Obesity Does Not Affect Survival Outcomes in Extremity Soft Tissue Sarcoma
Abstract
Background
Obesity is a growing epidemic and has been associated with an increased frequency of complications after various surgical procedures. Studies also have shown adipose tissue to promote a microenvironment favorable for tumor growth. Additionally, the relationship between obesity and prognosis of soft tissue sarcomas has yet to be evaluated.
Questions/purposes
We sought to assess if (1) obesity affects survival outcomes (local recurrence, distant metastasis, and death attributable to disease) in patients with extremity soft tissue sarcomas; and (2) whether obesity affected wound healing and other surgical complications after treatment.
Methods
A BMI of 30 kg/m or greater was used to define obesity. Querying our prospective database between 2001 and 2008, we identified 397 patients for the study; 154 were obese and 243 were not obese. Mean followup was 4.5 years (SD, 3.1 years) in the obese group and 3.9 years (SD, 3.2 years) in the nonobese group; the group with a BMI of 30 kg/m or greater had a higher proportion of patients with followups of at least 2 years compared with the group with a BMI less than 30 kg/m (76% versus 62%). Outcomes, including local recurrence, distant metastasis, and overall survival, were analyzed after patients were stratified by BMI. Multivariable survival models were used to identify independent predictors of survival outcomes. Wilcoxon rank sum test was used to compare continuous variables. Based on the accrual interval of 8 years, the additional followup of 5 years after data collection, and the median survival time for the patients with a BMI less than 30 kg/m of 3 years, we were able to detect true median survival times in the patients with a BMI of 30 kg/m of 2.2 years or less with 80% power and type I error rate of 0.05.
Results
Patients who were obese had similar survival outcomes and wound complication rates when compared with their nonobese counterparts. Patients who were obese were more likely to have lower-grade tumors (31% versus 20%; p = 0.021) and additional comorbidities including diabetes mellitus (26% versus 7%; p < 0.001), hypertension (63% versus 38%; p < 0.001), and smoking (49% versus 37%; p = 0.027). Regression analysis confirmed that even after accounting for certain tumor characteristics and comorbidities, obesity did not serve as an independent risk factor in affecting survival outcomes.
Conclusions
Although the prevalence of obesity continues to increase and lead to many negative health consequences, it does not appear to adversely affect survival, local recurrence, or wound complication rates for patients with extremity soft tissue sarcomas.
Level of Evidence
Level III, therapeutic study. See the Instructions for Authors for a complete description of levels of evidence.
Introduction
The prevalence of obesity has continued to increase during the previous four decades and currently is estimated to affect 32.2% of adult men and 35.5% of adult women in the United States [11]. In addition to being associated with various negative health consequences including increased incidence of cardiovascular disease, obesity has been shown to have an increased association with surgical complications. For example, in the case of total joint arthroplasties, some studies have reported increases in the postoperative complications such as acute renal insufficiency and increased incidences of pulmonary emboli [6, 27]. Additionally, patients who were obese had increased incidences of wound breakdown and infection and required readmissions for wound lavage and debridement [7, 17]. Obesity also has been correlated to increase the risk of soft tissue sarcomas [30]. Studies have revealed a link between a high fat diet and the formation of spontaneous liposarcomas through intermediaries such as overexpression of interleukin-22 [31].
In addition to increasing postoperative complications, obesity has been shown to increase the complexity of the surgical procedure. General surgical patients with obesity experience increased operative time, greater blood loss, and, if surgery is done laparoscopically, greater conversion to open procedures [20, 25]. Investigations into the link between obesity and tumor growth have become an active area of research and preliminary studies have shown an association between obesity and cancer incidence and worse prognosis in certain cancer types such as breast and colon cancer, but not in soft tissue sarcomas to our knowledge [3, 33].
In this study, we sought to assess the role that obesity plays in affecting prognosis of soft tissue sarcomas of the extremity. Specifically, we asked if (1) obesity affects survival outcomes (local recurrence, distant metastasis, and death attributable to disease) in patients with extremity soft tissue sarcomas; and (2) whether obesity affects wound healing and other surgical complications after treatment.
Patients and Methods
After obtaining institutional review board approval, a retrospective cohort study of a prospective database at a major sarcoma center was conducted to assess the role of obesity in affecting the survival outcomes in patients with extremity soft tissue sarcomas. Patients undergoing surgical resection of soft tissue sarcomas at our institution between January 2001 and December 2008 were identified based on a retrospective review and were considered for the study. This period was chosen to achieve 5 years followup. Patients were excluded from the study if they were younger than 18 years, if they lacked adequate medical records, and if they had a tumor with good prognosis and borderline malignancy (such tumors included desmoid tumors and dermatofibrosarcoma protuberans) [4]. A total of 397 patients met criteria for inclusion in this study.
