mason et al., Reference Mason, Manson, Sesso, Albert, Chown, Cook and Glynn2004; Psaty et al., Reference Psaty, Furberg, Kuller, Cushman, Savage, Levine and Lumley2001). Thus, across all levels, from normal to elevated, any degree of increase in blood pressure is equally problematic in terms of its prediction of long-term risks.
The overall magnitude of the predictions obtained is also noteworthy. Even after accounting for BMI and gender, the effect of adolescent romantic intensity accounted for 9% of the variance in SBP and 14% in DBP. This effect size translates into a 5–6 mm mercury difference in DBP between individuals who are one standard deviation above versus one standard deviation below the mean in adolescent romantic intensity. This effect size – derived from a single measure obtained more than a decade prior – is of the same approximate magnitude as the short-term effects of a number of current medications used to treat hypertension (Wald, Law, Morris, Bestwick, & Wald, Reference Wald, Law, Morris, Bestwick and Wald2009).
Romantic intensity in adolescence predicted both future romantic intensity in adulthood as well as difficulties in interactions with nonromantic peers. This provides additional evidence of the maladaptive nature of intense late-adolescent relationships and also suggests that there may be multiple pathways by which this intensity is linked to higher blood pressure. More generally, these findings highlight a fundamental question – Just what is it about romantic intensity in late adolescence that leads to these problematic associations? It is important to note that relationship intensity in adolescence is quite distinct from relationship quality (Collins, Reference Collins2003; Davila et al., Reference Davila, Mattanah, Bhatia, Latack, Feinstein, Eaton and Zhou2017). Although relationships can be sources of great support, they can also create great distress, and the evidence within adolescence suggests that more intense relationships are more apt to create such distress (Giordano et al., Reference Giordano, Soto, Manning and Longmore2010; Kaestle & Halpern, Reference Kaestle and Halpern2005). Giordano et al. (Reference Giordano, Soto, Manning and Longmore2010) found that the provision and receipt of more instrumental support within teen dating relationships was associated with an increased risk of dating violence – perhaps too much of a “good thing” could signal the presence of an inappropriately-consuming and demanding relationship. It also appears that it is intensity out of keeping with developmental norms in adolescence that is most problematic: adolescent relationships are often intense, but it was when intensity was higher than this normative intensity that it was predictive of future blood pressure.
However enticing an intense romantic relationship may be in late adolescence, it is nevertheless going to take away time and effort from close friendships and nonromantic socialization and activities, all of which appear to be particularly important to move development forward at this stage (Allen et al., Reference Allen, Narr, Kansky and Szwedo2020; Roisman, Aguilar, & Egeland, Reference Roisman, Aguilar and Egeland2004). Of note, in this study, to the extent this did not seem to be occurring (i.e., close friendship quality was rated as high by friends), the effects of intensity were greatly reduced or eliminated. Further, it is likely that the late adolescent will have difficulty managing the intense emotions generated by relationships that soak up large amounts of their time and incorporate relatively little interaction with others outside the relationship. This intensity may reflect a premature transfer of attachment functions from the parent to the romantic partner in a way that the partner is unlikely to be able to handle competently over time (Collins, Welsh, & Furman, Reference Collins, Welsh and Furman2009). The combined and potentially mildly suffocating effect of these factors is likely to give rise to conflict and negativity in the current relationship and to more intense distress if and when that relationship ends. Such conflict and rejection has been posited to lead to perceptions of lack of safety, which have been reliably linked to numerous long-term stress responses (Slavich, Reference Slavich2020). Given the formative nature of relationships during this period, the problems generated by relationship intensity are also likely to set a template for expectations of conflict, negativity, and threats to autonomy in future relationships. Together, these factors may account both for links to future relationship outcomes and potentially to a pattern for handling relationship stress that is linked to higher blood pressure.
It is notable that the links of intensity to adult blood pressure were not fully mediated via future relationship intensity or other later relationship quality measures, even though these were all assessed far more proximately to the blood pressure assessment. This finding is striking given the long time period across which predictions were observed, but it is also consistent with emerging research that has identified links from adolescent relationship behaviors to long-term health outcomes (e.g., interleukin-6 levels and reported overall health) that did not appear to be mediated by later relationship qualities (Allen et al., Reference Allen, Loeb, Tan and Narr2017; Reference Allen, Uchino and Hafen2015). It remains entirely possible, of course, that the effects of intensity were mediated by other relationship factors which were not measured across this period, as discussed further in the following. These findings are consistent, however, with findings from early childhood research suggesting that health-related effects of exposure to stress can exist even when such stress does not continue into later life (Pollitt et al., Reference Pollitt, Kaufman, Rose, Diez-Roux, Zeng and Heiss2007). Both animal and human research suggests that adolescence may be a period during which stress reaction systems are particularly malleable and vulnerable to aversive social experience (Quevedo, Johnson, Loman, Lafavor, & Gunnar, Reference Quevedo, Johnson, Loman, Lafavor and Gunnar2012; Romeo, Reference Romeo2010).
