Abbreviations:
BMI (body mass index), CI (confidence interval), CRP (C-reactive protein), DFS (disease-free survival), HR (hazard ratio), MSS (melanoma-specific survival), OS (overall survival), SNP (single-nucleotide polymorphism)Obesity is a known risk factor for cancer development (
Arnold et al., 2016
, Basen-Engquist and Chang, 2011
, Renehan et al., 2015
) and death (Calle et al., 2003
). Obesity has been associated with an increased risk of developing melanoma in men (Sergentanis et al., 2013
) and with thicker primary melanomas (Skowron et al., 2015
). The inflammatory adipokine leptin promotes melanoma progression in mice (Amjadi et al., 2011
, Brandon et al., 2009
, Gogas et al., 2008
); elevated leptin levels may predict melanoma sentinel node metastasis (Oba et al., 2016
). We hypothesized that elevated body mass index (BMI) would be associated with decreased melanoma patient survival through chronic inflammation, as indicated by levels of C-reactive protein (CRP). We also evaluated whether obesity-related genetic variants (single-nucleotide polymorphisms [SNPs]) play a role in melanoma patient outcomes.Among 1,804 patients with melanoma enrolled from 1998 to 2008, BMI information was available for 1,186 patients, and 725 underwent CRP determination (Supplementary Methods and Figure S1 online). A total of 2.65 million SNPs were available. The median BMI was 28.2 kg/m2, obtained at a median of 0.10 years after diagnosis (Table 1). The median CRP level was 1.71 mg/l. The raw CRP distribution was skewed (skewness = 8.06, kurtosis = 99.40, data not shown), but was close to normally distributed after log transformation (skewness = −0.01, kurtosis = 0.40, data not shown). Therefore, we used the log-transformed CRP (log[CRP]) in our analysis. Blood samples were obtained a median of 0.64 years after diagnosis, and the median follow-up duration was 8.3 years.
Table 1Relationship of demographic and clinical factors with body mass index in 1,186 patients with melanoma
| Characteristic | Total (N = 1,186) | Correlation with BMI | |
|---|---|---|---|
| Pearson r2 | P-value | ||
| BMI, median (IQR), kg/m2 | 28.2 (25.2–32.1) | – | – |
| Age at diagnosis, median (IQR), y | 52.4 (42.4–62.7) | 0.11 | 0.0002 |
| Sex, n (%) | |||
| Female | 471 (39.7) | <0.0001 | |
| Male | 715 (60.3) | ||
| Tumor thickness, median (IQR), mm | 1.10 (0.61–2.20) | 0.07 | |
| Stage at diagnosis, n (%) | |||
| I/II | 893 (75.3) | 0.1118 | |
| III/IV | 293 (24.7) | ||
| Time from diagnosis to blood draw, median (IQR), y | 0.64 (0.11–2.03) | – | – |
| CRP, median (IQR), mg/l | 1.71 (0.69–4.40) | 0.35 | <0.0001 |
| Follow-up time from diagnosis to disease relapse or censoring, median (IQR), y | 7.6 (3.81–10.11) | – | – |
| Follow-up time from diagnosis to death or censoring, Median (IQR), y | 8.3 (5.95–10.64) | – | – |
| Recurrence among all patients, n (%) | 334 (28.2) | – | – |
| Recurrence among stage I/II patients, n (%) | 197/893 (22.1) | – | – |
| Death from all causes, n (%) | 325 (27.4) | – | – |
| Death from melanoma, n (%) | 224 (18.9) | – | – |
Abbreviations: BMI, body mass index; CRP, C-reactive protein; IQR, interquartile range.
1 P-value from the Student t test for the association between categorical variables and BMI.
2 Correlation measured using BMI and logarithmic CRP levels.
Increased BMI was weakly associated with older age, increased tumor thickness, and increased log[CRP] (Table 1).