Patients were stratified into two groups based on their BMI; if their BMI was 30 kg/m or greater, they were classified as obese. Patients with a BMI less than 30 kg/m were classified as not being obese. A total of 397 patients were included in this study. Of these, 154 patients were classified as being obese (BMI ≥ 30 kg/m) and the remaining 243 patients were classified as not being obese (BMI < 30 kg/m). The prevalence of obesity among adults in Tennessee in 2010 was 31.7% [13]. Mean followup was 4.5 years (SD, 3.1 years) in the obese group and 3.9 years (SD, 3.2 years) in the nonobese group; the group with a BMI of 30 kg/m or greater had a larger proportion of patients with followup of at least 2 years compared with patients with a BMI less than 30 kg/m (76% versus 62%; p = 0.003). No patients were excluded owing to followup limitations.
Patient demographics and tumor characteristics collected included age at the time of surgery, sex, and race. Tumor characteristics consisted of size, depth (superficial or deep to the fascia of the underlying muscle), site (upper or lower extremity), grade (low, intermediate, or high), and histologic subtype. Staging of the patients also was performed per the guidelines recommended by the American Joint Committee on Cancer (AJCC) [8]. Margins either were recorded as positive or negative for each patient after definitive surgery. A positive margin was defined as the presence of malignant cells at the inked margin. Additionally, the excision status was noted for each patient; that is, if a patient presented for resection of a soft tissue sarcoma before excision, they were categorized as having a primary excision; if a patient presented after an unplanned excision elsewhere, the patient was categorized as having undergone a secondary excision. Additionally, whether they received radiotherapy was recorded. Chemotherapy, also noted, was administered to the patient at the discretion of the multidisciplinary oncology team consistent with current standards of care in our institution. Most instances of chemotherapy consisted of anthracycline-based regimens and were reserved for patients with Stage IV carcinoma with metastatic disease. Medical comorbidities such as the presence of hyperlipidemia, hypertension, diabetes mellitus, and chronic obstructive pulmonary disease (COPD) also were abstracted from the preoperative anesthesia clearance done before definitive resection. Smoking and alcohol abuse also were noted. After surgery, if patients had any wound complications develop (defined as hospitalization within 6 months postoperatively for a wound problem, wounds requiring irrigation and débridement, or infections treated with antibiotics on an outpatient basis), these incidences were recorded through a retrospective review of the medical record. Disease status, death resulting from soft tissue sarcoma, distant metastasis, and local recurrence were recorded. The time from the index surgery to each outcome listed was measured for all patients.
Patient demographics, tumor characteristics, and prognostic outcomes were compared across groups using Wilcoxon rank sum tests for continuous variables and chi-square or Fisher’s exact tests. Survival curves for disease-free survival and metastasis-free survival were calculated and presented using the Kaplan-Meier method [19].
The primary endpoint of the study was designated as death resulting from sarcoma for the disease-free survival curve. Death was treated as a censored observation for patients who died from a cause not directly related to their soft tissue sarcoma. Such censored deaths included patients who died of comorbidities during no-evidence-of-disease status. If patients had active disease at time of their death, their death was not treated as censored. Gray’s test was calculated and used to compare the disease-specific hazards of death [14, 15]. Log-rank test was used to compare the hazard of distant metastasis and local recurrence between the two groups (BMI ≥ 30 kg/m versus BMI < 30 kg/m). In addition to univariate comparisons, multivariable regression analyses were used to take into account potential confounders [10]. This hazard ratio model examined obesity as the main independent predictor of survival outcomes after controlling for certain tumor characteristics, demographics, and other medical comorbidities that such patients typically have.
Statistical Methods
Demographic and clinical variables including risk factors were summarized and compared for patients in each category as mentioned. Descriptive summaries of continuous variables were presented in terms of interquartile range, whereas discrete variables were summarized in terms of frequencies and percentage. Wilcoxon rank sum test and chi-square test were used to do statistical comparisons. Survival curves for disease-free survival were calculated using competing risk analyses. The primary endpoint of the study was designated as death resulting from sarcoma. Death was treated as a competing risk for patients who died from a cause not directly related to their soft tissue sarcoma. Gray’s test was calculated and used to compare the disease-specific cumulative incidence of death [14, 15]. In addition to univariate comparisons, multivariable regression analyses were used to take into account potential confounders such as tumor size, grade, and other comorbidities [10]. These models examined pertinent variables as the main independent predictor of survival outcomes after controlling for potentially confounding variables. Statistical software R (Version 1.15.1; R Foundation for Statistical Computing, Vienna, Austria) was used for all data analysis. Reported p values were two-sided and a p value less than 0.05 was considered significant.
Statistical Power
A post hoc power analysis was performed. Based on the accrual interval of 8 years, the additional followup of 5 years after data collection, and the median survival time for the group with a BMI less than 30 kg/m of 3 years, we were able to detect true median survival times for the patients with a BMI of 30 kg/m or greater of 2.2 years or less with 80% power and type I error rate of 0.05.
