Serum Urate Trajectory in young adulthood and incident Cardiovascular Disease Events by Middle Age: Coronary Artery Risk Development in Young Adults (CARDIA) Study

Abstract

Serum urate levels have been shown to be correlated with risk for incident cardiovascular disease (CVD) events among middle-aged or older adults. However, serum urate trajectory during young adulthood and its association with CVD events has been understudied. Using serum urate measurements collected at baseline and 10, 15, 20 years after baseline from 3563 Coronary Artery Risk Development in Young Adults (CARDIA) participants [mean age 25.1±3.6 (18–30) years at baseline (Year 0, 1985–86); 46.3% African American; 56.1% female], we determined sex-specific serum urate trajectories using SAS PROC TRAJ. We estimated hazard ratios (HR) for incident CVD events (coronary heart disease, heart failure, and stroke) occurring after the Year 20 exam through 2017. We identified 3 serum urate trajectories by sex, including low-stable (n=1251), moderate-stable (n=1761), and high-increasing (n=551). Over a median 10.6 years of follow-up, 157 incident CVD events occurred. Participants among the high-increasing trajectory group had 2.89 (95% confidence interval, 1.88–4.43) times greater risk for CVD compared to the low-stable trajectory group. The association was attenuated after adjustment for blood pressure levels during young adulthood. In conclusion, high-increasing serum urate trajectory during young adulthood was associated with incident CVD by middle age, and the association may be explained by blood pressure levels during the exposure period.

Summary

High-increasing serum urate trajectory during young adulthood was associated with incident CVD by middle age, and the substantial attenuation of this association by blood pressure levels during the exposure period suggested that blood pressure may be an important mediator.

Introduction

Cardiovascular disease (CVD) is a leading cause of death in the US. Higher serum urate or uric acid concentrations have become increasingly common in the US,¹ and associations with higher CVD incidence or mortality have been reported.^2–8 Most prior studies assessed serum urate or uric acid levels at a single time point in middle-aged and older adults. Few studies examined whether serum urate levels during young adulthood are associated with CVD events by middle age. Investigators from the Coronary Artery Risk Development in Young Adults (CARDIA) Study have reported an association between serum urate levels on a single occasion during young adulthood and CVD events.⁹ Serum urate levels tend to change with age, and patterns may differ among individuals.^1,10–13 However, little is known regarding whether long-term serum urate trajectory patterns exist from young adulthood to middle age, and how those patterns may be associated with CVD events. Characterizing trajectory patterns may help identify groups of young adults at high risk for CVD events who would benefit from interventions to alter the rise of serum urate and prevent CVD events.

Using a dataset from the CARDIA Study, we identified serum urate trajectories from young adulthood to midlife and assessed the associations with CVD events by middle age. We also assessesd whether serum urate trajectories would better predict incident CVD compared to a measurement of serum urate at a single time point.

Methods

Data and Material Disclosure Statement

Data documentation for CARDIA is publicly available at cardia.dopm.uab.edu. Data used in this paper are available on reasonable request from the cardia coordinating Center.

Study design and sample

CARDIA is a prospective cohort study designed to investigate the development and progression of cardiovascular risk factors and disease. A more thorough description of the study design and conduct has been previously published.^14–16 Briefly, in 1985–1986, 5115 black and white men and women aged 18 to 30 years were recruited from 4 urban sites across the United States, including Birmingham, Alabama; Chicago, Illinois; Minneapolis, Minnesota; and Oakland, California. CARDIA participants have been followed for over 30 years with detailed assessment of demographic and clinical factors, including self-reported anthropometric and laboratory measures. Race was determined by self-report at study commencement and verified at the Year 2 exam. Cohort members have been invited to participate in 9 examination cycles including the baseline exam (Year 0) and follow-up exams at Years 2, 5, 7, 10, 15, 20, 25, and 30. Retention rates have been high across these exams (91%, 86%, 81%, 77%, 74%, 72%, 72%, and 71%, respectively). Annual follow-up contact to identify clinical events has been successfully made in 91.2% of the participants during the past 5 years. The study was approved by the Institutional Review Board at all study sites, and all participants provided written informed consent.

