Thirty-Year Risk of Cardiovascular Disease Events in Adolescents with Severe Obesity
Abstract
Objective
Quantifying risk for cardiovascular disease (CVD) events among adolescents is difficult owing to the long latent period between risk factor development and disease outcomes. This study examined the 30-year CVD event risk among adolescents with severe obesity treated with and without metabolic and bariatric surgery (MBS), compared with youths with moderate obesity, overweight, or normal weight.
Methods
Cross-sectional and longitudinal comparisons of five frequency-matched (age and diabetes status) groups were performed: normal weight (n = 247), overweight (n = 54), obesity (n = 131), severe obesity without MBS (n = 302), and severe obesity undergoing MBS (n = 215). A 30-year CVD event score developed by the Framingham Heart Study was the primary outcome. Data are mean (SD) with differences between time points for MBS examined using linear mixed models.
Results
Preoperatively, the likelihood of CVD events was higher among adolescents undergoing MBS (7.9% [6.7%]) compared with adolescents with severe obesity not referred for MBS (5.5% [4.0%]), obesity (3.9% [3.0%]), overweight (3.1% [2.4%]), and normal weight (1.8% [0.8%]; all P < 0.001). At 1 year after MBS, event risk was significantly reduced (7.9% [6.7%] to 4.0% [3.4%], P < 0.0001) and was sustained for up to 5 years after MBS (P < 0.0001, all years vs. baseline).
Conclusions
Adolescents with severe obesity are at elevated risk for future CVD events. Following MBS, the predicted risk of CVD events was substantially and sustainably reduced.
Study Importance
What is already known?
- ► Adolescents with severe obesity often have elevated levels of many cardiovascular disease risk factors.
- ► Metabolic and bariatric surgery in this population produces significant weight reduction and improvements in cardiovascular disease risk factors.
What does this study add?
- ► This study demonstrates that adolescents with severe obesity have a significant risk of having a cardiovascular event prior to the age of 50 years.
- ► Metabolic and bariatric surgery for adolescents with severe obesity produces marked and sustained reduction in cardiovascular event risk.
Introduction
The long latent period between risk factor development and disease onset presents a challenge for evaluating cardiovascular disease (CVD) outcomes in pediatric cohorts. Evidence from several established longitudinal studies clearly demonstrates that obesity in childhood and adolescence is an independent risk factor for early CVD morbidity and mortality (1-7). However, this risk may be underestimated because the cohorts from which this evidence was derived were established prior to or early in the onset of the current obesity epidemic. Furthermore, the continued rise in obesity prevalence and the increased proportion of youths with severe obesity (i.e., BMI ≥ 120% of the 95th percentile or BMI ≥ 35 kg/m2) (8, 9) suggest that a current assessment of the long-term risks of CVD in youths with obesity is needed. Moreover, from a health care policy and economics standpoint, it would be very helpful to understand how effective obesity treatment impacts long-term CVD risk.
Measuring the actual impact of severe obesity and its treatment on CVD events is difficult until sufficient time has elapsed, usually requiring several decades for a critical mass of events to occur. However, the impact can be estimated using CVD prediction algorithms with relatively simple clinical data such as serum lipids, systolic blood pressure (SBP), diabetes status, and anthropometric measures. The Framingham Heart Study has developed a 30-year CVD event risk estimate (validated in 20- to 59-year-olds) that offers the opportunity to prospectively evaluate CVD event risk over the long term (10). Given the preponderance of evidence suggesting childhood obesity is a risk factor for early CVD morbidity and mortality (3, 4, 6, 7, 11), coupled with the elevated CVD risk profile exhibited among youths with severe obesity (12-18), it is probable that the risk of CVD events later in life is high among youths with severe obesity.
Currently, the only intervention with high rates of clinically meaningful and sustained weight loss among youths with severe obesity is metabolic and bariatric surgery (MBS) (12-14, 16-20). Data in adults undergoing MBS have demonstrated a significant reduction in CVD events and mortality compared with nonsurgical controls (21-23).Whether these findings translate to pediatric populations cannot be determined without long-term follow-up. However, the impact of MBS on long-term CVD event risk can be modeled using the Framingham methodology. Therefore, the aims of this analysis were to (1) examine the risk of developing a CVD event within 30 years among adolescents with severe obesity compared with contemporaries from differing weight strata, (2) determine the effect of MBS on 30-year CVD event risk among adolescents with severe obesity, and (3) evaluate the cost-effectiveness of MBS for CVD event reduction.