Statistical Methods
Demographic and clinical variables including risk factors were summarized and compared for patients in each category as mentioned. Descriptive summaries of continuous variables were presented in terms of interquartile range, whereas discrete variables were summarized in terms of frequencies and percentage. Wilcoxon rank sum test and chi-square test were used to do statistical comparisons. Survival curves for disease-free survival were calculated using competing risk analyses. The primary endpoint of the study was designated as death resulting from sarcoma. Death was treated as a competing risk for patients who died from a cause not directly related to their soft tissue sarcoma. Gray’s test was calculated and used to compare the disease-specific cumulative incidence of death [14, 15]. In addition to univariate comparisons, multivariable regression analyses were used to take into account potential confounders such as tumor size, grade, and other comorbidities [10]. These models examined pertinent variables as the main independent predictor of survival outcomes after controlling for potentially confounding variables. Statistical software R (Version 1.15.1; R Foundation for Statistical Computing, Vienna, Austria) was used for all data analysis. Reported p values were two-sided and a p value less than 0.05 was considered significant.
Statistical Power
A post hoc power analysis was performed. Based on the accrual interval of 8 years, the additional followup of 5 years after data collection, and the median survival time for the group with a BMI less than 30 kg/m of 3 years, we were able to detect true median survival times for the patients with a BMI of 30 kg/m or greater of 2.2 years or less with 80% power and type I error rate of 0.05.
Results
Survival Outcomes
There were no differences between patients who were obese or nonobese in terms of the proportion who experienced local recurrence (10% versus 16%; p = 0.14), distant metastasis (28% versus 30%; p = 0.59), and death secondary to their sarcoma (23% versus 24%; p = 0.55) (Table 1). Additionally, the rate to local recurrence (p = 0.07, log-rank test) (Fig. 1), distant metastasis (p = 0.26, log-rank test) (Fig. 2) and death attributable to sarcoma (p = 0.75, Gray’s test) (Fig. 3) were no different between patients who were obese or nonobese. After controlling for factors such as patient age, sex, tumor size, grade, medical comorbidities and wound complications, we were able to determine which factors were independent predictors of survival. Obesity was not a predictor of sarcoma-specific death (p = 0.22), distant metastasis (p = 0.395), or local relapse (p = 0.13) (Table 2).
Table 1
Survival outcomes
| Variable | BMI ≥ 30 kg/m (N = 154) | BMI < 30 kg/m (N = 243) | p value |
|---|---|---|---|
| Local recurrence | 0.14 | ||
| Yes | 16 (10%) | 38 (16%) | |
| No | 138 (90%) | 205 (84%) | |
| Distant metastasis | 0.59 | ||
| Yes | 43 (28%) | 74 (30%) | |
| No | 111 (72%) | 169 (70%) | |
| Survival status | 0.55 | ||
| Alive | 104 (69%) | 152 (65%) | |
| Died of sarcoma | 34 (23%) | 55 (24%) | |
| Died of other causes | 12 (8%) | 26 (11%) | |
Local recurrence-free survival curves are shown for patients who were obese versus patients who were nonobese.
Disease-free survival curves are shown for patients who were obese versus patients who were nonobese.
Table 2
Multivariable regression analysis of survival, distant metastasis, and local recurrence
| Variable | Sarcoma-specific death | Distant metastasis | Local recurrence | |||
|---|---|---|---|---|---|---|
| p value | Hazard ratio (2.5%, 97.5%) | p value | Hazard ratio (2.5%, 97.5%) | p value | Hazard ratio (2.5%, 97.5%) | |
| Obesity | 0.2200 | 1.40 (0.82–2.40) | 0.3952 | 0.82 (0.53–1.29) | 0.1300 | 0.60 (0.31–1.16) |
| Age | 0.0150 | 1.02 (1.00, 1.04) | 0.3978 | 0.99 (0.98–1.01) | 0.1538 | 1.01 (0.99–1.03) |
| Female sex (reference: male sex) | 0.0990 | 0.66 (0.40–1.08) | 0.0189 | 0.61 (0.40–0.92) | 0.4290 | 0.78 (0.42–1.45) |
| Tumor size(per 1-cm increase) | < 0.0001 | 1.07 (1.04–1.10) | < 0.0001 | 1.07 (1.04–1.10) | 0.7244 | 1.01 (0.97–1.05) |
| Tumor grade | ||||||
| Grade II (reference: Grade I) | 0.0024 | 7.90 (2.08–30.05) | < 0.0001 | 4.53 (1.56–13.11) | 0.1556 | 2.40 (0.85–6.75) |
| Grade III (reference: Grade I) | < 0.0001 | 12.47 (4.02–38.70) | < 0.0001 | 10.67 (4.59–24.83) | 0.1556 | 2.15 (0.95–4.87) |
