[Skip to Content]
[Skip to Content Landing]
Figure.
Cumulative Incidence Estimates of the Proportion of Participants Experiencing a Major Adverse Cardiovascular Event
Cumulative Incidence Estimates of the Proportion of Participants Experiencing a Major Adverse Cardiovascular Event

The solid line denotes the cumulative incidence function estimate accounting for competing risk. Shading around the line denotes the 95% CI. MACE indicates major adverse cardiovascular event.

Table 1.  
Baseline Characteristics of the Study Population
Baseline Characteristics of the Study Population
Table 2.  
Association Between Categorical Measures of Body Composition and Major Adverse Cardiovascular Events
Association Between Categorical Measures of Body Composition and Major Adverse Cardiovascular Events
Supplement.

eFigure 1. Inter-rater agreement of computed tomography derived measures of body composition was assessed in a random sample of 50 patients by two investigators who were blinded to outcome.

eFigure 2. Cumulative incidence functions of major adverse cardiovascular events, stratified by category of visceral adipose tissue (top panel) and muscle radiodensity (bottom panel).

eTable 1. Relationship between continuous measures of body composition and major adverse cardiovascular events.

eFigure 3. Multivariable-adjusted restricted cubic splines depicting the association between body mass index (top panel), visceral adipose tissue (middle panel), and muscle radiodensity (bottom panel) with risk of experiencing a major adverse cardiovascular event.

eTable 2. Relationship between categorical measures of body composition and major adverse cardiovascular events, additionally adjusted for body mass index.

eTable 3. Relationship between categorical measures of body composition and major adverse cardiovascular events, excluding 95 patients who experienced major adverse cardiovascular events within the first 12-months of follow-up.

eTable 4. Relationship between categorical measures of body composition and major adverse cardiovascular events, excluding 499 patients who had computed tomography imaging obtained after surgical resection.

eTable 5. Relationship between categorical measures of body composition and major adverse cardiovascular events, excluding 467 patients who received pre-operative chemotherapy and/or radiation therapy.

eTable 6. Relationship between categorical measures of body composition and major adverse cardiovascular events, using sex-specific categories.

eFigure4. E-value sensitivity analyses of the observed association of visceral adipose tissue area (top panel) and muscle radiodensity (bottom panel) with risk of experiencing a major adverse cardiovascular event.