Methods
Surgical study cohort and measurement time points
Data from the Teen-Longitudinal Assessment of Bariatric Surgery (Teen-LABS) study were used for this analysis (24). Teen-LABS is an ongoing, NIDDK-funded, prospective, longitudinal, multicenter observational study that enrolled consecutive adolescents (< 20 years of age) undergoing MBS at five clinical centers. Parental permission and participant assent (< 18 years old) and consent from older adolescents (≥ 18 years old) were obtained. Medical and surgical care were provided for each patient as specified by patient care pathways at each institution. No attempts were made to standardize or alter care within this observational research protocol (25). Data collection time points used for this analysis were baseline (preoperative) and annually up to 5 years of postoperative assessments.
Teen-LABS risk factor measurements
At each assessment, height was measured on a wall-mounted stadiometer, weight was measured on an electronic scale (Scale-Tronix 5200, White Plains, New York, or Tanita TBF-310, Arlington Heights, Illinois), and BMI was calculated. Diabetes was defined by using self-reported diagnoses, medical record review, medication use for the treatment of diabetes (excluding metformin with a concomitant diagnosis of polycystic ovary syndrome), hemoglobin A1c ≥ 6.5%, impaired fasting glucose (≥ 126 mg/dL), or 2-hour oral glucose ≥ 200 mg/dL within 2 weeks before enrollment. Total cholesterol (TC) and high-density lipoprotein cholesterol (HDL-c) were measured from fasting blood samples at a central laboratory (Northwest Lipid Metabolism and Diabetes Research Laboratories, Seattle, Washington). Average SBP was taken from ≥ 2 separate measurements obtained using a Welch Allyn Spot Vital Signs monitor (4200B; Hill-Rom, Batesville, Indiana). Subject-reported use of antihypertension (anti-HTN) medications and smoking status were documented.
Nonsurgical comparators
Cincinnati Children's Hospital nonsurgical cohort
Data were obtained from a study of youths aged 10 to 23 years who had been enrolled in a study of obesity and type 2 diabetes mellitus on the heart and vasculature. The detailed methodology of this study has been previously published (26). Diagnosis of type 2 diabetes mellitus was made by the participants' primary care provider. This study was approved by the institutional review board at Cincinnati Children's Hospital Medical Center.
Following an overnight fast (minimum 10 hours), participants underwent anthropometric, SBP, and laboratory assessments. Height, weight, and waist circumference were measured in a standardized manner, as previously published (26). Two measures of height obtained with a calibrated stadiometer (Veeder-Root, Elizabethtown, North Carolina) and two measures of weight obtained with a Health-O-Meter electronic scale (Jarden Consumer Solutions, Rye, New York) were averaged. SBP was measured according to standards described in the fourth report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Children and Adolescents (27). The average of three SBP measurements was taken using a mercury manometer. Fasting plasma lipid profiles (TC and HDL-c) were performed with standardized methods from the National Heart, Lung, and Blood Institute–Centers for Disease Control and Prevention. Smoking exposure and medication use were evaluated by self-report questionnaire.
University of Minnesota nonsurgical cohort
The University of Minnesota cohort included data from a cross-sectional and longitudinal study examining CVD risk factors in youths who ranged from having normal weight to having severe obesity. Height and weight were determined using a wall-mounted stadiometer and an electronic scale, respectively. Seated SBP was measured with an automated blood pressure cuff after the participant had been resting quietly (10 minutes) for three consecutive times at approximately 3-minute intervals, and the average of these measurements was used. Fasting blood samples (> 10 hours) were used to measure TC and HDL-c with standard methods by the Fairview Diagnostics Laboratories, Fairview-University Medical Center (Minneapolis, Minnesota), a Centers for Disease Control and Prevention–certified laboratory. Smoking exposure and medication use were evaluated by self-report questionnaire. Patients with diabetes were not enrolled into this Minnesota cohort.