| Margin status (positive; reference: negative margins) | 0.9300 | 1.03 (0.50–2.12) | 0.0880 | 1.65 (0.93–2.93) | < 0.0001 | 4.45 (2.29–8.63) |
| Smoking | 0.5900 | 0.88 (0.55–1.41) | 0.9323 | 0.98 (0.65–1.49) | 0.5451 | 0.82 (0.43–1.56) |
| Alcohol abuse | 0.1800 | 0.21 (0.02–2.09) | 0.2495 | 0.30 (0.04–2.31) | 0.1299 | 3.56 (0.69–18.43) |
| Diabetes mellitus | 0.4500 | 0.78 (0.40–1.50) | 0.7166 | 0.90 (0.50–1.60) | 0.6329 | 1.21 (0.56–2.59) |
| Hyperlipidemia | 0.5400 | 1.21 (0.65–2.26) | 0.6696 | 1.12 (0.67–1.87) | 0.1266 | 1.66 (0.87–3.17) |
| Hypertension | 0.2300 | 0.71 (0.41–1.24) | 0.9087 | 0.97 (0.61–1.56) | 0.8587 | 0.94 (0.49–1.82) |
| Chronic obstructive pulmonary disease | 0.8300 | 0.91 (0.40–2.10) | 0.3542 | 0.60 (0.21–1.76) | 0.5973 | 1.44 (0.37–5.60) |
| Wound complications | 0.3100 | 1.28 (0.79–2.08) | 0.6429 | 1.11 (0.72–1.69) | 0.2922 | 1.42 (0.74–2.71) |
Other Predictors of Survival Outcomes
After controlling for potentially confounding variables, as described above, we found that increasing age (p = 0.015), tumor size (p < 0.0001), and tumor grade (p < 0.0001) all increased the hazard of death attributable to sarcoma (Table 2). Sex, positive margins, and all comorbidities tested including smoking, alcohol abuse, diabetes mellitus, hyperlipidemia, hypertension, COPD, and wound complications failed to have a significant effect on sarcoma-specific death. In analyzing distant metastasis, male sex (p = 0.0189), increasing tumor size (p < 0.0001), and tumor grade (p < 0.0001) increased the hazard of distant metastasis. Age, margin status, and all comorbidities tested failed to have a significant effect on distant metastasis. In analyzing local recurrence, we found that the only factor that increased the hazard of local recurrence was a positive margin after definitive resection. Positive margins increased the hazard of local recurrence by a factor of 4.45 (p < 0.0001).
No differences were found in race (p = 0.35), sex (p = 0.52), tumor site (p = 0.84), stage (p = 0.18), or histology subtype (p = 0.72) between patients who were obese or nonobese (Table 3). Excision status (p = 0.70) and differences in microscopic margin status (p = 0.42) between patients who were obese or nonobese also were insignificant. Receipt of radiotherapy and chemotherapy were not different between the two groups (p = 0.84 and p = 0.60 respectively). Important differences were found in age (patients who were obese were younger compared with patients who were not nonobese, 54 years versus 60 years; p = 0.01) and tumor grade (31% of patients who were obese had a Grade 1 tumor, whereas only 20% of patients who were nonobese had a Grade 1 tumor; p = 0.02).
Table 3
Patient demographics and tumor characteristics
| Variable | BMI ≥ 30 kg/m (N = 154) | BMI < 30 kg/m (N = 243) | p value |
|---|---|---|---|
| Age, years, median (interquartile range) | 54 (42–66) | 60 (44–72) | 0.012 |
| Sex | 0.52 | ||
| Male | 85 (55%) | 126 (52%) | |
| Female | 69 (45%) | 117 (48%) | |
| Race | 0.35 | ||
| Caucasian | 132 (86%) | 217 (89%) | |
| African-American | 14 (9%) | 13 (5%) | |
| Others | 8 (5%) | 13 (5%) | |
| Excision status | 0.70 | ||
| Primary excision | 94 (61%) | 153 (63%) | |
| Secondary excision | 60 (39%) | 90 (37%) | |
| Site | 0.84 | ||
| Upper extremity | 41 (27%) | 67 (28%) | |
| Lower extremity | 113 (73%) | 176 (72%) | |
| Size, cm, median (interquartile range) | 10 (5–17) | 9 (5–15) | 0.29 |
| Depth | 0.93 | ||
| Superficial | 26 (17%) | 40 (17%) | |
| Deep | 128 (83%) | 202 (83%) | |
| Grade | 0.021 | ||
| I | 48 (31%) | 48 (20%) | |
| II | 21 (14%) | 30 (12%) | |
| III | 84 (55%) | 164 (68%) | |
| AJCC stage | 0.18 | ||
| I | 57 (37%) | 61 (25%) | |
| II | 27 (18%) | 47 (20%) | |
| III | 29 (19%) | 57 (24%) | |
| IV | 41 (27%) | 76 (32%) | |
| Histology type | 0.72 | ||
| Liposarcoma | 41(27%) | 51 (21%) | |
| Malignant fibrous histiocytoma | 56 (36%) | 93 (38%) | |
| Leiomyosarcoma | 13 (8%) | 25 (10%) | |
| Fibrosarcoma | 9 (6%) | 15 (6%) | |
| Others | 35 (22%) | 59 (24%) | |
| Microscopic margins | 0.42 | ||
| Negative | 139 (90%) | 212 (88%) | |
| Positive | 15 (15%) | 30 (12%) | |
| Radiotherapy | 0.84 | ||
| Preoperative | 20 (13%) | 29 (12%) | |
| Postoperative | 78 (51%) | 117 (48%) | |
| Preoperative and postoperative | 18 (12%) | 27 (11%) | |
| No | 38 (25%) | 70 (29%) | |
| Chemotherapy | 0.60 | ||
| Yes | 32 (21%) | 45 (19%) | |
| No | 121 (79%) | 195 (81%) | |
AJCC = American Joint Committee on Cancer.
