Obesity is Associated with Longer Survival Independent of Sarcopenia and Myosteatosis in Metastatic and/or Castrate-Resistant Prostate Cancer

Abstract

Purpose:

Obesity (body mass index 30 kg/m² or greater) is associated with better overall survival in metastatic prostate cancer. Conversely, low muscle mass (sarcopenia) and low muscle radiodensity (myosteatosis) are associated with worse overall survival in many cancers. This study seeks to evaluate the relationship of sarcopenia, myosteatosis and obesity with overall survival in men with metastatic or castrate-resistant prostate cancer.

Materials and Methods:

Retrospective analysis of men with metastatic or castrate-resistant prostate cancer and computerized tomography of abdomen/pelvis presenting to the Vanderbilt Comprehensive Prostate Cancer Clinic from 2012 to 2017 was performed. Demographic, pathological and survival data were described, with sarcopenia and myosteatosis determined from abdominal skeletal muscle area and skeletal muscle radiodensity, respectively. Kaplan-Meier curves and log-rank tests estimated the effect of body composition on survival. Multivariable Cox proportional hazard models were performed adjusting for age, Charlson comorbidity index, race and clinical stage. ANOVA was used to compare obese and nonobese men with and without sarcopenia or myosteatosis.

Results:

Of 182 men accrued, 37.4% were obese, 53.3% sarcopenic and 59.3% myosteatotic. Over a median followup of 33.9 months, body mass index was associated with reduced mortality (HR 0.93, p=0.02), as was visceral adiposity (HR 0.99, p=0.003). Men with high body mass index without sarcopenia/myosteatosis lived significantly longer than men with high body mass index with sarcopenia/myosteatosis or normal body mass index men (F[3,91]=4.03, p=0.01).

Conclusions:

Both high body mass index and visceral adiposity in metastatic or castrate-resistant prostate cancer are associated with reduced mortality, independent of sarcopenia and myosteatosis. Therefore, routine clinical workup should include calculation of body mass index and measurement of waist circumference. Morphometric analysis of computerized tomography imaging can identify patients at risk for poor prognosis.

PROSTATE cancer accounts for 1 in 5 new cancer diagnoses in men, with risk of progression to castrate resistance.¹ Castrate-resistant prostate cancer has a poor prognosis with a median survival of 9 to 30 months.² Progression to metastatic and/or CRPC and their disease-specific outcomes are driven by a complex interplay of metabolic factors and body composition. A greater understanding of these interactions could lead to more sophisticated risk stratification, treatment planning, prognostic assessment and disease monitoring.

The obesity paradox is at the center of investigation regarding the role of human metabolism and body composition in prostate cancer outcomes. In some studies, obesity (BMI ≥30 kg/m²) increases the risk for high grade prostate cancer³ as well as recurrence after prostatectomy.⁴ In contrast, obesity may be protective for patients with advanced disease. In men with metastatic hormone-sensitive prostate cancer, obesity was associated with better progression-free and overall survival.⁵ In nonmetastatic castrate-resistant prostate cancer, obesity was associated with reduced all cause mortality.⁶ In men with metastatic castrate-resistant prostate cancer, obesity was associated with reduced prostate cancer-specific mortality and all cause mortality.⁷ Thus, the use of BMI as an indicator of adiposity status and as a useful prognostic marker remains unclear.

Because BMI does not discriminate between the type or distribution of adipose tissue, the amount of subcutaneous and visceral adipose tissue may confound the role of BMI in mPC/CRPC risk and outcomes. Indeed, BMI, waist circumference, SAT and VAT have all been associated with increased prostate cancer fatality.⁸ Yet, high SAT is associated with higher OS and progression-free survival in CRPC.⁹ Increased VAT is associated with higher Gleason score at prostatectomy,¹⁰ and the association of high BMI with high grade prostate cancer may be mediated by high VAT.¹¹ Furthermore, in mCRPC high VAT was independently associated with reduced survival.¹² Moreover, a high ratio of VAT to SAT was associated with reduced survival in patients with mCRPC treated with docetaxel, but only in patients with normal BMI.¹³ As such, the relationship among BMI, adiposity and cancer outcomes is complex, and may vary with prostate cancer stage.