1.
Siegel  RL, Miller  KD, Fedewa  SA,  et al.  Colorectal cancer statistics, 2017.  CA Cancer J Clin. 2017;67(3):177-193. doi:10.3322/caac.21395PubMedGoogle ScholarCrossref
2.
Jemal  A, Ward  EM, Johnson  CJ,  et al.  Annual report to the nation on the status of cancer, 1975-2014, featuring survival.  J Natl Cancer Inst. 2017;109(9). doi:10.1093/jnci/djx030PubMedGoogle Scholar
3.
Baade  PD, Fritschi  L, Eakin  EG.  Non-cancer mortality among people diagnosed with cancer (Australia).  Cancer Causes Control. 2006;17(3):287-297. doi:10.1007/s10552-005-0530-0PubMedGoogle ScholarCrossref
4.
van Erning  FN, van Steenbergen  LN, Lemmens  VEPP,  et al.  Conditional survival for long-term colorectal cancer survivors in the Netherlands: who do best?  Eur J Cancer. 2014;50(10):1731-1739. doi:10.1016/j.ejca.2014.04.009PubMedGoogle ScholarCrossref
5.
Kenzik  KM, Balentine  C, Richman  J, Kilgore  M, Bhatia  S, Williams  GR.  New-onset cardiovascular morbidity in older adults with stage I to III colorectal cancer.  J Clin Oncol. 2018;36(6):609-616. doi:10.1200/JCO.2017.74.9739PubMedGoogle ScholarCrossref
6.
El-Shami  K, Oeffinger  KC, Erb  NL,  et al.  American Cancer Society colorectal cancer survivorship care guidelines.  CA Cancer J Clin. 2015;65(6):428-455. doi:10.3322/caac.21286PubMedGoogle ScholarCrossref
7.
Kim  SH, Després  JP, Koh  KK.  Obesity and cardiovascular disease: friend or foe?  Eur Heart J. 2016;37(48):3560-3568. doi:10.1093/eurheartj/ehv509PubMedGoogle ScholarCrossref
8.
Lee  CMY, Huxley  RR, Wildman  RP, Woodward  M.  Indices of abdominal obesity are better discriminators of cardiovascular risk factors than BMI: a meta-analysis.  J Clin Epidemiol. 2008;61(7):646-653. doi:10.1016/j.jclinepi.2007.08.012PubMedGoogle ScholarCrossref
9.
Romero-Corral  A, Somers  VK, Sierra-Johnson  J,  et al.  Accuracy of body mass index in diagnosing obesity in the adult general population.  Int J Obes (Lond). 2008;32(6):959-966. doi:10.1038/ijo.2008.11PubMedGoogle ScholarCrossref
10.
Prado  CM, Heymsfield  SB.  Lean tissue imaging: a new era for nutritional assessment and intervention.  JPEN J Parenter Enteral Nutr. 2014;38(8):940-953. doi:10.1177/0148607114550189PubMedGoogle ScholarCrossref
11.
Brown  JC, Cespedes Feliciano  EM, Caan  BJ.  The evolution of body composition in oncology-epidemiology, clinical trials, and the future of patient care: facts and numbers.  J Cachexia Sarcopenia Muscle. 2018;9(7):1200-1208. doi:10.1002/jcsm.12379PubMedGoogle ScholarCrossref
12.
World Health Organization.  Obesity: Preventing and Managing the Global Epidemic. Geneva, Switzerland: World Health Organization; 2000.
13.
Shen  W, Punyanitya  M, Wang  Z,  et al.  Total body skeletal muscle and adipose tissue volumes: estimation from a single abdominal cross-sectional image.  J Appl Physiol (1985). 2004;97(6):2333-2338. doi:10.1152/japplphysiol.00744.2004PubMedGoogle ScholarCrossref
14.
Aubrey  J, Esfandiari  N, Baracos  VE,  et al.  Measurement of skeletal muscle radiation attenuation and basis of its biological variation.  Acta Physiol (Oxf). 2014;210(3):489-497. doi:10.1111/apha.12224PubMedGoogle ScholarCrossref
15.
Charlson  M, Szatrowski  TP, Peterson  J, Gold  J.  Validation of a combined comorbidity index.  J Clin Epidemiol. 1994;47(11):1245-1251. doi:10.1016/0895-4356(94)90129-5PubMedGoogle ScholarCrossref
16.
US Department of Health and Human Services, US Food and Drug Administration, Center for Drug Evaluation and Research. Guidance for industry: diabetes mellitus—evaluating cardiovascular risk in new antidiabetic therapies to treat type 2 diabetes. https://www.fda.gov/downloads/Drugs/Guidances/ucm071627.pdf. Accessed April 2, 2019.
17.
McCormick  N, Bhole  V, Lacaille  D, Avina-Zubieta  JA.  Validity of diagnostic codes for acute stroke in administrative databases: a systematic review.  PLoS One. 2015;10(8):e0135834. doi:10.1371/journal.pone.0135834PubMedGoogle ScholarCrossref
18.
McCormick  N, Lacaille  D, Bhole  V, Avina-Zubieta  JA.  Validity of myocardial infarction diagnoses in administrative databases: a systematic review.  PLoS One. 2014;9(3):e92286. doi:10.1371/journal.pone.0092286PubMedGoogle ScholarCrossref
19.
Austin  PC, Fine  JP.  Practical recommendations for reporting Fine-Gray model analyses for competing risk data.  Stat Med. 2017;36(27):4391-4400. doi:10.1002/sim.7501PubMedGoogle ScholarCrossref
20.