No differences were found in the proportion of patients with hyperlipidemia (p = 0.21), alcohol abuse (p = 0.52), or wound complications (p = 0.25) between patients who were obese or nonobese (Table 4). However, patients who were obese had a greater proportion of diabetes and hypertension. Twenty-six percent of patients who were obese also had diabetes, whereas only 7% of patients who were nonobese had diabetes (p < 0.001). Similarly, 63% of patients who were obese had hypertension, whereas only 38% of the patients who were nonobese had hypertension (p < 0.001). There was a greater prevalence of smoking among patients who were obese (49% among patients who were obese versus 37% among patients who were nonobese; p = 0.027). In analyzing COPD, patients who were nonobese had a greater prevalence of the disease compared with patients who were obese (5% versus 1%; p = 0.038).
Table 4
Obesity and other comorbid conditions
| Variable | BMI ≥ 30 (N = 154) | BMI < 30 (N = 243) | p value |
|---|---|---|---|
| Smoking | 0.027 | ||
| Yes | 75 (49%) | 91 (37%) | |
| No | 79 (51%) | 152 (63%) | |
| Alcohol abuse | 0.52 | ||
| Yes | 4 (3%) | 4 (2%) | |
| No | 144 (9%) | 227 (98%) | |
| Diabetes mellitus | < 0.001 | ||
| Yes | 40 (26%) | 17 (7%) | |
| No | 114 (74%) | 226 (93%) | |
| Hyperlipidemia | 0.21 | ||
| Yes | 38 (25%) | 47 (19%) | |
| No | 116 (75%) | 196 (81%) | |
| Hypertension | < 0.001 | ||
| Yes | 96 (63%) | 92 (38%) | |
| No | 57 (37%) | 151 (62%) | |
| Chronic obstructive pulmonary disease | 0.038 | ||
| Yes | 2 (1%) | 12 (5%) | |
| No | 152 (99%) | 229 (95%) | |
| Wound complications | 0.25 | ||
| Yes | 42 (27%) | 54 (22%) | |
| No | 112 (73%) | 189 (78%) | |
Survival Outcomes
There were no differences between patients who were obese or nonobese in terms of the proportion who experienced local recurrence (10% versus 16%; p = 0.14), distant metastasis (28% versus 30%; p = 0.59), and death secondary to their sarcoma (23% versus 24%; p = 0.55) (Table 1). Additionally, the rate to local recurrence (p = 0.07, log-rank test) (Fig. 1), distant metastasis (p = 0.26, log-rank test) (Fig. 2) and death attributable to sarcoma (p = 0.75, Gray’s test) (Fig. 3) were no different between patients who were obese or nonobese. After controlling for factors such as patient age, sex, tumor size, grade, medical comorbidities and wound complications, we were able to determine which factors were independent predictors of survival. Obesity was not a predictor of sarcoma-specific death (p = 0.22), distant metastasis (p = 0.395), or local relapse (p = 0.13) (Table 2).
Table 1
Survival outcomes
| Variable | BMI ≥ 30 kg/m (N = 154) | BMI < 30 kg/m (N = 243) | p value |
|---|---|---|---|
| Local recurrence | 0.14 | ||
| Yes | 16 (10%) | 38 (16%) | |
| No | 138 (90%) | 205 (84%) | |
| Distant metastasis | 0.59 | ||
| Yes | 43 (28%) | 74 (30%) | |
| No | 111 (72%) | 169 (70%) | |
| Survival status | 0.55 | ||
| Alive | 104 (69%) | 152 (65%) | |
| Died of sarcoma | 34 (23%) | 55 (24%) | |
| Died of other causes | 12 (8%) | 26 (11%) | |
Local recurrence-free survival curves are shown for patients who were obese versus patients who were nonobese.
Metastasis-free survival curves are shown for patients who were obese versus nonobese.
Disease-free survival curves are shown for patients who were obese versus patients who were nonobese.