Elevated BMI was associated with shorter overall survival (OS) (hazard ratio [HR] = 1.20 per 5 kg/m2 increase in BMI, 95% confidence interval [CI] 1.11–1.29, P < 0.0001). In the multivariable analysis, after adjustment for age, sex, and stage, increased BMI remained associated with poorer OS (HR = 1.14, 95% CI 1.06–1.23, P = 0.0005) (Table 2). However, when log[CRP] was included, elevated CRP remained an independent predictor of poorer OS (HR = 1.21 per unit increase in log[CRP], 95% CI 1.10–1.34, P = 0.0001), but BMI was no longer associated with OS (HR = 1.08, 95% CI 0.97–1.19, P = 0.1531) (Table 2).
Table 2Association between body mass index and patient outcome in 1,186 patients with melanoma
| Melanoma-specific survival | Overall survival | |||||||
|---|---|---|---|---|---|---|---|---|
| Per 5 kg/m2 increase in BMI | Categorical BMI (≥30 kg/m2 versus <30 kg/m2) | Per 5 kg/m2 increase in BMI | Categorical BMI (≥30 kg/m2 versus <30 kg/m2) | |||||
| HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value | |
| Univariate analysis | ||||||||
| BMI | 1.20 (1.09–1.31) | 0.0002 | 1.65 (1.27–2.16) | 0.0002 | 1.20 (1.11–1.29) | <0.0001 | 1.56 (1.25–1.94) | <0.0001 |
| Multivariable analysis | ||||||||
| BMI | 1.15 (1.05–1.26) | 0.0037 | 1.47 (1.12–1.91) | 0.0049 | 1.14 (1.06–1.23) | 0.0005 | 1.45 (1.16–1.81) | 0.0010 |
| Age | 1.02 (1.01–1.03) | 0.0007 | 1.02 (1.01–1.03) | 0.0004 | 1.04 (1.03–1.05) | <0.0001 | 1.04 (1.03–1.05) | <0.0001 |
| Sex: male versus female | 1.29 (0.97–1.72) | 0.0833 | 1.27 (0.95–1.69) | 0.1052 | 1.31 (1.03–1.67) | 0.0272 | 1.29 (1.02–1.65) | 0.0374 |
| Stage: III/IV versus I/II | 4.63 (3.54–6.04) | <0.0001 | 4.57 (3.50–5.97) | <0.0001 | 2.91 (2.33–3.64) | <0.0001 | 2.89 (2.31–3.60) | <0.0001 |
| Multivariable analysis adjusted for CRP | ||||||||
| BMI | 1.08 (0.96–1.21) | 0.2268 | 1.33 (0.96–1.85) | 0.0858 | 1.08 (0.97–1.19) | 0.1531 | 1.32 (1.01–1.73) | 0.0455 |
| Logarithmic CRP | 1.26 (1.12–1.41) | 0.0002 | 1.26 (1.12–1.41) | 0.0002 | 1.21 (1.10–1.34) | 0.0001 | 1.21 (1.10–1.34) | 0.0001 |
| Age | 1.01 (1.00–1.03) | 0.0334 | 1.01 (1.00–1.03) | 0.0261 | 1.04 (1.03–1.05) | <0.0001 | 1.04 (1.03–1.05) | <0.0001 |
| Sex: male versus female | 1.56 (1.11–2.21) | 0.0115 | 1.56 (1.10–2.20) | 0.0122 | 1.61 (1.20–2.15) | 0.0014 | 1.60 (1.20–2.14) | 0.0015 |
| Stage: III/IV versus I/II | 3.67 (2.65–5.10) | <0.0001 | 3.67 (2.65–5.09) | <0.0001 | 2.27 (1.73–2.97) | <0.0001 | 2.26 (1.73–2.97) | <0.0001 |
Abbreviations: BMI, body mass index; CI, confidence interval; CRP, C-reactive protein; HR, hazard ratio.
1 Cox proportional hazards analyses.
2 725 patients with CRP data available.