Fine  JP, Gray  RJ.  A proportional hazards model for the subdistribution of a competing risk.  J Am Stat Assoc. 1999;94(446):496-509. doi:10.1080/01621459.1999.10474144Google ScholarCrossref
21.
Caan  BJ, Meyerhardt  JA, Kroenke  CH,  et al.  Explaining the obesity paradox: the association between body composition and colorectal cancer survival.  Cancer Epidemiol Biomarkers Prev. 2017;26(7):1008-1015. doi:10.1158/1055-9965.EPI-17-0200PubMedGoogle ScholarCrossref
22.
Kroenke  CH, Prado  CM, Meyerhardt  JA,  et al.  Muscle radiodensity and mortality in patients with colorectal cancer.  Cancer. 2018;124(14):3008-3015. doi:10.1002/cncr.31405PubMedGoogle ScholarCrossref
23.
Austin  PC, Lee  DS, Fine  JP.  Introduction to the analysis of survival data in the presence of competing risks.  Circulation. 2016;133(6):601-609. doi:10.1161/CIRCULATIONAHA.115.017719PubMedGoogle ScholarCrossref
24.
Dignam  JJ, Zhang  Q, Kocherginsky  M.  The use and interpretation of competing risks regression models.  Clin Cancer Res. 2012;18(8):2301-2308. doi:10.1158/1078-0432.CCR-11-2097PubMedGoogle ScholarCrossref
25.
Kronmal  RA.  Spurious correlation and the fallacy of the ratio standard revisited.  J R Stat Soc Ser A Stat Soc. 1993;156(3):379-392. doi:10.2307/2983064Google ScholarCrossref
26.
Koene  RJ, Prizment  AE, Blaes  A, Konety  SH.  Shared risk factors in cardiovascular disease and cancer.  Circulation. 2016;133(11):1104-1114. doi:10.1161/CIRCULATIONAHA.115.020406PubMedGoogle ScholarCrossref
27.
Shi  Q, Andre  T, Grothey  A,  et al.  Comparison of outcomes after fluorouracil-based adjuvant therapy for stages II and III colon cancer between 1978 to 1995 and 1996 to 2007: evidence of stage migration from the ACCENT database.  J Clin Oncol. 2013;31(29):3656-3663. doi:10.1200/JCO.2013.49.4344PubMedGoogle ScholarCrossref
28.
van Leersum  NJ, Janssen-Heijnen  ML, Wouters  MW,  et al.  Increasing prevalence of comorbidity in patients with colorectal cancer in the South of the Netherlands 1995-2010.  Int J Cancer. 2013;132(9):2157-2163. doi:10.1002/ijc.27871PubMedGoogle ScholarCrossref
29.
Hawkes  AL, Lynch  BM, Owen  N, Aitken  JF.  Lifestyle factors associated concurrently and prospectively with co-morbid cardiovascular disease in a population-based cohort of colorectal cancer survivors.  Eur J Cancer. 2011;47(2):267-276. doi:10.1016/j.ejca.2010.10.002PubMedGoogle ScholarCrossref
30.
Pouliot  MC, Després  JP, Lemieux  S,  et al.  Waist circumference and abdominal sagittal diameter: best simple anthropometric indexes of abdominal visceral adipose tissue accumulation and related cardiovascular risk in men and women.  Am J Cardiol. 1994;73(7):460-468. doi:10.1016/0002-9149(94)90676-9PubMedGoogle ScholarCrossref
31.
de Koning  L, Merchant  AT, Pogue  J, Anand  SS.  Waist circumference and waist-to-hip ratio as predictors of cardiovascular events: meta-regression analysis of prospective studies.  Eur Heart J. 2007;28(7):850-856. doi:10.1093/eurheartj/ehm026PubMedGoogle ScholarCrossref
32.
Goodpaster  BH, Krishnaswami  S, Resnick  H,  et al.  Association between regional adipose tissue distribution and both type 2 diabetes and impaired glucose tolerance in elderly men and women.  Diabetes Care. 2003;26(2):372-379. doi:10.2337/diacare.26.2.372PubMedGoogle ScholarCrossref
33.
Gustafson  B, Smith  U.  Cytokines promote Wnt signaling and inflammation and impair the normal differentiation and lipid accumulation in 3T3-L1 preadipocytes.  J Biol Chem. 2006;281(14):9507-9516. doi:10.1074/jbc.M512077200PubMedGoogle ScholarCrossref
34.
Iliodromiti  S, Celis-Morales  CA, Lyall  DM,  et al.  The impact of confounding on the associations of different adiposity measures with the incidence of cardiovascular disease: a cohort study of 296 535 adults of white European descent.  Eur Heart J. 2018;39(17):1514-1520. doi:10.1093/eurheartj/ehy057PubMedGoogle ScholarCrossref
35.
Coleman  KJ, Ngor  E, Reynolds  K,  et al.  Initial validation of an exercise “vital sign” in electronic medical records.  Med Sci Sports Exerc. 2012;44(11):2071-2076. doi:10.1249/MSS.0b013e3182630ec1PubMedGoogle ScholarCrossref
36.
Birman-Deych  E, Waterman  AD, Yan  Y, Nilasena  DS, Radford  MJ, Gage  BF.  Accuracy of ICD-9-CM codes for identifying cardiovascular and stroke risk factors.  Med Care. 2005;43(5):480-485. doi:10.1097/01.mlr.0000160417.39497.a9PubMedGoogle ScholarCrossref
Views 1,604
Citations 0
Original Investigation
May 16, 2019