Table 2
Multivariable regression analysis of survival, distant metastasis, and local recurrence
| Variable | Sarcoma-specific death | Distant metastasis | Local recurrence | |||
|---|---|---|---|---|---|---|
| p value | Hazard ratio (2.5%, 97.5%) | p value | Hazard ratio (2.5%, 97.5%) | p value | Hazard ratio (2.5%, 97.5%) | |
| Obesity | 0.2200 | 1.40 (0.82–2.40) | 0.3952 | 0.82 (0.53–1.29) | 0.1300 | 0.60 (0.31–1.16) |
| Age | 0.0150 | 1.02 (1.00, 1.04) | 0.3978 | 0.99 (0.98–1.01) | 0.1538 | 1.01 (0.99–1.03) |
| Female sex (reference: male sex) | 0.0990 | 0.66 (0.40–1.08) | 0.0189 | 0.61 (0.40–0.92) | 0.4290 | 0.78 (0.42–1.45) |
| Tumor size(per 1-cm increase) | < 0.0001 | 1.07 (1.04–1.10) | < 0.0001 | 1.07 (1.04–1.10) | 0.7244 | 1.01 (0.97–1.05) |
| Tumor grade | ||||||
| Grade II (reference: Grade I) | 0.0024 | 7.90 (2.08–30.05) | < 0.0001 | 4.53 (1.56–13.11) | 0.1556 | 2.40 (0.85–6.75) |
| Grade III (reference: Grade I) | < 0.0001 | 12.47 (4.02–38.70) | < 0.0001 | 10.67 (4.59–24.83) | 0.1556 | 2.15 (0.95–4.87) |
| Margin status (positive; reference: negative margins) | 0.9300 | 1.03 (0.50–2.12) | 0.0880 | 1.65 (0.93–2.93) | < 0.0001 | 4.45 (2.29–8.63) |
| Smoking | 0.5900 | 0.88 (0.55–1.41) | 0.9323 | 0.98 (0.65–1.49) | 0.5451 | 0.82 (0.43–1.56) |
| Alcohol abuse | 0.1800 | 0.21 (0.02–2.09) | 0.2495 | 0.30 (0.04–2.31) | 0.1299 | 3.56 (0.69–18.43) |
| Diabetes mellitus | 0.4500 | 0.78 (0.40–1.50) | 0.7166 | 0.90 (0.50–1.60) | 0.6329 | 1.21 (0.56–2.59) |
| Hyperlipidemia | 0.5400 | 1.21 (0.65–2.26) | 0.6696 | 1.12 (0.67–1.87) | 0.1266 | 1.66 (0.87–3.17) |
| Hypertension | 0.2300 | 0.71 (0.41–1.24) | 0.9087 | 0.97 (0.61–1.56) | 0.8587 | 0.94 (0.49–1.82) |
| Chronic obstructive pulmonary disease | 0.8300 | 0.91 (0.40–2.10) | 0.3542 | 0.60 (0.21–1.76) | 0.5973 | 1.44 (0.37–5.60) |
| Wound complications | 0.3100 | 1.28 (0.79–2.08) | 0.6429 | 1.11 (0.72–1.69) | 0.2922 | 1.42 (0.74–2.71) |
Other Predictors of Survival Outcomes
After controlling for potentially confounding variables, as described above, we found that increasing age (p = 0.015), tumor size (p < 0.0001), and tumor grade (p < 0.0001) all increased the hazard of death attributable to sarcoma (Table 2). Sex, positive margins, and all comorbidities tested including smoking, alcohol abuse, diabetes mellitus, hyperlipidemia, hypertension, COPD, and wound complications failed to have a significant effect on sarcoma-specific death. In analyzing distant metastasis, male sex (p = 0.0189), increasing tumor size (p < 0.0001), and tumor grade (p < 0.0001) increased the hazard of distant metastasis. Age, margin status, and all comorbidities tested failed to have a significant effect on distant metastasis. In analyzing local recurrence, we found that the only factor that increased the hazard of local recurrence was a positive margin after definitive resection. Positive margins increased the hazard of local recurrence by a factor of 4.45 (p < 0.0001).
No differences were found in race (p = 0.35), sex (p = 0.52), tumor site (p = 0.84), stage (p = 0.18), or histology subtype (p = 0.72) between patients who were obese or nonobese (Table 3). Excision status (p = 0.70) and differences in microscopic margin status (p = 0.42) between patients who were obese or nonobese also were insignificant. Receipt of radiotherapy and chemotherapy were not different between the two groups (p = 0.84 and p = 0.60 respectively). Important differences were found in age (patients who were obese were younger compared with patients who were not nonobese, 54 years versus 60 years; p = 0.01) and tumor grade (31% of patients who were obese had a Grade 1 tumor, whereas only 20% of patients who were nonobese had a Grade 1 tumor; p = 0.02).