Other possible confounders in the role of obesity in patients with mPC/CRPC is the quantity and quality of muscle mass. The progressive atrophy of muscle mass (ie sarcopenia) is associated with many adverse health outcomes, including increased mortality in patients with CRPC undergoing chemotherapy.¹² The coexistence of sarcopenia and obesity (sarcopenic obesity) is prevalent in men on antigen deprivation therapy.¹⁴ In a variety of cancers, including gynecologic, gastrointestinal, head/neck and lung, patients with sarcopenic obesity have a worse prognosis than those with sarcopenia alone.¹⁵ Myosteatosis, defined as the excess accumulation of ectopic fat in muscle and detected radiographically as low muscle density, worsens OS in lymphoma, renal, gastrointestinal and gynecologic cancers.¹⁶ Further, myosteatosis is strongly associated with frailty and poor aerobic fitness,¹⁷ increasing susceptibility to adverse outcomes. Notably, relationships between adiposity, sarcopenia, myosteatosis and prostate cancer-specific outcomes have not been fully explored.

Greater understanding of these relationships may provide insights to help elucidate the obesity paradox in mPC/CRPC. This study was designed to investigate the roles of obesity (defined by BMI), adiposity (defined by abdominal SAT and VAT), sarcopenia and myosteatosis on survival in men with metastatic and/or castrate-resistant prostate cancer. We hypothesized that increased VAT, sarcopenia and myosteatosis, but not BMI, would be associated with decreased overall survival in this population.

MATERIALS AND METHODS

Study Population and Electronic Medical Record Abstraction

Retrospective analysis of CT imaging and electronic medical record data was performed for all men with mPC/CRPC (188) who presented to the Vanderbilt University Medical Center Comprehensive Prostate Cancer Clinic from May 2012 to December 2017. Patients without CT imaging data and who were not treated by a urologist (6) were excluded from the final cohort (182). Castrate resistance was defined according to the Prostate Cancer Clinical Trials Working Group¹⁸ and clinical staging was determined by the attending urologist based on radiographic evidence of metastatic disease. Demographic, pathological and survival information were obtained via electronic medical record review. The duration of followup was the number of months from first presentation to the Comprehensive Prostate Cancer Clinic to the date of death for deceased patients or the date of last documented encounter for surviving patients. The study was approved by the Vanderbilt University Medical Center Institutional Review Board (IRB #160431).

Imaging Analysis

CT images of the abdomen/pelvis, acquired upon presentation to the clinic as part of routine cancer workup, were used for morphometric analysis via an automated version of sliceOmatic software.¹⁹ One axial slice at the level of the L3 vertebra was used to quantify total abdominal muscle, VAT and SAT areas (cm²) along with abdominal muscle density in Hounsfield units. To standardize these measures across patients, abdominal muscle, VAT and SAT were indexed to the patient’s height (cm²/m²). Previously established cut points were utilized to categorize patients as sarcopenic or nonsarcopenic based on the skeletal muscle index (SMI <43 cm²/m² for BMI <25 and SMI <53 cm²/m² for BMI ≥25).²⁰ Myosteatosis was defined based on skeletal muscle radiodensity of less than 41 HU for BMI less than 25 and less than 33 HU for BMI 25 or greater.²⁰ Sarcopenic obesity was defined as meeting criteria for obesity (BMI ≥30 kg/m²) and sarcopenia. All morphometric analyses were performed by 1 trained investigator who was blinded to BMI and age to minimize bias from technical or intrapersonal variability.

Statistical Analysis

Data analyses were performed using SPSS® v26 (IBM®, Armonk, New York) and figures created using Prism v8 (GraphPad®, La Jolla, California). All tests were 2-tailed, and the probability of Type 1 error was 0.05. The primary outcome was survival, defined as time to death in months. Mantel-Cox log-rank tests were performed to compare survival between groups. Multivariable Cox proportional hazards models were utilized to identify morphometric and clinicopathological factors associated with mortality from prostate cancer. One-way analysis of variance (ANOVA) was used to compare obese men with and without sarcopenia or myosteatosis, and nonobese men with and without sarcopenia and myosteatosis.

RESULTS

Patient Characteristics

Of the 182 men with complete data 156 (85.7%) were Caucasian, 23 (12.6%) were African American and the remaining were Hispanic, Native American or of unknown ethnicity (table 1). Median age at presentation was 71.5 years (range 50–97 years) with a median BMI of 28.8 kg/m² (range 17.8–54.7). Median Charlson comorbidity index was 6.0 (range 0–14). Patients were followed longitudinally for a median of 33.9 months (range 0–176), with 90 (47.9%) men alive at the longest point of followup (176 months).