Body Composition and Cardiovascular Events in Patients With Colorectal Cancer: A Population-Based Retrospective Cohort Study

Author Affiliations
  • 1Pennington Biomedical Research Center, Baton Rouge, Louisiana
  • 2Stanley S. Scott Cancer Center, Louisiana State University Health Sciences Center, New Orleans
  • 3Kaiser Permanente Northern California, Oakland
  • 4University of Alberta, Edmonton, Alberta, Canada
  • 5Dana-Farber Cancer Institute, Boston, Massachusetts
JAMA Oncol. 2019;5(7):967-972. doi:10.1001/jamaoncol.2019.0695
Key Points

Question  Which measures of body composition are associated with major adverse cardiovascular events in patients with colorectal cancer (CRC)?

Findings  In this population-based cohort study of 2839 patients with CRC, body composition measured by visceral adiposity and muscle radiodensity was associated with major adverse cardiovascular events, whereas body composition measured by body mass index was not associated with these events.

Meaning  Body composition measures collected using routine computed tomographic images, including visceral adiposity and muscle radiodensity, can be used to assess cardiac risk in patients with CRC; however, body mass index may have limited use for assessing cardiovascular risk in this patient population.

Abstract

Importance  Patients with colorectal cancer (CRC) are up to 4-fold more likely than individuals without a history of cancer to develop cardiovascular disease. Clinical care guidelines recommend that physicians counsel patients with CRC regarding the association between obesity (defined using body mass index [BMI] calculated as weight in kilograms divided by height in meters squared) and cardiovascular disease risk; however, this recommendation is based on expert opinion.

Objective  To determine which measures of body composition are associated with major adverse cardiovascular events (MACEs) in patients with CRC.

Design, Setting, and Participants  Population-based retrospective cohort study of 2839 patients with stage I to III CRC diagnosed between January 2006 and December 2011 at an integrated health care system in North America.

Exposures  The primary exposures were BMI and computed tomography–derived body composition measurements (eg, adipose tissue compartments and muscle characteristics) obtained at the diagnosis of CRC.

Main Outcomes and Measures  The primary outcome was time to the first occurrence of MACE after diagnosis of CRC, including myocardial infarction, stroke, and cardiovascular death.