Measurements of serum urate

In preparation for each exam, participants were asked to fast for at least 12 hours and to avoid smoking and heavy physical activity. Fasting blood was drawn, aliquoted, and serum samples were shipped (in dry ice) to a central laboratory and stored at −70℃ until analysis.¹⁷ Serum urate concentration was measured at the Year 0 exam using the uricase method.¹⁸ A colorimetric assay (a modification of the uricase method, in which urate is oxidized to peroxide)¹⁹ was used at the Year 10,15, and 20 exams. At Year 20, serum urate was assessed as a part of the Young Adult Longitudinal Trends in Antioxidants (YALTA) ancillary study to CARDIA, wherein a validation study of 105 samples from Year 0 were re-assayed at Year 17 using Year 15 methods, and a correlation of 0.99 was observed, compared with the original baseline data. Recalibrated values were used for the Year 10, 15 and 20 serum urate levels.²⁰

Other covariates

At each examination, the CARDIA Study collected age, sex, race, and clinical characteristics: body mass index (BMI) (calculated as weight in kilograms divided by height in meters squared), current smoking, systolic blood pressure (SBP), diastolic blood pressure (DBP), anti-hypertensive medication use, total cholesterol, high-density lipoprotein (HDL)-cholesterol, diabetes (DM) and estimated glomerular filtration rate (eGFR). The CKD-EPI equation was used to calculate eGFR.²¹ Female participants were asked about their menopausal state at the Year 20 exam.

Assessment of any fatal or nonfatal CVD events

Events were ascertained during annual participant contacts; death certificates and relevant medical and hospital records were obtained. The CARDIA Endpoints Surveillance and Adjudication Committee then adjudicated CVD events according to the study protocol.²² Incident CVD was defined as adjudicated definite or possible coronary heart disease (CHD; including myocardial infarction, angina pectoris, and death due to CHD), stroke, transient ischemic attack, heart failure (HF), or peripheral artery disease. No participant at Year 0 had a history of one or more clinical CVD events. Follow-up time used in our analyses was calculated as the number of days between the Year 20 examination date and the day of incident CVD, the last day the participant was contacted, or August 31, 2017.

Statistical analysis

Each trajectory group was defined using at least three measurements obtained at baseline and Year 10, 15, and/or 20. We excluded 1457 participants who had fewer than three measurements of serum urate. Serum urate trajectories were modeled among the remaining 3657 CARDIA participants. The Year 20 exam date was defined as the time origin for time-to-event analysis. We used latent class models (SAS PROC TRAJ) to identify subgroups that share a similar underlying trajectory in serum urate with examination age as the time scale.^23–25 Because serum urate substantially differs by sex,^26–27 we also determined serum urate trajectories by sex. Based on the Bayesian Information Criterion, we selected the number and shape of trajectories that best represented distinct patterns of change in serum urate. Model adequacy was assessed by comparing posterior predicted probabilities of final group assignment and comparing odds ratios of final group classification (and 95% confidence intervals [CIs]). Trajectories were labeled by initial serum urate levels (low, moderate or high) and pattern of change in serum urate with age (increasing, stable or decreasing). Trajectories were depicted by plots which represented the mean ± standard error of serum urate in each group in 3-year bins, starting at 18–20 years old. Trajectory names were selected to describe these plots.

Analyses of variance and chi-square tested the significance of differences in continuous and categorical variables, respectively. Cox proportional hazards models were used to estimate adjusted hazard ratios (HRs) and 95% CIs for incident CVD. Because age at baseline, sex and race are associated with both exposure and outcomes,^10,26–29 these factors were adjusted in Model 1. Further adjustments were made for means of BMI; SBP; DBP; total cholesterol and HDL cholesterol collected at the Year 0, 10, 15, and 20 exams; the maximum of current smoking state across exams (no=0 or yes=1); use of anti-hypertensive medication ever (no=0 or yes=1); ever DM (no=0 or yes=1) at Year 0, 10, 15, or 20; and the lowest eGFR at Years 0, 10, 15, and 20.