It is noteworthy that the simple presence of a romantic relationship in adolescence has been found to predict future depressive symptoms, with adolescents who are more preoccupied with such relationships being most likely to experience such symptoms (Davila et al., Reference Davila, Steinberg, Kachadourian, Cobb and Fincham2004). Given that even intense adolescent romantic relationships are likely to ultimately end, it may be that the strong distress experienced upon the ending of such a relationship is actually a factor leading to higher adult blood pressure. The previously observed links of hyperarousal following romantic dissolution to changes in patterns of blood pressure reactivity in adulthood add plausibility to this notion (Sbarra, Law, Lee, & Mason, Reference Sbarra, Law, Lee and Mason2009). Stressors in adolescence, such as may be created by overly intense relationships or their eventual dissolution, have the potential not only to alter metabolic systems and anatomic structures, but also more basic social and physiological patterns of stress responsivity (Ben-Shlomo & Kuh, Reference Ben-Shlomo and Kuh2002; Goff et al., Reference Goff, Gee, Telzer, Humphreys, Gabard-Durnam, Flannery and Tottenham2013; McEwen, Reference McEwen2007).
An alternative explanation of the observed findings, however, is that intensity in adolescence is a marker of a broader pattern of disturbed relationships, which is a primary driver of later blood pressure. The finding that romantic intensity was predicted by psychologically controlling maternal behavior is consistent with a perspective in which intensity reflects a pattern of overinvolvement in intimate relationships, perhaps learned from parents and then cascading forward into new relationships (Oudekerk et al., Reference Oudekerk, Allen, Hessel and Molloy2015). It may be that a tendency toward dependence, possessiveness, and/or lack of boundaries in relationships learned in one's family of origin drives both romantic intensity as well as future difficulties. The finding that intensity was much less predictive for adolescents with strong close friendships is consistent with the notion that intensity may be predictive primarily when it reflects a broader pattern of relationship disturbance. It should be noted, however, that neither the most widely used measures of personality in adulthood nor other markers of social functioning in adolescence were predictive of adult blood pressure; this makes it more difficult to argue that intensity is simply a marker for a broad personality factors as opposed to more subtle relational processes.
Overall, it is perhaps most plausible that multiple explanations for the intensity predictions apply: intensity may mark a problematic relationship tendency and it may also establish or further a pattern of problematic relationship behavior that ultimately influences blood pressure. One piece of evidence weighing in favor of the latter possibility is that romantic intensity predicted precisely the type of relationship factors in adulthood (e.g., evidence of conflict and negativity) that have been linked to adult blood pressure in shorter-term, cross-lagged studies (Holt-Lunstad et al., Reference Holt-Lunstad, Uchino, Smith, Olson-Cerny and Nealey-Moore2003). Clearly, further research is needed to understand which of a variety of possible weathering and mediated mechanisms best account for the intensity effects observed.
Several limitations of this study also warrant consideration. As should be clear from the above discussion, although the results are consistent with findings from research in adulthood and extend these to a far earlier stage of the life span, direct causal inferences cannot be supported by these data. Relatedly, this study also did not have measures of blood pressure available from the adolescent era. It is thus possible that blood pressure levels were already heightened in adolescence and that what was observed in this study may be linked to other prior experiences, to a reverse causal process in which blood pressure somehow influences relationship qualities, or to unmeasured temperamental or genetic predispositions that influence both blood pressure and relationship experiences. Of course, even if intensity were simply a marker of risk and not a causal agent, it would still warrant consideration given that it was assessed at an age when blood pressure is likely to appear in the normal range and thus not warrant attention.
In addition, the assessment of romantic intensity was made via a novel measure. This measure, comprising the duration of the relationship and time spent alone with the partner, certainly captures some aspects of intensity and was quite similar to the way others have conceptualized problematically intense relationships (Giordano et al., Reference Giordano, Soto, Manning and Longmore2010). The combination of relationship duration and amount of time spent alone with a partner also seems likely to forecast the degree of distress felt upon relationship dissolution; this was not assessed in this study but would be a promising topic for future research. The measurement of romantic intensity is not well developed overall, however, and other measures of intensity (e.g., self- and partner-reports of the experience within the relationship, delineation of activities together, observational data, etc.) will be needed to fully flesh out this construct.
It should also be noted that although this study posits a developmental perspective on intensity – suggesting it is more problematic in adolescence than later in life – even in adulthood long-duration relationships that involved very large amounts of time spent alone with a partner were somewhat linked to higher blood pressure. This raises the possibility that there may be a facet of such relationships, such as emotional overload, that is linked to higher blood pressure at any age, though this is also a question for future research to pursue. An additional limitation of this study is that we did not have access to measures of biological markers of stress response patterns that might further elucidate the nature of the links observed.
Overall, however, the findings of this study contribute to a developing picture suggesting robust links between relational markers of developmental psychopathology in adolescence and long-term physical health outcomes. Unfortunately, to date, current national recommendations from the primary government agencies tasked with the prevention of age-related disease fail to consider any social relationship factors in adolescence or early adulthood (Centers for Disease Control and Prevention, 2017; Holt-Lunstad, Robles, & Sbarra, Reference Holt-Lunstad, Robles and Sbarra2017). The present findings, together with similar findings from related research, suggest an urgent need for this to change going forward.
Funding Statement
This study was supported by grants from the National Institute of Child Health and Human Development and the National Institute of Mental Health (5R37HD058305-23, R01 HD058305-16A1, and R01-MH58066).
Conflicts of Interest
None.