Results  In this population-based cohort study of 2839 participants with CRC (1384 men and 1455 women), the average age (SD) was 61.9 (11.5) years (range, 19-80 years). A substantial number of patients were former (1127; 40%) or current smokers (340; 12%), with hypertension (1150; 55%), hyperlipidemia (1389; 49%), and type 2 diabetes (573; 20%). The cumulative incidence of MACE 10 years after diagnosis of CRC was 19.1%. Body mass index was positively correlated with some computed tomography-derived measures of body composition. However, BMI was not associated with MACE; contrasting BMI categories of greater than or equal to 35 vs 18.5 to 24.9, the hazard ratio (HR) was 1.23 (95% CI, 0.85-1.77; P = .50 for trend). Visceral adipose tissue area was associated with MACE; contrasting the highest vs lowest quintile, the HR was 1.54 (95% CI, 1.02-2.31; P = .04 for trend). Subcutaneous adipose tissue area was not associated with MACE; contrasting the highest vs lowest quintile, the HR was 1.15 (95% CI, 0.78-1.69; P = .65 for trend). Muscle mass was not associated with MACE; contrasting the highest vs lowest quintile, the HR was 0.96 (95% CI, 0.57-1.61; P = .92 for trend). Muscle radiodensity was associated with MACE; contrasting the highest (ie, less lipid stored in the muscle) vs lowest quintile, the HR was 0.67 (95% CI, 0.44-1.03; P = .02 for trend).

Conclusions and Relevance  Visceral adiposity and muscle radiodensity appear to be risk factors for MACE. Body mass index may have limited use for determining cardiovascular risk in this patient population.

Introduction

Colorectal cancer (CRC) is the fourth most common malignant neoplasm in the United States.1 Five-year survival for patients with CRC has increased by 33% over the past 4 decades.2 Patients with CRC are now more susceptible to competing causes of morbidity and mortality, such as those from cardiovascular disease (CVD).3,4 Patients with CRC are 2-fold to 4-fold more likely than individuals without a history of cancer to develop CVD.5 Given the high risk of CVD in patients with CRC, evidence is necessary to inform cardiovascular management in this susceptible population.

The American Cancer Society’s CRC survivorship care guidelines recommend that physicians counsel patients with CRC regarding the association between obesity (defined using body mass index [BMI], which is calculated as weight in kilograms divided by height in meters squared) and CVD risk, a recommendation based exclusively on expert opinion.6 Moreover, uncertainty exists regarding the use of BMI for optimal cardiovascular risk management.7-9 At CRC diagnosis, patients undergo radiologic imaging with computed tomography (CT) to characterize the disease stage. Using commercially available automated analysis methods, CT images can be used to quantify body composition, including visceral and subcutaneous adiposity and muscle mass and radiodensity (a measure of lipid deposition into skeletal muscle).10 Quantification of body composition may improve prognostication of overall and cancer-specific survival11; however, the incremental utility to guide CVD risk management is unknown.

This study aimed to achieve 3 objectives using a population-based retrospective cohort of 2839 patients with CRC who were treated with curative intent. The first objective was to quantify the incidence of cardiovascular events up to 10 years after diagnosis of CRC. The second objective was to quantify the correlation between BMI and other measures of body composition. The third objective was to determine if specific BMI categories and CT-derived measures of body composition are risk factors for cardiovascular events independent of traditional risk factors, including smoking, hypertension, hyperlipidemia, and type 2 diabetes.

Methods
Study Population and Design

The Colorectal, Sarcopenia, Cancer And Near-term Survival (C-SCANS) cohort was derived from the Kaiser Permanente Northern California (KPNC) cancer registry, with ascertainment of all patients aged 18 to 80 years who were diagnosed with stage I to III invasive CRC from 2006 to 2011 and underwent surgical resection for CRC (n = 4465). We excluded 693 patients without abdominal or pelvic CT images, 411 patients without valid measures of body mass, 99 patients whose CT images were unreadable owing to poor image quality, and 423 patients who had a history of myocardial infarction or stroke documented in the electronic medical record (EMR) prior to CRC diagnosis. The final analytic sample included 2839 patients. A waiver of written patient informed consent was obtained by the study investigators, and this study was approved by the KPNC and University of Alberta institutional review boards. Data analyses were performed from March 2018 to September 2018.

Measures of Body Composition

Height in meters and weight in kilograms were measured by medical assistants at the time of diagnosis. Body mass index was calculated as weight in kilograms divided by height in meters squared and categorized using the World Health Organization classifications.12 Body composition was measured using a single-slice transverse CT image of the third lumbar vertebra and analyzed with sliceOmatic software V5.0 (TomoVision).13 Tissues were demarcated with a semiautomated procedure using Hounsfield unit thresholds of −29 to 150 for muscle, −150 to −50 for visceral adipose tissue, and −190 to −30 for subcutaneous adipose tissue. Muscle radiodensity quantifies the average radiation attenuation rate (in Hounsfield units) and is a radiologic measure of lipid deposition into skeletal muscle.14 A randomly selected subsample of 50 CT images were analyzed by 2 research staff members blinded to outcome (eFigure 1 in the Supplement), and the remaining CT images were analyzed by a single trained research staff member blinded to outcome.

Covariates

The KPNC EMR was used to obtain baseline information on age, sex, self-reported race and ethnicity, and CVD risk factors, including smoking history, hypertension, hyperlipidemia, and type 2 diabetes.15 Cardiovascular disease risk factors were obtained using a 36-month lookback period from the time of CRC diagnosis in the EMR. The KPNC cancer registry provided information on the anatomical site of cancer, cancer stage, and the administration of chemotherapy and radiation. Covariate data was 99.9% complete (2 missing observations for self-reported race and ethnicity and 3 missing observations for smoking history).