We conducted five sensitivity analyses. First, we adjusted for the highest creatinine measurement recorded across Year 0, 10, 15, and 20 exams instead of eGFR. Second, we assessed whether serum urate levels at Year 20, instead of serum urate trajectories, were associated with incident CVD events. Serum urate levels at Year 20 were used as both continuous variables and categorical variables, which were divided into 3 groups based on sex-specific cutpoints selected to match the distributions of the corresponding sex-specific trajectories. We calculated C-statistics (95% CIs) as described by Uno et al.³⁰ to evaluate the predictive utility of serum urate trajectory groups and categorical serum urate at Year 20 in Cox proportional hazards models for CVD events. To quantify how the model reclassified individuals after including the trajectory group, we calculated the net reclassification improvement. Third, we divided participants into four groups based on whether they were above or below the median of serum urate at Years 0 and 20, respectively, or above or below the reference value of serum urate (i.e., 6.8mg/dL) defined in the 2020 American College of Rheumatology Guideline for the Management of Gout^31–32 at Years 0 and 20, respectively. Fourth, we adjusted for menopause at Year 20 when modeling for women, since menopause may change serum urate trajectories.³³ Fifth, we excluded participants who reported taking diuretics (at any clinical exam—baseline, Year 10, 15, or 20). All analyses were completed using SAS software version 9.4 (SAS Institute Inc). Two-sided P<.05 was considered statistically significant.

Results

We excluded 1457 participants who had fewer than 3 measurements of serum urate through Year 20. Participants excluded in the current analysis were more likely to be male, African American, and current smokers, and have higher BMI and prevalent hypertension compared to those who were included (Table S1, please see http://hyper.ahajournals.org.). Of the 3657 remaining participants, we excluded 94 who had evidence of CVD by Year 20, leaving 3563 participants for the analysis.

Mean levels of serum urate (mg/dl) were 5.2±1.4 (Year 0: n=5048), 5.4±1.4 (Year 10: n=3869), 5.5±1.5 (Year 15: n=3604) and 5.7±1.5 (Year 20: n=3147), respectively. We identified serum urate trajectories by sex: 3 trajectories in men and 3 similarly-shaped trajectories in women, who as a group had lower concentrations of serum urate compared to men. Sex-specific serum urate levels at each examination among each trajectory group are shown in Table 1. The three identified trajectory groups included low-stable (men: n=452, 28.9%; women: n=799, 40.0%), moderate-stable (men: n=838, 53.5%; women: n=923, 46.2%), and high-increasing (men: n=275, 17.6%; women: n=276, 13.8%) (Figure 1). Among both men and women, the high-increasing group had higher BMI, and more prevalence of hypertension, DM, and metabolic syndrome at Years 0 and 20 compared to the low- and moderate-stable groups. Among women, those of African American race and current smokers were more prevalent in the high-increasing group compared to the low- and moderate-stable groups (Table 2).

Table 1.

Serum urate levels at each examination by sex-specific serum trajectory groups

	Serum urate, mean (SD), mg/dL
Exam year	Men			Women
	Low-stable (n=452)	Moderate-stable (n=838)	High-increasing (n=275)	Low-stable (n=799)	Moderate-stable (n=923)	High-increasing (n=276)
Year 0	5.16 (0.88)	6.25 (0.83)	7.57 (1.05)	3.76 (0.65)	4.75 (0.74)	5.60 (1.00)
Year 10	5.17 (0.66)	6.44 (0.75)	8.01 (0.99)	3.86 (0.56)	4.85 (0.61)	6.16 (0.95)
Year 15	5.20 (0.65)	6.48 (0.75)	8.15 (0.97)	3.87 (0.56)	5.01 (0.65)	6.40 (0.94)
Year 20	5.34 (0.69)	6.73 (0.82)	8.38 (1.07)	4.06 (0.63)	5.26 (0.73)	6.95 (1.09)

Abbreviations: SD=standard deviation

Figure 1. Sex-specific trajectories of serum urate in CARDIA across age.