Main statistical analysis
Standard descriptive statistics were used to summarize participant characteristics at baseline. Categorical variables were calculated as frequencies and percentages; continuous variables were calculated as means and standard deviations. The analysis used two models for estimating full CVD event risks, including coronary death, myocardial infarction, stroke (fatal and nonfatal), angina pectoris, intermittent claudication, and congestive heart failure: (1) the first model included BMI but excluded lipids (TC and HDL-c), sex, age, SBP, anti-HTN treatment, smoking status, and diabetes status; (2) the second model excluded BMI but included lipids (TC and HDL-c), sex, age, SBP, anti-HTN treatment, smoking status, and diabetes status. Age, SBP, and TC and HDL-c levels were treated as continuous variables. Sex, anti-HTN treatment, smoking status, and diabetes status were treated as categorical variables. The CVD event models from the Framingham Heart Study were developed in, and modeled for, persons aged 20 to 59 years. Because of the model constraint and to eliminate the bias, ages less than 20 were adjusted to 20 for all groups in this study. Subjects with ages > 20 in comparison groups were removed to match the age range of the Teen-LABS baseline (13 to < 20 years old). Because of group differences in diabetes status and low prevalence, participants with normal weight and diabetes were removed from the study. Next, we matched the proportion of participants with diabetes who had overweight, obesity, and severe obesity to the proportion with diabetes in the Teen-LABS cohort. Missing data patterns and percentages were examined before calculating risk scores. We detected no evidence of informative missing data. Prior to matching, only complete data were used for calculating risk scores. Main analyses were performed using SAS software version 9.4. (SAS Institute Inc., Cary, North Carolina). Data were presented as means with 95% confidence intervals. The statistical significance level was set at α = 0.05. The Tukey adjustment method for multiple testing was used for pairwise post hoc comparisons between groups with normal weight, overweight, obesity, or severe obesity and the Teen-LABS baseline cohort for each hypothesis considered in this study. A linear mixed model was used for risk score comparisons between baseline and follow-up time points.
Effectiveness and cost-effectiveness modeling
A state-transition model was developed to assess the long-term and cost-effectiveness of MBS for CVD event risk reduction. The analyses used Teen-LABS baseline and postsurgical results compared with modeled nonsurgical comparators using Teen-LABS baseline characteristics. The analysis had a time horizon of 30 years, and the model cycle length was 1 year, with modeling beginning at 5 years following MBS. To minimize potential bias favoring our hypothesis, we created the statistical model under the conservative assumption that nonsurgical comparators would not gain weight over time (despite the existence of data to the contrary) (14, 19) but instead would remain at their initial BMI of 53 kg/m2. In the MBS group, we modeled receiving either Roux-en-Y gastric bypass (RYGB; 71%) or vertical sleeve gastrectomy (VSG; 29%) based on Teen-LABS distributions. We modeled BMI data showing a decline at 1 year with a slight increase each subsequent year and a plateau at 10 years based on the results of the Swedish Obese Subjects Study (21). Transitions from the event-free state to CVD event states were based on the 30-year composite risk score measured in the Teen-LABS study. Event states included coronary death, myocardial infarction, stroke (fatal and nonfatal), angina pectoris, intermittent claudication, and congestive heart failure. The distribution of CVD events for males and females was based on incidence rates from the National Heart, Lung, and Blood Institute (28). The probability of death from CVD events was estimated from the American Heart Association event statistics (29). The model also incorporated costs of MBS and CVD events. CVD event costs and utilities were estimated separately for acute and post-first-year events (11, 30-39). Both cohorts had a baseline quality-adjusted life year (QALY) of 0.72, and the MBS group gained 0.0056 QALYs per unit of BMI lost (28, 40). All costs from prior years were adjusted to 2015-year dollars using the Consumer Price Index. End points included QALYs, total costs (US $2015), and incremental cost-effectiveness ratios. A willingness-to-pay threshold of $100,000/QALY was used to determine cost-effectiveness. Costs and QALYs were each discounted at 3%. All cost-effectiveness analyses were conducted using the R statistical platform (R version 3.2.3; R Foundation for Statistical Computing, Vienna, Austria).