We also observed associations between BMI and melanoma-specific survival (MSS) in both univariate (HR = 1.20, 95% CI 1.09–1.31, P = 0.0002) and multivariable (HR= 1.15, 95% CI 1.05–1.26, P = 0.0037) analyses (Table 2). In addition, we found associations between BMI and disease-free survival (DFS) in patients with stage I/II melanoma (HR = 1.17, 95% CI 1.05-1.30, P = 0.0033 in univariate analysis; HR = 1.10, 95% CI 0.98–1.23, P = 0.0994 in multivariable analysis) (Supplementary Table S1 online). As in our OS analysis, after incorporation of log[CRP], BMI was no longer associated with MSS (HR = 1.08, 95% CI 0.96–1.21, P = 0.2268) (Table 2) or DFS (HR = 0.99, 95% CI 0.86–1.13, P = 0.8623) (Supplementary Table S1).
We further evaluated BMI by dichotomizing the variable at the standard cutoff for obesity, 30 kg/m2. Patients with a BMI ≥ 30 kg/m2 had poorer OS (HR = 1.56, 95% CI 1.25–1.94, P < 0.0001; Table 2) and MSS (HR = 1.65, 95% CI 1.27–2.16, P = 0.0002; Table 2) than did patients with a BMI < 30 kg/m2. Patients with stage I/II melanoma and BMI ≥ 30 kg/m2 had poorer DFS (HR = 1.48, 95% CI 1.11–1.96, log-rank P = 0.0077; Supplementary Table S1). The associations between BMI and the three outcome measures were weaker after adjustment for age, sex, and stage (OS: HR = 1.45, 95% CI 1.16–1.81, P = 0.0010; MSS: HR = 1.47, 95% CI 1.12–1.91, P = 0.0049; DFS: HR = 1.29, 95% CI 0.97–1.72, P = 0.0836 [nonsignificant]) (Table 2, Supplementary Table S1). When log[CRP] was incorporated, the association between BMI and OS became weaker (HR= 1.32, 95% CI 1.01–1.73, P = 0.0455), and the relationships between BMI and MSS (HR = 1.33, 95% CI 0.96–1.85, P = 0.0858) and BMI and DFS (HR = 1.12, 95% CI 0.79–1.59, P = 0.5304) lost significance (Table 2, Supplementary Table S1).
We next excluded patients whose first BMI measurement or blood draw was performed more than 1 year after diagnosis and again assessed the relationship between BMI and melanoma outcomes for the remaining 760 patients. We observed the same pattern of associations as in the full patient cohort (Supplementary Tables S2 and S3 online).
Although underweight patients in our population tended to have a lower risk of death than did normal-weight patients (HR close to 0, likely because sample size was small; Supplementary Table S4 online), overweight patients had a trend toward an elevated risk of disease recurrence and death. These risks were significantly increased in patients who were obese (P < 0.05).
Previously validated BMI-associated SNPs were assessed for association with melanoma risk and patient outcomes. None of the 82 examined SNPs significantly predicted any outcome after correction for multiple testing (P = 0.05/82 = 6.10 × 10−4) (Supplementary Table S5 online). Fifteen SNPs from different gene regions reached nominal significance (P < 0.05) in predicting BMI. In addition, several BMI-associated SNPs were associated with melanoma risk or outcome (P < 0.05). In particular, the C allele in the rs17782313 SNP (within MC4R, the melanocortin 4 receptor) was nominally associated with increased BMI (beta coefficient = 0.65, P = 0.0249), showed a trend toward association with elevated CRP (beta coefficient = 0.13, P = 0.0653), and was associated with poorer OS (HR = 1.11, 95% CI 1.00–1.23, P = 0.0422) and poorer MSS (HR = 1.16, 95% CI 1.03–1.30, P = 0.0135), but not DFS (HR = 1.08, 95% CI 0.95–1.22, P = 0.237), among patients with stage I/II melanoma (Supplementary Table S5).