Table 1.

Demographics and clinical characteristics of men presenting to Comprehensive Prostate Cancer Clinic

Median yrs age (IQR):	71.5	(64.9–76.1)
≤60	27	(14.8)
61–70	57	(31.3)
71–80	71	(39.0)
>80	27	(14.8)
No. race (%):
Caucasian	156	(85.7)
African American/Other	26	(14.3)
No. Charlson comorbidity index (%):
0 to 2	45	(24.7)
3 to 5	30	(16.5)
>5	107	(58.8)
No. clinical state (%):
Metastatic hormone-sensitive	55	(30.2)
Nonmetastatic castrate-resistant	40	(22.0)
Metastatic castrate-resistant	87	(47.8)
No. kg/m2 BMI (%):
Normal (<25)	34	(18.6)
Overweight (25 to <30)	80	(44.0)
Obese (30 to <35)	44	(24.2)
Morbidly obese (≥35)	24	(13.2)
No. skeletal muscle index (%):
Sarcopenic*	97	(53.3)
Nonsarcopenic	85	(46.7)
Skeletal muscle density (HU):
Myosteatotic*	108	(59.3)
Nonmyosteatotic	74	(40.7)
Median mos followup (IQR):†	33.9	(20.4–55.2)
Alive	87	(47.8)
Deceased	95	(52.2)

^*Cut point for sarcopenia: SMI <43 cm²/m² for BMI <25 and <53 cm²/m² for BMI ≥25); cut point for myosteatosis: SMD <41 HU for BMI <25 and <33 HU for BMI ≥25).

^†Defined as date of death for deceased patients and date of last documented encounter prior to October 2019 in electronic medical record for surviving patients.

Association of Prostate Cancer Characteristics with Survival

A total of 87 (47.8%) men were diagnosed with mCRPC, 55 (30.2%) with mHSPC and 40 (22%) with nmCRPC. Men with mCRPC had a median OS of 35.6 months. In comparison, men with mHSPC had a median OS of 57.0 months and those with nmCRPC had a median OS of 63.5 months. At initial pathological diagnosis of localized disease 101 (55.4%) men had Gleason score 8 or greater, 54 (29.7%) had Gleason 7 and 5 (2.7%) had Gleason 6. Both clinical stage of mPC/CRPC (χ²[2]=9.67, p=0.008) and higher Gleason score at initial prostate cancer diagnosis (χ²[4]=12.0, p=0.02) were associated with worse OS (supplementary figure, parts a and b, https://www.jurology.com).

Association of Obesity, Adipose Distribution, Sarcopenia and Myosteatosis with Survival

Of 182 men 34 (18.6%) were underweight or normal weight, 80 (44.0%) were overweight and 68 (37.4%) were obese. OS for obese men (median 61.7±7.81 months) was significantly longer than OS for nonobese men (median 47.6±6.38 months, HR 5.02, p=0.02) (see part a of figure ). On univariate analysis the association between higher BMI categories and longer survival trended toward significance (c2 [3]=7.41, p=0.06), as did the association between higher SAT quartiles and longer survival (c2[3] =7.28, p=0.06). Univariate analysis did not show a significant association between higher VAT quartile (c2[3]=3.44, p=0.33) or VAT/SAT ratio (c2[3] =2.39, p=0.50) and longer survival.

Figure.

Kaplan-Meier analysis of overall survival for obesity (HR 5.02, p = 0.025) (a), sarcopenia (HR 0.628, p = 0.43) (b) and myosteatosis (HR 7.18, p = 0.007) (c).

In this cohort 97 (53.3%) met SMI criteria for sarcopenia, 108 (59.3%) met SMD criteria for myosteatosis and 69 (37.9%) met criteria for sarcopenic obesity. Median survival for men with sarcopenia (50.6±6.13 months) was not significantly different from those without sarcopenia (55.5±5.81 months; HR 0.628, p=0.43) (see part b of figure). However, men with myosteatosis had reduced overall survival (42.1±7.09 months) compared to those without myosteatosis (61.7±3.96; HR 7.18, p=0.007) (see part c of figure).