Study Outcomes

The primary end point was defined as the time from cancer diagnosis to the first occurrence of any component of the composite major adverse cardiovascular event (MACE) outcome, including death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke (3-component MACE). The 3-component MACE is recommended by the US Food and Drug Administration for use in cardiovascular safety studies.16 Deaths were identified from the California State death registry, the National Death Index (using Social Security Administration data), and KPNC electronic mortality files. Deaths were classified as cardiovascular specific if a cardiovascular cause was documented as an underlying or contributing cause of death on the death certificate through January 31, 2015. Validated International Classification of Diseases codes were used to identify nonfatal myocardial infarction and nonfatal stroke in the EMR.17,18 The end point event database was constructed by investigators blinded to BMI and body composition values.

Statistical Analysis

Two time-to-event regression models were used to estimate hazard ratios (HRs) and 95% CIs for each body composition variable. The first regression model estimated the cause-specific hazard using a Cox proportional hazards regression model. The cause-specific hazard is interpreted as the magnitude of the relative change in the instantaneous rate of the occurrence of MACE in patients who are event free.19 The second regression model estimated the subdistribution hazard using a Fine-Gray competing risk model.20 The subdistribution hazard is interpreted as the magnitude of relative change in the instantaneous rate of the occurrence of MACE in patients who are event free or who have experienced a competing event (eg, death from noncardiovascular causes, such as CRC, which we previously reported is associated with body composition).21,22 Detailed comparisons of these 2 regression models are described elsewhere.23,24 Contrasts were estimated to test for trends across categories.

Covariates were chosen a priori and included age, sex, race, ethnicity, cancer site, cancer stage, cancer treatment, smoking history, hypertension, hyperlipidemia, and type 2 diabetes; analyses of CT-derived measures of body composition were adjusted for patient height.25 One subgroup was specified a priori, to test if sex modified any associations between BMI and body composition with the risk of MACE. Effect modification was examined by adding a statistical interaction term to the regression model. Correlations between BMI and measures of body composition were quantified using the Pearson correlation coefficient with 95% CI.

Results
Characteristics of the Study Cohort

Of 2839 participants, 1384 were men and 1455 were women with an average (SD) age of 61.9 (11.5) years (range, 19-80 years). Many participants were former (n = 1127, 40%) or current smokers (n = 340, 12%) and had hypertension (n = 1150, 55%), hyperlipidemia (n = 1389, 49%), and type 2 diabetes (n = 573, 20%) (Table 1).

Computed tomographic images were obtained at a median of 6 days (interquartile range [IQR], 0-13) after results of a biopsy confirmed diagnosis of CRC. By a median follow-up of 6.8 years (IQR, 5.2- 8.3), MACE had occurred in 366 participants (12.9%). The cumulative incidence of MACE at 1, 3, and 10 years after diagnosis was 3.4%, 5.9%, 19.1%, respectively (Figure).

Correlation Between BMI and Body Composition

Body mass index was positively correlated with visceral adipose tissue area (r = 0.61; 95% CI, 0.59-0.63), subcutaneous adipose tissue area (r = 0.83; 95% CI, 0.82-0.85), and muscle mass (r = 0.41; 95% CI, 0.38-0.44). Body mass index was negatively correlated with muscle radiodensity (r = −0.33; 95% CI, −0.37 to −0.30).

BMI, Body Composition, and Major Adverse Cardiovascular Events

Body mass index was not associated with risk of MACE. For BMI categories greater than 35 vs 18.5 to 24.9, the HR was 1.23 (95% CI, 0.85-1.77; P = .50 for trend) (Table 2). Contrasting the highest to lowest quintile of visceral adipose tissue area, the multivariable-adjusted cause-specific HR for MACE was 1.54 (95% CI, 1.02-2.31; P = .04 for trend) (eFigure 2 in the Supplement). Conversely, subcutaneous adipose tissue area was not associated with risk of MACE. Contrasting the highest vs the lowest quintile, the HR for MACE was 1.15 (95% CI, 0.78-1.69; P = .65 for trend). Muscle mass was not associated with risk of MACE; comparing the highest vs lowest quintile, the HR for MACE was 0.96 (95% CI, 0.57-1.61; P = .92 for trend). Contrasting the highest (eg, less lipid stored in the muscle) to lowest quintile of muscle radiodensity, the multivariable-adjusted cause-specific HR for MACE was 0.67 (95% CI, 0.44-1.03; P = .02 for trend). Sex did not modify the association between any body composition measure and risk of MACE (results not shown). Effect estimates did not meaningfully differ when body composition measures were analyzed in their continuous form (eTable 1 and eFigure 3 in the Supplement), when body composition measures were additionally adjusted for BMI (eTable 2 in the Supplement), or in a variety of sensitivity analyses (eTables 3-6 and eFigure 4 in the Supplement).

Discussion

In this population-based cohort, 1 of 5 patients experienced MACE within 10 years after CRC diagnosis. Visceral adiposity but not subcutaneous adiposity was statistically significantly associated with the risk of MACE. To our knowledge, we are among the first to study muscle mass and muscle radiodensity in relation to MACE in CRC; of these muscle characteristics, muscle radiodensity was statistically significantly associated with the risk of MACE whereas muscle mass was not. These associations were independent of other established cardiovascular risk factors, including smoking, hypertension, hyperlipidemia, and type 2 diabetes. Surprisingly, BMI was not associated with the risk of MACE in this cohort.