Results
Descriptive and clinical characteristics of the nonsurgical and MBS cohorts used for these analyses are presented in Table 1. Of note, participants were matched on the age range of Teen-LABS (13 to < 20 years old) and prevalence of diabetes among the differing obesity groups.
| Normal weight | Overweight | Obesity | Severe obesity | Teen-LABS preop | |
|---|---|---|---|---|---|
| n | 247 | 54 | 131 | 302 | 215 |
| Age (y) | 15.9 (2.5) | 16.8 (2.2) | 16.3 (2.3) | 16.4 (2.0) | 17.0 (1.5) |
| Sex (male) (%) | 41 | 37 | 44 | 38 | 25 |
| Race (%) | |||||
| White | 56 | 56 | 47 | 55 | 72 |
| Black | 42 | 37 | 51 | 43 | 22 |
| Other | 2 | 7 | 2 | 2 | 6 |
| Height (cm) | 164.7 (10.8) | 168.8 (8.9) | 167.2 (10.2) | 168.6 (9.0) | 168.2 (8.6) |
| Weight (kg) | 56.7 (11.5) | 76.8 (10.9) | 87.3 (13.5) | 117.8 (20.8) | 149.9 (30.5) |
| BMI (kg/m2) | 20.7 (2.4) | 26.9 (2.2) | 31.1 (2.5) | 41.3 (5.8) | 52.9 (9.4) |
| Percent of 95th BMI percentile | 73.5 (6.9) | 92.9 (5.1) | 109.3 (5.7) | 144.8 (21.1) | 168.0 (30.6) |
| SBP (mm Hg) | 107.5 (10.2) | 113.3 (8.9) | 116.8 (10.2) | 121.5 (11.8) | 125.4 (13.4) |
| TC (mg/dL) | 158 (29) | 168 (34) | 167 (32) | 167 (32) | 157 (30) |
| HDL-c (mg/dL) | 56 (12) | 53 (13) | 46 (10) | 44 (12) | 38 (9) |
| Smoking (%) | 1.6 | 1.9 | 0.0 | 2.7 | 1.9 |
| Diabetes (%) | 0 | 13.0 | 13.7 | 13.6 | 13.5 |
| Anti-HTN meds (%) | 0 | 5.6 | 3.8 | 5.3 | 22.8 |
- Data presented as means (SD) and percentages.
- Anti-HTN meds, prescribed antihypertension medications; HDL-c, high-density lipoprotein cholesterol; preop, before surgery; SBP, systolic blood pressure; Teen-LABS, Teen-Longitudinal Assessment of Bariatric Surgery; TC, total cholesterol.
Baseline (before surgery [preop]) and postsurgical follow-up data are presented in Table 2 with data by surgical type presented as Supporting Information (RYGB: Supporting Information Table S1; VSG: Supporting Information Table S2). The sample size at each time point represents those with complete data. The nadir for postsurgical weight loss (BMI reduction and weight loss) was achieved at 1 year with small increases in 3-year, 4-year, and 5-year weight loss. SBP and TC decreased at 1 year and increased in each subsequent year. HDL-c increased at 1 year with a further increase at 2 years, which was sustained out to 5 years. Diabetes prevalence decreased after MBS (13.5% to 2.2% prevalence from baseline to 1 year), and the decrease was sustained to 5 years after surgery. A reduction in blood pressure medication use was also observed. Smoking prevalence increased from baseline.
| Teen-LABS baseline and postsurgical follow-up time points | ||||||
|---|---|---|---|---|---|---|
| Preop | 1 year | 2 years | 3 years | 4 years | 5 years | |
| n | 215 | 181 | 168 | 162 | 166 | 158 |
| Age (y) | 17.0 (1.5) | 18.1 (1.5) | 19.3 (1.5) | 20.2 (1.5) | 21.1 (1.5) | 22.1 (1.6) |
| Sex (male) (%) | 25 | 26 | 23 | 24 | 24 | 25 |
| Race (%) | ||||||
| White | 72 | 72 | 69 | 70 | 71 | 72 |
| Black | 22 | 23 | 25 | 24 | 23 | 22 |
| Other | 6 | 5 | 6 | 6 | 6 | 6 |
| Height (cm) | 168.2 (8.6) | 168.2 (9.1) | 168.8 (8.9) | 168.1 (9.5) | 168.8 (9.1) | 168.3 (10.0) |
| Weight (kg) | 149.9 (30.5) | 104.8 (27.0) | 105.5 (28.7) | 108.5 (31.7) | 111.5 (33.9) | 113.9 (32.0) |
| BMI (kg/m2) | 52.9 (9.4) | 37.0 (8.9) | 37.0 (9.8) | 38.4 (10.9) | 39.1 (11.4) | 40.3 (10.9) |
| SBP (mmHg) | 125 (13) | 117 (13) | 119 (13) | 118 (14) | 120 (14) | 122 (14) |
| TC (mg/dL) | 157 (30) | 147 (30) | 150 (29) | 151 (29) | 156 (35) | 154 (29) |
| HDL-c (mg/dL) | 38 (9) | 48 (11) | 53 (12) | 53 (14) | 55 (15) | 54 (15) |
| Smoking (%) | 1.9 | 3.3 | 6.6 | 8.6 | 6.6 | 10.8 |
| Diabetes (%) | 13.5 | 2.2 | 1.8 | 0.6 | 2.4 | 2.5 |
| Anti-HTN meds (%) | 22.8 | 8.3 | 4.8 | 5.6 | 6.0 | 5.1 |
- Data presented as means (SD) and percentages.