Overall Associations with Survival

Multivariable Cox proportional hazard modeling for the relationship between BMI, VAT index and SAT index with survival showed that the most parsimonious model for predicting survival included age, BMI, SMD, SMI, Charlson comorbidity index, race and clinical stage at presentation (table 2). In this modeling we found that increased age was associated with increased mortality (HR 1.04, p=0.01). Increased BMI (HR 0.93, p=0.02) and less advanced clinical stage (HR 0.54, p=0.04) were significantly associated with reduced mortality. Lower SMD (indicating myosteatosis) trended toward reduced survival (HR 0.97, p=0.06). SMI, Charlson comorbidity index, clinical stage at presentation and race were not significant predictors of OS.

Table 2.

Multivariable analysis using Cox proportional hazards for mortality in 182 males with advanced prostate cancer

	HR	95% CI	p Value
Body mass index	0.93	0.88–0.99	0.02
Skeletal muscle index (cm2/m2)	1.02	0.99–1.06	0.19
Skeletal muscle density (HU)	0.97	0.94–1.00	0.08
Age (yrs)	1.03	1.00–1.06	0.04
Charlson comorbidity index	1.05	0.98–1.13	0.20
Race (Caucasian vs Other)	1.31	0.69–2.49	0.41
mHSPC vs nmCRPC	0.55	0.33–0.91	0.02
mHSPC vs mCRPC	0.54	0.30–0.98	0.04

Additional multivariate models using VAT index, SAT index and VAT/SAT in place of BMI showed that VAT index was associated with reduced mortality (HR 0.99, p=0.003). SAT index trended toward a significant association with reduced mortality (HR 0.99, p=0.06), but the ratio of VAT/SAT did not (HR 0.65, p=0.21) (supplementary table, https://www.jurology.com). To further evaluate the relationship between BMI and increased survival, obese men with and without sarcopenia or myosteatosis were compared with nonobese men with and without sarcopenia and myosteatosis (table 3). There were statistically significant differences between groups (F[3,91]=4.03, p=0.01) indicating high BMI (with or without sarcopenia or myosteatosis) was associated with longer survival than sarcopenia or myosteatosis without obesity.

Table 3.

ANOVA of mean overall survival by morphometric status in 95 patients with prostate cancer who were deceased at last followup

	No. Pts	Mean yrs	SE	95% CI
No obesity, sarcopenia or myosteatosis*	2	1.72	0.76	−7.92–11.35
Obesity alone	14	3.69	0.43	2.76–4.62
Sarcopenia + myosteatosis without obesity	20	1.81	0.29	1.19–2.43
Sarcopenia + myosteatosis with obesity	59	2.89	0.23	2.43–3.35

^*Obesity defined as BMI ≥30; sarcopenia defined by skeletal muscle index; myosteatosis defined by skeletal muscle density (HU).

DISCUSSION

The novel finding of this study is that men with mPC/CRPC and high BMI had significantly longer overall survival independent of low muscle mass (sarcopenia) or low muscle radiodensity (myosteatosis). This finding suggests that calculating BMI remains a useful tool in evaluating mPC/CRPC prognosis. Previous studies have suggested that the association between BMI and longer overall survival may be attributed to higher muscle mass, as BMI does not delineate between types of tissue.²¹ When controlling for the presence of sarcopenia or myosteatosis, high BMI was still associated with longer survival, suggesting that obesity may be protective against mortality in mPC/CRPC.

These data align with prior studies suggesting an inverse association between high BMI and mortality across the spectrum of mPC/CRPC.^5–7 This association may be secondary to increased conversion of testosterone to estrogen, as estrogen may suppress growth of castrate-resistant cancer.²² Patients with high BMI may also have higher caloric reserve (storage of fat as an energy source), which could be protective against the catabolism typically occurring in advanced cancer stages.⁷ An association between caloric intake in men with lower BMI, tumor related growth factor IGF-1 and fatal prostate cancer has been reported,²³ suggesting certain metabolic profiles may favor tumor growth over adiposity.

However, further examination of the role of obesity in this patient cohort, by assessing the type of fat stored, showed that having high VAT significantly associated with reduced mortality. These results contrast with a prior retrospective study showing high VAT in 63 patients with mCRPC treated with docetaxel shortened survival.¹² In the present study, patients with HSPC were included, as well as other treatment modalities. While we did not find that the amount of SAT or the VAT/SAT ratio were significantly associated with mortality, a prior study in mCRPC showed worse prognosis for higher VAT/SAT ratio.^9,13 Overall, these data indicate that the interplay between high BMI and the type and distribution of adipose tissue require further examination in patients with prostate cancer. Beyond calculating BMI, measuring waist circumference in routine clinical practice as a proxy for VAT may be a valuable method of determining long-term risk in patients with prostate cancer.