Serum urate trajectories in each sex were identified by latent class models (SAS PROC TRAJ), using serum urate levels at the Year 0, 10, 15 and 20 exams. Men and women were categorized into 3 trajectories: the low-stable, moderate-stable, and high-increasing groups. Numbers (sex-specific %) of CARDIA participants in each class were described under the figure. The plots in each trajectory represent mean ± standard error of serum urate in 3 year bins, starting at 18–20 years old.

Table 2.

Demographic characteristics of participants by sex-specific serum urate trajectory groups

Characteristics	Men				Women
	Low-stable (n=452)	Moderate-stable (n=838)	High-increasing (n=275)	p	Low-stable (n=799)	Moderate-stable (n=923)	High-increasing (n=276)	p
Age at baseline, mean (SD), years	25.3 (3.56)	24.9 (3.48)	25.0 (3.6)	0.09	25.3 (3.6)	25.0 (3.6)	24.8 (3.9)	0.2
African American, %	45.1	40.6	46.9	0.1	44.6	48.2	63.0	<.0001
Body mass index at the Year 0, mean (SD), kg/m2	23.3 (3.2)	24.3 (3.5)	26.4 (4.8)	<.0001	22.6 (4.1)	24.8 (5.5)	28.4 (6.6)	<.0001
Body mass index at the Year 20, mean (SD), kg/m2	27.0 (7.0)	28.8 (5.2)	32.1 (6.8)	<.0001	26.8 (6.2)	30.0 (7.2)	36.6 (9.0)	<.0001
Current smoker at the Year 0, %	28.6	27.3	24.4	0.5	22.5	27.2	34.4	0.0004
Current smoker at the Year 20, %	21.2	19.8	20.0	0.8	15.0	18.8	23.3	0.008
Hypertension at the Year 0, %	3.3	3.7	7.3	0.02	1.9	2.9	6.2	0.001
Hypertension at the Year20, %	10.3	20.1	36.0	<.0001	11.8	22.7	43.5	<.0001
Diabetes at the Year 0, %	0.9	0.4	0.0	0.2	0.3	0.4	1.1	0.2
Diabetes at the Year 20, %	3.9	6.8	13.2	<.0001	2.3	7.0	14.6	<.0001
Total cholesterol level at the Year 0, mean (SD), mg/dL	171.4 (32.1)	176.5 (32.4)	184.9 (38.6)	<.0001	177.4 (33.1)	176.9 (30.6)	180.8 (33.9)	0.2
Total cholesterol level at the Year 20, mean (SD), mg/dL	181.4 (35.3)	186.8 (36.4)	190.4 (40.2)	0.007	185.7 (33.4)	184.8 (32.4)	189.7 (35.8)	0.1
Estimated glomerular filtration rate at the Year 0, mean (SD), mL/min/1.73m2	126.1 (21.9)	122.8 (19.5)	118.7 (21.5)	<.0001	118.8 (24.3)	115.9 (21.2)	118.4 (22.2)	0.02
Estimated glomerular filtration rate at the Year 20, mean (SD), mL/min/1.73m2	99.2 (22.3)	96.5 (21.8)	95.8 (23.9)	0.08	98.8 (24.1)	96.3 (23.1)	97.1 (29.5)	0.1

Abbreviations: SD=standard deviation

During a median follow-up of 10.6 years, 157 CVD events occurred between Year 20 and the end of event follow-up. There was no significant multiplicative sex interaction in the association between serum urate trajectories and CVD events (interaction p=0.68). Thus, we conducted aggregated analyses as well as analyses by sex. As shown in Table 3, men in the high-increasing trajectory group had 2.46 (95% CI, 1.39–4.35) times greater HR for incident CVD compared to those in the low-stable group; the corresponding HR for this association in women was 2.94 (95% CI, 1.53–5.65). Among men, CVD risk associated with the high-increasing trajectory was attenuated after adjustment for SBP, DBP and use of anti-hypertensive medication. Among women, CVD risk associated with the high-increasing trajectory was attenuated after adjustment for BMI and current smoking. Overall, the high-increasing trajectory group had greater HR for incident CVD compared to the low-stable trajectory group (HR: 2.89; 95% CI, 1.88–4.43). The association remained statistically significant after adjustments for BMI, smoking and DM (Models 3 and 4), but was attenuated after adjustment for SBP, DBP, and antihypertensive medication use (Model 5). In the full model including blood lipids, BP, DM and eGFR (Model 6), the high-increasing trajectory was not significantly associated with a higher risk for incident CVD events compared to the low-stable trajectory (HR: 1.34; 95% CI, 0.83–2.17).