To our knowledge, this investigation is the first to report associations between elevated BMI and poorer melanoma patient outcomes after adjustment for sex, age, and stage. A previous population-based cohort study detected no association between elevated BMI and melanoma mortality (
Calle et al., 2003
), but that study was small, did not assess melanoma stage or recurrence, and did not include biomarker data. Furthermore, because prior investigations have identified strong associations between obesity and elevated levels of inflammatory markers, including CRP (Ellulu et al., 2016
, Oba et al., 2016
), our finding that the outcome associations identified in the current investigation were weakened or became insignificant after adjustment for CRP suggests that systemic inflammation and/or metabolic syndrome may be involved in BMI-associated melanoma progression. Finally, our data suggest that genetic variations underlying elevated BMI and CRP might also contribute to poorer melanoma patient survival. Further investigation is needed to confirm these findings and to determine whether control of body weight and/or interventions to reduce chronic inflammation and metabolic syndrome could be beneficial to patients with melanoma.All individuals gave written informed consent to participate under an Institutional Review Board-approved protocol.
Conflict of Interest
JEG claims a consulting role for Merck. JW claims roles for the following companies: Honoraria for Dava Oncology, H. Lee Moffitt Cancer Center and Research Institute; Consulting or Advisory Role for Genentech, GlaxoSmithKline, Novartis; Speakers’ Bureau for Illumina, Bristol-Myers Squibb, Dava Oncology; Research Funding for Genentech, GlaxoSmithKline, Bristol-Myers Squibb; Travel, Accommodations, Expenses for Bristol-Myers Squibb, JP Morgan. All other authors state no conflict of interest.
Acknowledgments
We thank the individuals who volunteered to participate in this project. We also thank Dr Amy Ninetto in the Department of Scientific Publications at The University of Texas MD Anderson Cancer Center who edited the manuscript. This work was supported by the National Cancer Institute of the National Institutes of Health through SPORE grant P50 CA093459 and Cancer Center Support Grant P30 CA016672 (Clinical Trials Support Resource), as well as by philanthropic contributions to The University of Texas MD Anderson Cancer Center Moon Shots Program, The University of Texas MD Anderson Cancer Center Various Donors Melanoma and Skin Cancers Priority Program Fund; the Miriam and Jim Mulva Research Fund; the McCarthy Skin Cancer Research Fund and the Marit Peterson Fund for Melanoma Research.
Supplementary Material
- Supplementary Data
References
- Leptin promotes melanoma tumor growth in mice related to increasing circulating endothelial progenitor cells numbers and plasma NO production.J Exp Clin Cancer Res. 2011; 30: 21
- Duration of Adulthood Overweight, Obesity, and Cancer Risk in the Women's Health Initiative: A Longitudinal Study from the United States.PLoS Med. 2016; 13: e1002081
- Obesity and cancer risk: recent review and evidence.Curr Oncol Rep. 2011; 13: 71-76
- Obesity promotes melanoma tumor growth: role of leptin.Cancer Biol Ther. 2009; 8: 1871-1879
- Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults.N Engl J Med. 2003; 348: 1625-1638
- Obesity can predict and promote systemic inflammation in healthy adults.Int J Cardiol. 2016; 215: 318-324
- Melanoma risk in association with serum leptin levels and lifestyle parameters: a case-control study.Ann Oncol. 2008; 19: 384-389
- Elevated serum leptin levels are associated with an increased risk of sentinel lymph node metastasis in cutaneous melanoma.Medicine. 2016; 95: e3073
- Adiposity and cancer risk: new mechanistic insights from epidemiology.Nat Rev Cancer. 2015; 15: 484-498
- Obesity and risk of malignant melanoma: a meta-analysis of cohort and case-control studies.Eur J Cancer. 2013; 49: 642-657
- Role of obesity on the thickness of primary cutaneous melanoma.J Eur Acad Dermatol Venereol. 2015; 29: 262-269
Article info
Publication history
Published online: April 22, 2017
Accepted manuscript published online 23 April 2017; corrected proof published online 16 June 2017Identification
Copyright
© 2017 The Authors. Published by Elsevier, Inc. on behalf of the Society for Investigative Dermatology.
User license
Elsevier user license | How you can reuse 
Elsevier's open access license policy
Elsevier user license
Permitted
For non-commercial purposes:
- Read, print & download
- Text & data mine
- Translate the article
Not Permitted
- Reuse portions or extracts from the article in other works
- Redistribute or republish the final article
- Sell or re-use for commercial purposes
Elsevier's open access license policy