There are limited data regarding the role of sarcopenia in mPC/CRPC. However, it is surprising that sarcopenia was not associated with reduced survival in the present cohort. Sarcopenia is associated with reduced OS in bladder, kidney and upper tract urothelial cancers, as well as reduced cancer-specific survival in bladder cancer.²⁴ Furthermore, high protein oral liquid nutrition supplementation in patients with bladder cancer treated with radical cystectomy reduced the prevalence of sarcopenia and may have reduced postoperative complications and readmissions.²⁵ Importantly, the development of sarcopenia, which would promote decline in muscle strength, power and function, can limit the duration of ADT treatment.²⁶ The presence of sarcopenia has been associated with chemotherapy toxicity and shorter OS in men with mCRPC treated with docetaxel,¹² although the finding of shorter OS may have been a consequence of inadequate treatment, as studies conducted with post-chemotherapy patients with mCRPC showed no effect on OS.²⁷

Another interesting finding was that low skeletal muscle density (myosteatosis) was associated with shorter OS. However, multivariable analysis adjusting for age, BMI or VAT, and clinical stage showed only a trend toward significance for the relationship between myosteatosis and shorter OS. Although less frequently studied, myosteatosis is as prevalent as sarcopenia in patients with advanced cancer across all BMI categories.²⁸ In overweight and obese patients with head/neck cancer, myosteatosis was present at different levels of nutrition risk and independently predicted reduced survival.²⁹ Further, myosteatosis was associated with a 75% increase in mortality risk in lymphomas, gynecologic, renal and gastrointestinal cancers.¹⁶ The mechanism behind this reduced survival is not well understood. One hypothesis is that ectopic fat in skeletal muscle impedes blood flow to muscle while secreting adipokines and cytokines that promote inflammation and increased catabolism from cancer cachexia.¹⁶ In addition, myosteatosis is more robustly associated with physical function impairments and frailty than sarcopenia.¹⁷ This relationship may explain why myosteatosis is associated with worse prognosis in this cohort while sarcopenia is not. The specific role of myosteatosis and effect on outcomes in mPC/CRPC merits further study.

There are several strengths to this study. All men presenting to the Comprehensive Prostate Cancer Clinic were included. This cohort represents men with mHSPC, nmCRPC and mCRPC followed longitudinally with a median followup of nearly 3 years. As far as we know, this study is the most comprehensive investigation of relationships between BMI, VAT, SAT, sarcopenia, myosteatosis and mPC/CRPC to date. Further, a rigorously validated CT protocol was employed to measure body composition and used to quantify total abdominal musculature, not simply calculate the psoas area, to identify patients with sarcopenia. Limitations of utilizing psoas muscle alone include poor association with whole body skeletal muscle area, lack of standardization in how the psoas muscle index is measured, and that the psoas muscle index can be disproportionally affected by localized pathology.³⁰

Nevertheless, limitations of the study must be considered. First, is the selection bias inherent to a retrospective study derived from a single academic medical center. In addition, all subjects were from the Southeast region of the United States, thus there were few normal weight men. The findings of this study indicate a larger sample size may be necessary to determine the specific roles of BMI and morphometric features, and whether high SAT and/or myosteatosis are independently associated with shorter OS.

CONCLUSION

In men with mPC/CRPC high BMI and high VAT were significantly associated with longer OS. In contrast, the presence of myosteatosis may be associated with shorter OS. Careful measurement of body weight and height to accurately calculate BMI and measurement of waist circumference, as a proxy of VAT, are indicated in routine clinical workup to determine potential prognosis. Furthermore, incorporating morphometric analysis of CT imaging can be used to identify patients at high risk for poor prognosis.

Abbreviations and Acronyms

BMI

body mass index

CRPC

castrate-resistant prostate cancer

CT

computerized tomography

HR

hazard ratio

mCRPC

metastatic castrate-resistant prostate cancer

mHSPC

metastatic hormone sensitive prostate cancer

mPC/CRPC

metastatic prostate cancer/castrate resistant prostate cancer

nmCRPC

nonmetastatic castrate resistant prostate cancer

OS

overall survival

SAT

subcutaneous adipose tissue

SMD

skeletal muscle density

SMI

skeletal muscle index

VAT

visceral adipose tissue