Table 3.

Hazard ratios for incident cardiovascular disease in sex-specific and pooled serum urate trajectory groups

Trajectory group	% (n/N)*	Hazard ratio (95% CI)
		Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Men  Low-stable	4.42 (20/452)	reference	reference	reference	reference	reference	reference
Moderate-stable	5.01 (42/838)	1.14 (0.67–1.94)	1.16 (0.68–1.98)	1.07 (0.63–1.83)	1.05 (0.61–1.79)	0.92 (0.53–1.58)	0.85 (0.49–1.47)
High-increasing	10.55 (29/275)	2.46 (1.39–4.35)	2.43 (1.38–4.30)	1.93 (1.05–3.54)	1.89 (1.03–3.47)	1.59 (0.86–2.94)	1.44 (0.77–2.71)
Women  Low-stable	2.25 (18/799)	reference	reference	reference	reference	reference	reference
Moderate-stable	3.25 (30/923)	1.45 (0.81–2.61)	1.43 (0.80–2.57)	1.28 (0.71–2.33)	1.17 (0.64–2.14)	1.07 (0.58–1.96)	0.94 (0.50–1.74)
High-increasing	6.52 (18/276)	2.94 (1.53–5.65)	2.61 (1.35–5.05)	1.99 (0.97–4.10)	1.56 (0.75–3.26)	1.41 (0.68–2.93)	1.15 (0.54–2.46)
Pooled  Low-stable	3.04 (38/1251)	reference	reference	reference	reference	reference	reference
Moderate-stable	4.09 (72/1761)	1.36 (0.92–2.01)	1.30 (0.87–1.92)	1.17 (0.78–1.74)	1.11 (0.75–1.66)	0.97 (0.64–1.45)	0.89 (0.59–1.34)
High-increasing	8.53 (47/551)	2.89 (1.88–4.43)	2.57 (1.67–3.95)	2.01 (1.26–3.19)	1.85 (1.16–2.96)	1.51 (0.95–2.42)	1.34 (0.83–2.17)

Participants were categorized into three trajectory groups (low-stable, moderate-stable, and high-increasing groups) by sex and pooled sex-specific groups using serum urate levels collected at baseline and at the Years 10, 15, and 20 after baseline.

^*Cardiovascular disease event number divided by total number in each group.

Model 1: Unadjusted

Model 2: Adjusted for age at baseline, sex (when applicable) and race

Model 3: Adjusted for variables in Model 2 + body mass index and current smoker

Model 4: Adjusted for variables in Model 3 + diabetes

Model 5: Adjusted for variables in Model 3 + systolic blood pressure, diastolic blood pressure and anti-hypertensive medication

Model 6: Adjusted for variables in Model 5 + total cholesterol, high-density lipoprotein cholesterol, diabetes and estimated glomerular filtration rate