Table 3
Patient demographics and tumor characteristics
| Variable | BMI ≥ 30 kg/m (N = 154) | BMI < 30 kg/m (N = 243) | p value |
|---|---|---|---|
| Age, years, median (interquartile range) | 54 (42–66) | 60 (44–72) | 0.012 |
| Sex | 0.52 | ||
| Male | 85 (55%) | 126 (52%) | |
| Female | 69 (45%) | 117 (48%) | |
| Race | 0.35 | ||
| Caucasian | 132 (86%) | 217 (89%) | |
| African-American | 14 (9%) | 13 (5%) | |
| Others | 8 (5%) | 13 (5%) | |
| Excision status | 0.70 | ||
| Primary excision | 94 (61%) | 153 (63%) | |
| Secondary excision | 60 (39%) | 90 (37%) | |
| Site | 0.84 | ||
| Upper extremity | 41 (27%) | 67 (28%) | |
| Lower extremity | 113 (73%) | 176 (72%) | |
| Size, cm, median (interquartile range) | 10 (5–17) | 9 (5–15) | 0.29 |
| Depth | 0.93 | ||
| Superficial | 26 (17%) | 40 (17%) | |
| Deep | 128 (83%) | 202 (83%) | |
| Grade | 0.021 | ||
| I | 48 (31%) | 48 (20%) | |
| II | 21 (14%) | 30 (12%) | |
| III | 84 (55%) | 164 (68%) | |
| AJCC stage | 0.18 | ||
| I | 57 (37%) | 61 (25%) | |
| II | 27 (18%) | 47 (20%) | |
| III | 29 (19%) | 57 (24%) | |
| IV | 41 (27%) | 76 (32%) | |
| Histology type | 0.72 | ||
| Liposarcoma | 41(27%) | 51 (21%) | |
| Malignant fibrous histiocytoma | 56 (36%) | 93 (38%) | |
| Leiomyosarcoma | 13 (8%) | 25 (10%) | |
| Fibrosarcoma | 9 (6%) | 15 (6%) | |
| Others | 35 (22%) | 59 (24%) | |
| Microscopic margins | 0.42 | ||
| Negative | 139 (90%) | 212 (88%) | |
| Positive | 15 (15%) | 30 (12%) | |
| Radiotherapy | 0.84 | ||
| Preoperative | 20 (13%) | 29 (12%) | |
| Postoperative | 78 (51%) | 117 (48%) | |
| Preoperative and postoperative | 18 (12%) | 27 (11%) | |
| No | 38 (25%) | 70 (29%) | |
| Chemotherapy | 0.60 | ||
| Yes | 32 (21%) | 45 (19%) | |
| No | 121 (79%) | 195 (81%) | |
AJCC = American Joint Committee on Cancer.
No differences were found in the proportion of patients with hyperlipidemia (p = 0.21), alcohol abuse (p = 0.52), or wound complications (p = 0.25) between patients who were obese or nonobese (Table 4). However, patients who were obese had a greater proportion of diabetes and hypertension. Twenty-six percent of patients who were obese also had diabetes, whereas only 7% of patients who were nonobese had diabetes (p < 0.001). Similarly, 63% of patients who were obese had hypertension, whereas only 38% of the patients who were nonobese had hypertension (p < 0.001). There was a greater prevalence of smoking among patients who were obese (49% among patients who were obese versus 37% among patients who were nonobese; p = 0.027). In analyzing COPD, patients who were nonobese had a greater prevalence of the disease compared with patients who were obese (5% versus 1%; p = 0.038).
Table 4
Obesity and other comorbid conditions
| Variable | BMI ≥ 30 (N = 154) | BMI < 30 (N = 243) | p value |
|---|---|---|---|
| Smoking | 0.027 | ||
| Yes | 75 (49%) | 91 (37%) | |
| No | 79 (51%) | 152 (63%) | |
| Alcohol abuse | 0.52 | ||
| Yes | 4 (3%) | 4 (2%) | |
| No | 144 (9%) | 227 (98%) | |
| Diabetes mellitus | < 0.001 | ||
| Yes | 40 (26%) | 17 (7%) | |
| No | 114 (74%) | 226 (93%) | |
| Hyperlipidemia | 0.21 | ||
| Yes | 38 (25%) | 47 (19%) | |
| No | 116 (75%) | 196 (81%) | |
| Hypertension | < 0.001 | ||
| Yes | 96 (63%) | 92 (38%) | |
| No | 57 (37%) | 151 (62%) | |
| Chronic obstructive pulmonary disease | 0.038 | ||
| Yes | 2 (1%) | 12 (5%) | |
| No | 152 (99%) | 229 (95%) | |
| Wound complications | 0.25 | ||
| Yes | 42 (27%) | 54 (22%) | |
| No | 112 (73%) | 189 (78%) | |
Discussion
Obesity has increased in prevalence worldwide [1, 11, 18]. Apart from increasing the risk of cardiovascular disease, stroke, obstructive sleep apnea [21, 32], obesity has been found to increase the complexity and complications from surgical procedures such as total joint arthroplasties [22, 23]. It also has been linked to breast, colon, and endometrial cancers [28]. Adiponcosis is a recently proposed term to describe the link between obesity and cancer [2]. It was reported that adipocyte tissue expansion associated with obesity results in increased infiltration of immune cells and release of associated proinflammatory and growth factors that creates a microenvironment favorable for tumor growth [5]. We sought to assess the relationship between obesity and survival outcomes in patients with soft tissue sarcomas of the extremity as, to our knowledge, the relationship between obesity and survival has not been evaluated in these tumors.