Incident CRC and CVD share many risk factors, including excess adiposity.26 Improvements in early cancer detection and chemotherapy efficacy have reduced the risk of disease recurrence and cancer-specific mortality in CRC survivors.27 However, the increasing burden of CVD in this population may compromise improvements in overall survival.28 Among 1966 Australian patients with CRC, the 3-year cumulative incidence of CVD after diagnosis of CRC was 16%.29 Among 72 408 Medicare beneficiaries with CRC, the 10-year cumulative incidence of CVD after diagnosis of CRC was 57%.5 Physicians should be aware that patients diagnosed with CRC are not only at risk for cancer recurrence but are at high risk of developing CVD at some point during their survivorship trajectory. Furthermore, this CVD risk is not described by BMI alone, and other body composition measures should be considered to identify patients most in need of preventive cardiovascular care.

Visceral adipose tissue area and muscle radiodensity were identified as risk factors for MACE. The risk of MACE was higher among patients in the highest quintile of visceral adiposity, compared with those in the lowest quintile. These data are consistent with observations that waist circumference (an anthropometric proxy measure for visceral adipose tissue area) is independently associated with CVD.30 A meta-analysis of 15 studies with 258 114 participants demonstrated that each 1-cm increase in waist circumference increased the risk of a CVD event by 2%.31

The risk of MACE was lower among patients in the highest quintile of muscle radiodensity (ie, less lipid stored in the skeletal muscle), compared with those in the lowest quintile. The exact mechanisms linking muscle radiodensity to MACE are not clear. Muscle is an ectopic adiposity depot, and the accumulation of adiposity within skeletal muscle alters whole-body metabolism, manifesting in insulin resistance,14 impaired glucose tolerance, and type 2 diabetes.32 In addition, intermuscular adiposity is positively associated with a variety of proinflammatory mediators that are associated with CVD risk, such as C-reactive protein, interleukin-6, and tumor necrosis factor.33 Further research is necessary to better understand the mechanisms linking muscle radiodensity to MACE.

In CRC survivors, we found no association between BMI and the risk of MACE. Uncertainty exists regarding the use of BMI as the optimal measure for cardiovascular risk stratification.34 In a nationally representative sample of 13 601 adults, obesity defined as a BMI category of 30 or more had a high sensitivity (≥95%) but poor specificity (36%-49%) in identification of obesity defined using percent body fat with bioelectric impedance analysis.9 A meta-analysis of 10 studies demonstrated that anthropometric measures of visceral adiposity are superior to BMI in identifying cardiovascular risk factors, including hyperlipidemia, hypertension, and type 2 diabetes.8

Limitations

The main limitation of this study is the observational design, which precludes our ability to rule out residual confounding. The data used in our analysis were collected for clinical care purposes. The reliance of this study on administrative codes within the EMR precluded our ability to obtain information on patient behaviors such as physical activity, dietary patterns, and other behaviors or health conditions that may influence body composition and the risk of MACE. The measurement of physical activity was adopted across the KPNC health care system beginning in October 2009.35 Consequently, only 230 (8%) of our cohort had physical activity measures available at the time of CRC diagnosis. Moreover, using administrative codes for the identification of CVD risk factors has high specificity (>0.95), but low sensitivity (<0.76) compared with manual medical record review.36 Measures of body composition were obtained at the third lumbar vertebrae at a solitary time point. Although this anatomical region is correlated with whole-body tissue volumes,13 it is not known how whole-body tissue volume or changes in this volume are associated with cardiovascular risk.

Conclusions

Because patients with CRC are at a higher risk of developing CVD than the general population, physicians may wish to refine cardiovascular risk management by integrating quantitative measures of body composition that can be derived automatically from CT scans that are routinely obtained during CRC diagnosis. This precision prevention approach to cardiovascular risk management may help to cost-effectively allocate limited resources such as dietary and physical activity counseling to patients who may be most likely to benefit from lifestyle counseling. Our findings suggest that body composition measures that are collected using routine CT images, including visceral adiposity and muscle radiodensity, can be used to assess risk for MACEs in patients with CRC, whereas BMI may have limited use for determining cardiovascular risk in this patient population.

Back to top
Article Information

Accepted for Publication: February 11, 2019.

Corresponding Author: Justin C. Brown, PhD, Pennington Biomedical Research Center, 6400 Perkins Rd, Baton Rouge, LA 70808 (justin.brown@pbrc.edu).

Published Online: May 16, 2019. doi:10.1001/jamaoncol.2019.0695

Author Contributions: Drs Brown and Weltzien had full access to all the data in the study and take responsibility for the integrity of the data and accuracy of the data analysis.

Study concept and design: Brown, Caan, Prado, Cespedes Feliciano, Kroenke.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Brown, Cespedes Feliciano, Kroenke, Meyerhardt.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Brown, Weltzien, Cespedes Feliciano, Kroenke.

Obtained funding: Brown, Caan, Prado.

Administrative, technical, or material support: Brown, Caan, Cespedes Feliciano.

Study supervision: Brown, Caan, Prado, Meyerhardt.

Conflict of Interest Disclosures: Dr Brown reports grants from the National Cancer Institute (paid to his institution). Dr Prado reports personal fees from Abbott Nutrition outside the submitted work. Dr Cespedes Feliciano reports grants from National Cancer Institute (K01-CA226155, R01CA175011) during the conduct of the study. No other disclosures were reported.

Funding/Support: Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under award numbers K99-CA218603, R01-CA175011, and R25-CA203650.