Sensitivity analyses

We adjusted for serum creatinine levels instead of eGFR, and the point estimates for CVD events associated with serum urate trajectories were similar to those in the primary analyses (Table S2). We also calculated the association between a single measure of serum urate at Year 20 and CVD events. In the unadjusted model, higher serum urate at Year 20 was associated with CVD events (HR: 1.27; 95% CI, 1.14–1.41), and participants in the high group of serum urate at Year 20 had a higher CVD risk compared to the low group (HR: 2.09; 95% CI, 1.35–3.23). However, the association was attenuated after adjustment for BMI and smoking (Table S3). In models adjusted for age, sex, race, BMI, smoking, SBP, DBP, antihypertensive medication use, total cholesterol, HDL cholesterol, DM, and eGFR, the C-statistic was 0.77 (95% CI, 0.73–0.81), and this did not increase in the models that included serum urate trajectories or categorical serum urate at Year 20 (C-statistic: 0.77 (95% CI, 0.73–0.81), p=0.82, and 0.76 (95% CI, 0.72–0.81), p=0.99, respectively). Net reclassification improvement (NRI) revealed that serum urate trajectories had a slightly better predictive value for incident CVD than serum urate at Year 20 (NRI: 0.14; 95% CI, 0.05 – 0.24). Participants were divided into four groups (low-low (n=1187), low-high (n=385), high-low (n=364), and high-high (n=1187)) based on whether their measurements were above or below the median of serum urate levels at Years 0 and 20. In an unadjusted model, HRs for CVD events associated with the low-high group, the high-low group, and the high-high group were 1.88 (95% CI: 1.10, 3.21), 1.18 (95% CI: 0.63, 2.23), and 1.82 (95% CI: 1.21, 2.74), respectively, compared to the low-low group (Table S4). All associations were attenuated after adjustment for age and sex. Participants were also divided into four groups based on whether their measurements were above or below the reference of serum urate (6.8mg/dL) defined in the 2020 American College of Rheumatology Guideline for the Management of Gout. In an unadjusted model, HRs for CVD events associated with the low-high group (n=428), the high-low group (n=124), and the high-high group (n=221) were 1.83 (95% CI: 1.18–2.85), 1.73 (95% CI: 0.80–3.74), and 2.67 (95% CI: 1.62–4.41), respectively, compared to the low-low group (n=2298) (Table S5). All associations were attenuated after adjustment for age and sex, BMI, and current smoking. The proportions of postmenopausal women at Year 20 in low-stable, moderate-stable, and high-increasing serum urate trajectories were 12.4%, 11.1%, and 13.0%, respectively. When we adjusted for menopause, the point estimates for CVD events associated with serum urate trajectories were similar to those in the primary analyses (Table S6). Prevalence of diuretic use at each exam year were very low: 9/5114 (0.18%) at Year 0, 15/3949 (0.38%) at Year 10, 34/3671 (0.93%) at Year 15 and 63/3549 (1.78%) at Year 20. Serum urate levels were higher in participants taking diuretics (Table S7). After excluding participants taking diuretics, high-increasing serum urate trajectory was associated with CVD events (unadjusted HR: 2.71; 95% CI, 1.75–4.19). However, adjustment for BP attenuated this association (Table S8).

Discussion

In a community-based sample of African Americans and whites, we identified three serum urate trajectory patterns from young adulthood to midlife: low-stable, moderate-stable, and high-increasing. The high-increasing versus low-stable serum urate trajectory was associated with a higher risk for incident CVD events in later life, independently of BMI, smoking, diabetes, and lipid parameters during the exposure period (i.e., at baseline and Years 10, 15, and 20). However, the association of high-increasing serum urate trajectory with CVD events was attenuated to non-significance after adjustment for BP levels during the exposure period. Serum urate trajectories did not predict incident CVD better than serum urate at a single time point.

In a systematic review and meta-analysis conducted by Kim et al.,³⁴ which included 26 prospective cohort studies of 402997 adults, hyperuricemia (defined as 5.6–7.7 mg/dl in men and 4.7–7.0 mg/dl in women) was associated with a 34% greater risk for incident coronary heart disease (CHD) events and a 46% greater risk for CHD mortality. The association was independent from age, sex, hypertension, diabetes, smoking and hypercholesterolemia. In a meta-analysis that included prospective cohort studies of adults (sample sizes ≥100) with no CVD at baseline and a follow-up of at least 1 year, Braga et al. demonstrated that hyperuricemia (defined as 6.6–6.8 mg/dL in men and 5.2–6.3 mg/dl in women) was associated with higher risk for incident CHD, independently of age, sex, hypertension, diabetes, smoking, hypercholesterolemia, and alcohol drinking.³⁵ Most studies included in these meta-analyses enrolled middle-aged and older adults, and serum uric acid levels at a single time point were used for the analyses. The current study extends knowledge by demonstrating that the high-increasing versus low-stable serum urate trajectory from young adulthood to middle age was associated with a higher risk for incident CVD events by middle age, but the association was attenuated after adjustment for BP levels during the exposure period.