- Anti-HTN meds, prescribed antihypertension medications; HDL-c, high-density lipoprotein cholesterol; preop, before surgery; SBP, systolic blood pressure; Teen-LABS, Teen-Longitudinal Assessment of Bariatric Surgery; TC, total cholesterol.
Figure 1 displays individual 30-year risk of CVD events (with BMI included in the model; Supporting Information Figure S1 shows risk without BMI in the model) for each comparison group and Teen-LABS preop versus BMI. Individual variation was observed in each group and was most pronounced among adolescents with obesity and severe obesity (Table 3). For the Teen-LABS preop group, 52.6% of the cohort presented with > 5% risk of CVD events, 22.3% presented with > 10% risk, 10.7% presented with > 15% risk, and 5.6% presented with > 20% risk. For the group with severe obesity, 38.7% of the cohort presented with > 5% risk of CVD events, 8.9% presented with > 10% risk, and 3.6% presented with > 15% risk. For the group with obesity, 13.0% of the cohort presented with > 5% risk of CVD events, 4.6% presented with > 10% risk, and 2.3% presented with > 15% risk. For the group with overweight, 9.3% with overweight presented with > 5% risk of CVD events and 1.9% had > 10% risk of CVD events. Conversely, only 2 out of 247 (0.8%) youths with normal weight presented with > 5% risk of CVD events, and no youths with normal weight had > 10% risk of CVD events.
| Risk level | Normal weight (n = 247) |
Overweight (n = 54) |
Obesity (n = 131) | Severe obesity (n = 302) | Teen-LABS preop (n = 215) | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| n | % | n | % | n | % | n | % | n | % | |
| > 5% | 2 | 0.8 | 5 | 9.3 | 17 | 13 | 117 | 38.7 | 113 | 52.6 |
| > 10% | 0 | 0.0 | 1 | 1.9 | 6 | 4.6 | 27 | 8.9 | 48 | 22.3 |
| > 15% | 0 | 0.0 | 1 | 1.9 | 3 | 2.3 | 11 | 3.6 | 23 | 10.7 |
| > 20% | 0 | 0.0 | 0 | 0.0 | 1 | 0.8 | 6 | 2.0 | 12 | 5.6 |
- Preop, before surgery; Teen-LABS, Teen-Longitudinal Assessment of Bariatric Surgery.
Figure 2 depicts the 30-year risk of CVD events (with BMI included in the model; Supporting Information Figure S2 shows risk without BMI in the model) of each nonsurgical cohort group and Teen-LABS at each time point. Post hoc pairwise comparisons were conducted (using the Tukey method) between groups. No statistical differences in 30-year CVD event risk were observed between youths with normal weight (1.8% [0.8%]) and youths with overweight (3.1% [2.4%]; P = 0.24) or between youths with overweight and youths with obesity (3.9% [3.0%]; P = 0.76). The Teen-LABS group at baseline (preop) had a significantly higher 30-year CVD event risk (7.9% [6.7%]) as compared with all comparison groups (all P < 0.0001). Youths with severe obesity had a higher 30-year CVD event risk (5.5% [4.0%]) than youths with normal weight, overweight, and obesity (all P < 0.01). Following MBS in Teen-LABS participants, a statistically significant reduction in 30-year CVD event risk was observed at 1 year, which was sustained at each time point up to 5 years after surgery (P < 0.0001 all). The proportion of participants with elevated CVD risk scores fell (> 5% [52.6% to 22.1%], > 10% [22.3% to 4.4%], and > 15% [10.7% to 1.7%]) at 1 year (Supporting Information Table S3). These reductions were sustained with no statistically significant changes by years 4 or 5.