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

References
1.
Siegel  RL, Miller  KD, Fedewa  SA,  et al.  Colorectal cancer statistics, 2017.  CA Cancer J Clin. 2017;67(3):177-193. doi:10.3322/caac.21395PubMedGoogle ScholarCrossref
2.
Jemal  A, Ward  EM, Johnson  CJ,  et al.  Annual report to the nation on the status of cancer, 1975-2014, featuring survival.  J Natl Cancer Inst. 2017;109(9). doi:10.1093/jnci/djx030PubMedGoogle Scholar
3.
Baade  PD, Fritschi  L, Eakin  EG.  Non-cancer mortality among people diagnosed with cancer (Australia).  Cancer Causes Control. 2006;17(3):287-297. doi:10.1007/s10552-005-0530-0PubMedGoogle ScholarCrossref
4.
van Erning  FN, van Steenbergen  LN, Lemmens  VEPP,  et al.  Conditional survival for long-term colorectal cancer survivors in the Netherlands: who do best?  Eur J Cancer. 2014;50(10):1731-1739. doi:10.1016/j.ejca.2014.04.009PubMedGoogle ScholarCrossref
5.
Kenzik  KM, Balentine  C, Richman  J, Kilgore  M, Bhatia  S, Williams  GR.  New-onset cardiovascular morbidity in older adults with stage I to III colorectal cancer.  J Clin Oncol. 2018;36(6):609-616. doi:10.1200/JCO.2017.74.9739PubMedGoogle ScholarCrossref
6.
El-Shami  K, Oeffinger  KC, Erb  NL,  et al.  American Cancer Society colorectal cancer survivorship care guidelines.  CA Cancer J Clin. 2015;65(6):428-455. doi:10.3322/caac.21286PubMedGoogle ScholarCrossref
7.
Kim  SH, Després  JP, Koh  KK.  Obesity and cardiovascular disease: friend or foe?  Eur Heart J. 2016;37(48):3560-3568. doi:10.1093/eurheartj/ehv509PubMedGoogle ScholarCrossref
8.
Lee  CMY, Huxley  RR, Wildman  RP, Woodward  M.  Indices of abdominal obesity are better discriminators of cardiovascular risk factors than BMI: a meta-analysis.  J Clin Epidemiol. 2008;61(7):646-653. doi:10.1016/j.jclinepi.2007.08.012PubMedGoogle ScholarCrossref
9.
Romero-Corral  A, Somers  VK, Sierra-Johnson  J,  et al.  Accuracy of body mass index in diagnosing obesity in the adult general population.  Int J Obes (Lond). 2008;32(6):959-966. doi:10.1038/ijo.2008.11PubMedGoogle ScholarCrossref
10.
Prado  CM, Heymsfield  SB.  Lean tissue imaging: a new era for nutritional assessment and intervention.  JPEN J Parenter Enteral Nutr. 2014;38(8):940-953. doi:10.1177/0148607114550189PubMedGoogle ScholarCrossref
11.
Brown  JC, Cespedes Feliciano  EM, Caan  BJ.  The evolution of body composition in oncology-epidemiology, clinical trials, and the future of patient care: facts and numbers.  J Cachexia Sarcopenia Muscle. 2018;9(7):1200-1208. doi:10.1002/jcsm.12379PubMedGoogle ScholarCrossref
12.
World Health Organization.  Obesity: Preventing and Managing the Global Epidemic. Geneva, Switzerland: World Health Organization; 2000.
13.
Shen  W, Punyanitya  M, Wang  Z,  et al.  Total body skeletal muscle and adipose tissue volumes: estimation from a single abdominal cross-sectional image.  J Appl Physiol (1985). 2004;97(6):2333-2338. doi:10.1152/japplphysiol.00744.2004PubMedGoogle ScholarCrossref
14.
Aubrey  J, Esfandiari  N, Baracos  VE,  et al.  Measurement of skeletal muscle radiation attenuation and basis of its biological variation.  Acta Physiol (Oxf). 2014;210(3):489-497. doi:10.1111/apha.12224PubMedGoogle ScholarCrossref
15.
Charlson  M, Szatrowski  TP, Peterson  J, Gold  J.  Validation of a combined comorbidity index.  J Clin Epidemiol. 1994;47(11):1245-1251. doi:10.1016/0895-4356(94)90129-5PubMedGoogle ScholarCrossref
16.
US Department of Health and Human Services, US Food and Drug Administration, Center for Drug Evaluation and Research. Guidance for industry: diabetes mellitus—evaluating cardiovascular risk in new antidiabetic therapies to treat type 2 diabetes. https://www.fda.gov/downloads/Drugs/Guidances/ucm071627.pdf. Accessed April 2, 2019.
17.
McCormick  N, Bhole  V, Lacaille  D, Avina-Zubieta  JA.  Validity of diagnostic codes for acute stroke in administrative databases: a systematic review.  PLoS One. 2015;10(8):e0135834. doi:10.1371/journal.pone.0135834PubMedGoogle ScholarCrossref
18.
McCormick  N, Lacaille  D, Bhole  V, Avina-Zubieta  JA.  Validity of myocardial infarction diagnoses in administrative databases: a systematic review.  PLoS One. 2014;9(3):e92286. doi:10.1371/journal.pone.0092286PubMedGoogle ScholarCrossref
19.
Austin  PC, Fine  JP.  Practical recommendations for reporting Fine-Gray model analyses for competing risk data.  Stat Med. 2017;36(27):4391-4400. doi:10.1002/sim.7501PubMedGoogle ScholarCrossref
20.
Fine  JP, Gray  RJ.  A proportional hazards model for the subdistribution of a competing risk.  J Am Stat Assoc. 1999;94(446):496-509. doi:10.1080/01621459.1999.10474144Google ScholarCrossref
21.
Caan  BJ, Meyerhardt  JA, Kroenke  CH,  et al.  Explaining the obesity paradox: the association between body composition and colorectal cancer survival.  Cancer Epidemiol Biomarkers Prev. 2017;26(7):1008-1015. doi:10.1158/1055-9965.EPI-17-0200PubMedGoogle ScholarCrossref
22.
Kroenke  CH, Prado  CM, Meyerhardt  JA,  et al.  Muscle radiodensity and mortality in patients with colorectal cancer.  Cancer. 2018;124(14):3008-3015. doi:10.1002/cncr.31405PubMedGoogle ScholarCrossref
23.
Austin  PC, Lee  DS, Fine  JP.  Introduction to the analysis of survival data in the presence of competing risks.  Circulation. 2016;133(6):601-609. doi:10.1161/CIRCULATIONAHA.115.017719PubMedGoogle ScholarCrossref
24.
Dignam  JJ, Zhang  Q, Kocherginsky  M.  The use and interpretation of competing risks regression models.  Clin Cancer Res. 2012;18(8):2301-2308. doi:10.1158/1078-0432.CCR-11-2097PubMedGoogle ScholarCrossref
25.
Kronmal  RA.  Spurious correlation and the fallacy of the ratio standard revisited.  J R Stat Soc Ser A Stat Soc. 1993;156(3):379-392. doi:10.2307/2983064Google ScholarCrossref
26.
Koene  RJ, Prizment  AE, Blaes  A, Konety  SH.  Shared risk factors in cardiovascular disease and cancer.  Circulation. 2016;133(11):1104-1114. doi:10.1161/CIRCULATIONAHA.115.020406PubMedGoogle ScholarCrossref
27.
Shi  Q, Andre  T, Grothey  A,  et al.  Comparison of outcomes after fluorouracil-based adjuvant therapy for stages II and III colon cancer between 1978 to 1995 and 1996 to 2007: evidence of stage migration from the ACCENT database.  J Clin Oncol. 2013;31(29):3656-3663. doi:10.1200/JCO.2013.49.4344PubMedGoogle ScholarCrossref
28.
van Leersum  NJ, Janssen-Heijnen  ML, Wouters  MW,  et al.  Increasing prevalence of comorbidity in patients with colorectal cancer in the South of the Netherlands 1995-2010.  Int J Cancer. 2013;132(9):2157-2163. doi:10.1002/ijc.27871PubMedGoogle ScholarCrossref
29.
Hawkes  AL, Lynch  BM, Owen  N, Aitken  JF.  Lifestyle factors associated concurrently and prospectively with co-morbid cardiovascular disease in a population-based cohort of colorectal cancer survivors.  Eur J Cancer. 2011;47(2):267-276. doi:10.1016/j.ejca.2010.10.002PubMedGoogle ScholarCrossref
30.
Pouliot  MC, Després  JP, Lemieux  S,  et al.  Waist circumference and abdominal sagittal diameter: best simple anthropometric indexes of abdominal visceral adipose tissue accumulation and related cardiovascular risk in men and women.  Am J Cardiol. 1994;73(7):460-468. doi:10.1016/0002-9149(94)90676-9PubMedGoogle ScholarCrossref
31.
de Koning  L, Merchant  AT, Pogue  J, Anand  SS.  Waist circumference and waist-to-hip ratio as predictors of cardiovascular events: meta-regression analysis of prospective studies.  Eur Heart J. 2007;28(7):850-856. doi:10.1093/eurheartj/ehm026PubMedGoogle ScholarCrossref
32.
Goodpaster  BH, Krishnaswami  S, Resnick  H,  et al.  Association between regional adipose tissue distribution and both type 2 diabetes and impaired glucose tolerance in elderly men and women.  Diabetes Care. 2003;26(2):372-379. doi:10.2337/diacare.26.2.372PubMedGoogle ScholarCrossref
33.
Gustafson  B, Smith  U.  Cytokines promote Wnt signaling and inflammation and impair the normal differentiation and lipid accumulation in 3T3-L1 preadipocytes.  J Biol Chem. 2006;281(14):9507-9516. doi:10.1074/jbc.M512077200PubMedGoogle ScholarCrossref
34.
Iliodromiti  S, Celis-Morales  CA, Lyall  DM,  et al.  The impact of confounding on the associations of different adiposity measures with the incidence of cardiovascular disease: a cohort study of 296 535 adults of white European descent.  Eur Heart J. 2018;39(17):1514-1520. doi:10.1093/eurheartj/ehy057PubMedGoogle ScholarCrossref
35.
Coleman  KJ, Ngor  E, Reynolds  K,  et al.  Initial validation of an exercise “vital sign” in electronic medical records.  Med Sci Sports Exerc. 2012;44(11):2071-2076. doi:10.1249/MSS.0b013e3182630ec1PubMedGoogle ScholarCrossref
36.
Birman-Deych  E, Waterman  AD, Yan  Y, Nilasena  DS, Radford  MJ, Gage  BF.  Accuracy of ICD-9-CM codes for identifying cardiovascular and stroke risk factors.  Med Care. 2005;43(5):480-485. doi:10.1097/01.mlr.0000160417.39497.a9PubMedGoogle ScholarCrossref
×