In a previous CARDIA Study investigation, Gaffo et al found that higher serum urate levels at the Year 0 exam were associated with a 25% higher risk for developing hypertension by the 20-year follow-up in men.³⁶ In a prospective cohort study, which included 5758 participants aged between 30 and 60 years without hypertension, Ma et al. identified five distinct urate trajectories, including low-stable, low-increasing, moderate-increasing, high-decreasing, and high-stable patterns.³⁷ Compared with the low-stable group, the moderate-increasing group had the highest risk, with adjusted HRs of 2.48 (95%CI, 1.64–3.74) in women and 1.84 (95% CI, 1.43–2.35) in men, followed by the high-stable group, with adjusted HRs of 1.97 (95% CI, 1.29–3.01) in women and 1.45 (95% CI, 1.15–1.88) in men, and the low-increasing group, with adjusted HRs of 1.83 (95% CI, 1.20–2.79) in women and 1.42 (95% CI, 1.10–1.83) in men. However, it remains uncertain whether higher BP among individuals with higher serum urate or uric acid levels could explain their CVD risk. Most of the previous studies that have shown an association between serum urate or uric acid levels and CVD events adjusted for a single measurement of BP levels, which might not reflect the potential effects of long-term exposure to BP on CVD risk. The current study extends knowledge by demonstrating that CVD risk associated with high-increasing versus low-stable serum urate trajectories from young adulthood to middle age might be explained by BP levels during the exposure period.

Strengths of this study include characterizing long-term trajectories during young adulthood before the onset of CVD in a large cohort of black and white Americans. This study has several limitations. First, the amount of missing data for serum urate differed across each examination. Serum urate at Year 0 was missing for 1.3% of participants, and for 38.5% of participants at Year 20. However, the average number of serum urate measurements available for inclusion in our analyses was similar across trajectory groups (3.74, 3.71 and 3.71 in the low-stable, moderate-stable and high-increasing groups, respectively). Second, we had no information regarding participants’ serum urate patterns prior to their entry into the CARDIA Study. Third, there is variability in serum urate, depending on factors like timing of collection. Potential variability and diurnal fluctuations in serum urate measurements within individuals were not assessed in the current study. Fourth, residual confounding may not be ruled out, as in all observational studies.

Perspectives

A high-increasing trajectory of serum urate levels during young adulthood is a strong risk factor for CVD events in middle age. BP levels during the exposure period may be an important mediator of this association. In the clinical setting, electronic health record systems could be programmed to calculate a person’s serum urate trajectory over time using multiple measurements, which could help clinicians identify individuals at high risk for CVD events. Further studies are warranted to determine whether the assessment of serum urate trajectories in addition to assessing serum urate levels in clinical practice improves detection and subsequent medical management of young adults at higher risk for CVD events.

Conclusion

A high-increasing serum urate trajectory during young adulthood was associated with incident CVD by middle age. Further studies are needed to evaluate whether the assessment of serum urate trajectories can be clinically useful to detect and make a decision to start to treat young adults at higher risk for CVD events.

Novelty and Significance.

What is New?

• The current study attempts to qualify the role of serum urate trajectory on CVD development in young adults. Three trajectory patterns of serum urate during young adulthood were identified. High-increasing serum urate trajectory during young adulthood increased CVD events risk in both sex.

What is Relevant?

• Concurrent hypertension during young adulthood may have a large effect on the association between serum urate trajectory and incident CVD. If further confirmed, the findings of the study will encourage clinicians to monitor hypertension while assessing CVD risk in individuals with high-increasing serum urate trajectory.