Figure 3 compares the 30-year risk of CVD events for two models (model 1 with BMI excluding HDL and TC; model 2 without BMI but including TC and HDL) for the Teen-LABS cohort at each time point. Although CVD events were reduced in both models following MBS and this was sustained for 3 years, the estimates of CVD event risk were lower without the inclusion of BMI in the model. Analyses by surgical type were conducted as well, showing a similar pattern (RYGB: Supporting Information Figure S3; VSG: Supporting Information Figure S4). It should be noted that these analyses are underpowered because of sample size limitations.
Table 4 shows effectiveness models of MBS versus nonsurgical controls for CVD events and deaths over 30 years. The number of qualifying youths with severe obesity needing to be treated with MBS to prevent one CVD event was 36.5, and the number needing to be treated to prevent one CVD death was 379. These analyses were expanded (Supporting Information Table S4) to examine the cost-effectiveness of MBS for preventing CVD events. Despite the initial cost of MBS ($24,675.59), the MBS models indicate higher QALYs (18.5 vs. 16.8) than control, with an incremental cost-effectiveness ratio of $13,432.64/QALY. Using a willingness-to-pay threshold of $100,000/QALY, MBS was cost-effective, as it was well below the cutoff. Detailed CVD event probability and costs are included in Supporting Information Table S5.
| Outcome | Group | Estimated event/death rate (%) | Number needed to treat |
|---|---|---|---|
| Total CVD events | Control | 7.31 | 36.5 |
| Bariatric surgery | 4.58 | ||
| CVD deaths | Control | 0.70 | 379 |
| Bariatric surgery | 0.44 |
Discussion
The main results of these analyses indicate that adolescents with severe obesity, especially those qualifying for MBS, appear to be at pronounced risk for a major CVD event within 30 years and that by undergoing MBS as an adolescent, this risk of hard cardiovascular end points was substantially mitigated, despite a postoperative plateau of BMI that did not drop below the cutoff for severe obesity for most participants. From a health care cost perspective, despite the initial cost of surgery, these data indicate that MBS in adolescents with severe obesity is cost-effective by preventing later CVD events and early CVD deaths.
Prior to undergoing MBS, adolescents with severe obesity face a greater than fourfold higher risk of CVD events within 30 years compared with peers with normal weight. Even when compared with age- and diabetes frequency–matched adolescents with severe obesity who did not undergo MBS, a statistically significant and clinically meaningfully elevated risk of 30-year CVD events was found prior to surgery. This finding emphasizes the truly unique nature of adolescents who qualify for and elect to undergo MBS, as they appear to harbor a very concerning CVD risk profile. However, despite having lower estimated CVD event risk than the Teen-LABS cohort, youths with severe obesity who did not undergo MBS (the comparison group) were also at pronounced risk for CVD events when compared with peers with lesser obesity or normal weight. These results support the position that more effective treatment options are needed in this group to mitigate this risk (20).
The findings from this study, despite differences in methodology, are similar to data in adults undergoing MBS, demonstrating significant reductions in measured CVD events and mortality (21-23). Our study is limited by using estimated risk models; however, the models have been validated in adults using hard endpoints (e.g., stroke, heart failure). In addition, our use of contemporaneous comparison groups demonstrates striking treatment effects of surgery, irrespective of the residual high BMI values in many of the postoperative cohort members. Despite adult studies being able to measure actual events, the similarity in terms of relative treatment effects are striking. The Swedish Obese Subjects Study demonstrated after a median follow-up of 14.7 years (range: 0-20 years) that MBS reduced the risk for total CVD events by 33%, whereas our analysis of these adolescent data estimated a 39% risk reduction at 30 years. Participants in the Swedish Obese Subjects Study showed that the number of persons needed to be treated to prevent 1 event was 50 (21), whereas our modeling with adolescent data estimated needing only 36.5 persons. In addition, another cohort composed entirely of adults undergoing RYGB has demonstrated similar findings, with a reduction in composite CVD outcomes (hazard ratio: 0.58) and a reduced 10-year Framingham Risk Score at 5 years when compared with nonsurgical controls (23). Collectively, our data using risk models are consistent with these findings using hard endpoints, despite the limitations inherent in a modeling approach to prediction of long-term CVD outcomes.
Importantly, the reduced risk was a stable finding in spite of the majority of surgical participants remaining in the severe obesity category (50.8% [92/181]) following MBS. The reduced weight, combined with improvements in dyslipidemia, reductions in blood pressure, and lower diabetes prevalence all contributed to the risk reduction. Moreover, the sustained risk reduction was observed despite a significant increase in smoking prevalence from preop (1.9%) to 5 years (10.8%). These findings further highlight the multifaceted metabolic effects of MBS, which are more profound than weight loss alone. Given these findings, it will be important to continue to follow adolescents after MBS over the long term in order to better understand the implications, particularly considering the relatively high degree of heterogeneity that was projected. Future work could focus on better understanding the sources of variability and whether patient selection in terms of number and severity of comorbidities may play a role. Furthermore, these data highlight the need for creation of CVD risk factor scoring lasting longer than 30 years and encompassing adolescent age groups.
This study has several strengths, including a large sample size of the surgical and comparison cohorts, longitudinal follow-up for 5 years after surgery with excellent cohort retention, and the use of clinically relevant outcomes and measures. However, there are several limitations inherent in this work as well. The algorithm used to calculate risk reduction makes certain assumptions that deserve mention. The 30-year risk scores were originally developed for persons 20 to < 50 years old; therefore, our models used an age of 20 for calculations despite primary data from individuals with mean ages of ~17 years. Although this assumption represents a constraint and a study limitation, there is little reason to believe that there are substantive differences in CVD event risk between the mean age of 17 and the 20 years of age used to conduct analysis; therefore, this necessary assumption likely does not alter the risk assessment. The racial distribution of the Framingham cohort is largely white, and the diversity difference documented in our cohort may lead to bias in calculation of risk. Although the risk scores were developed in the Framingham Heart Study, whose benefits and limitations have been well documented (41, 42), the predicted risk of events does not always equate to actual event occurrence and may be an overestimation as presented in this study. We conducted analysis examining differences by surgical procedure; however, because of sample size, these data are underpowered and should be viewed with caution. It will be important for future studies to examine these important differences.
Conclusion
Adolescents with severe obesity are at a much higher risk of experiencing a major CVD event within 30 years in comparison with peers with normal weight, overweight, or obesity. Following MBS, risk of CVD events was substantially and durably reduced. The estimated number of adolescents with severe obesity who needed to be treated in order to prevent 1 event was 36.5. MBS appears to be cost-effective, due in part to CVD event prevention.
Acknowledgments
We would like to thank all of the adolescents and young adults who participated in these studies.
Funding agencies
Funding for Teen-LABS was provided by the NIH (U01DK072493/UM1 DK072493 to THI) and the National Center for Research Resources and the National Center for Advancing Translational Sciences, NIH (8UL1TR000077). Support also came from the National Center for Research Resources and the National Center for Advancing Translational Sciences, NIH (UL1TR000114). Nonsurgical comparison cohorts were provided in part by funding from NIH grants F32-HL127881 (JRR), R01-HL110957 (ASK), and R01-HL105591 (EMU). Research reported in this publication was also supported by the National Center for Advancing Translational Sciences of the NIH award number UL1TR000114 (JRR).
Disclosure
Dr. Ryder receives support from Boehringer Ingelheim Pharmaceuticals in the form of drugs/placebos. Dr. Inge has served as a consultant for Standard Bariatrics and Independent Medical Expert Consulting Services, both unrelated to this project. Dr. Kelly serves as a consultant for Vivus Pharmaceuticals, Novo Nordisk Pharmaceuticals, and WW but does not accept personal or professional income for his services. Dr. Kelly also receives research support from AstraZeneca Pharmaceuticals in the form of drugs/placebos. Dr Urbina receives compensation from Astellas for participation on a data safety monitoring board for a drug study unrelated to this study. The other authors declared no conflict of interest.
Author contributions
JRR conceptualized the framework of the manuscript, drafted the initial manuscript, and reviewed and revised the manuscript. EMU, ASK, and THI obtained funding for cohorts used, supervised data collection, and reviewed and revised the manuscript. CH, PX, TMJ, CX, ML, and JC carried out statistical analysis and reviewed and revised the manuscript. MPM supervised data collection and reviewed and revised the manuscript. The content is solely the responsibility of the authors, the final manuscript was approved by all authors as submitted, and all authors agree to be accountable for all aspects of the work.
Clinical trial registration
ClinicalTrials.gov identifier NCT00474318.