This study had some limitations. Patients in this study were classified as being obese if their BMI was greater than 30 kg/m. BMIs of patients were collected at the time of definitive surgery. However, BMI is a dynamic and not a static variable and has the potential to change with time. Thus, change in weight with time has the possibility to affect outcomes for some patients. BMI also was a surrogate for measuring obesity as the local area of surgery might not always correlate with BMI. Measuring the fat layer at the site of surgery might provide additional information regarding the relationship between survival outcomes and obesity and deserves further study. There was differential followup between our two groups with the obese group having a longer followup. This would have made it more likely to detect complications and survival endpoints in this group. This group also was not further stratified into subgroups such as morbidly obese (BMI > 35 kg/m) or super obese (BMI > 40 kg/m), and results in those subgroups might be different and deserve further study. Additional limitations include the inclusion of all patients from one institution, a higher proportion of Grade 1 tumors in the obese group which might yield a favorable survival profile, and the invariable heterogeneity in adjuvant treatment received by the patients. Despite these limitations, our study is based on a large and heterogeneous group of patients from a large catchment area, and as such, the relationships and conclusions postulated remain highly plausible.
No differences were seen in the proportion or rate of distant metastasis between patients who were obese or nonobese. Additionally, no differences in disease-free survival and local recurrence were seen. Only certain cancers such as esophageal, colon, liver, endometrial, and postmenopausal breast cancers have been found to have a poorer survival profile associated with obesity [26, 29]. These associated cancer subtypes primarily are epithelial-based carcinomas. In addition to not having any survival differences, as seen in our study, patients who are obese and have a soft tissue sarcoma also had a lower proportion of high-grade tumors compared with their nonobese counterparts. Perhaps the biologic features of a soft tissue sarcoma might be more resistant to the proinflammatory cytokines released in the tumor microenvironment by the hypertrophied adipocytes in patients who are obese. However, additional studies are needed to further evaluate the responses seen in soft tissue sarcomas exposed to the proinflammatory state created with obesity.
Obesity has been reported to increase the complexity of surgical procedures. For example, in the case of colorectal cancer resections in patients who are obese, the surgical complexity has been reported to be much greater compared with surgery for patients who are nonobese [20]. Similarly, in the case of total joint arthroplasties, patients who are obese have been reported to have more intraoperative complications [9, 16, 24]. Presumably owing to the hypertrophied adipose tissue surrounding and abutting the surgical site and complicating the surgical procedure, increases in the complexity and complications from the surgery were reported in these studies [9, 16, 24]. However, in our study, obesity did not seem to have an effect on the quality of the surgical resection. The frequencies of patients with positive margins were comparable between the two groups. Additionally, frequencies of patients having wound complications after resection, including wound dehiscence and wound infections, were comparable. This may be attributable to the high rates of wound complications inherent in sarcoma resection in all patients owing to adjuvant therapies such as radiation and large dissection planes required for tumor removal [12]. Thus, although obesity might increase the complexity of certain procedures such as the insertion of prosthetic implants and might lead to early implant failure in such patients after resection, it does not seem to affect the quality of resection of the malignant neoplasm, at least for soft tissue sarcomas that do not involve resection of bone or require implants.
Obesity continues to become more common worldwide and causes increases in the frequency of numerous medical conditions, including cardiovascular disease, stroke, and obstructive sleep apnea. However, we found obesity was not associated with poorer survival in patients with soft tissue sarcomas or with increased frequency of positive margins or wound complications.
Footnotes
Each author certifies that he or she, or a member of his or her immediate family, has no funding or commercial associations (eg, consultancies, stock ownership, equity interest, patent/licensing arrangements, etc) that might pose a conflict of interest in connection with the submitted article.
All ICMJE Conflict of Interest Forms for authors and Clinical Orthopaedics and Related Research editors and board members are on file with the publication and can be viewed on request.
Each author certifies that his or her institution approved the human protocol for this investigation, that all investigations were conducted in conformity with ethical principles of research, and that informed consent for participation in the study was obtained.
References
- 1. Atek M, Traissac P, El Ati J, Laid Y, Aounallah-Skhiri H, Eymard-Duvernay S, Mezimeche N, Bougatef S, Beji C, Boutekdjiret L, Martin-Prevel Y, Lebcir H, Gartner A, Kolsteren P, Delpeuch F, Romdhane HB, Maire BObesity and association with area of residence, gender and socio-economic factors in Algerian and Tunisian adults. PloS One. 2013;8:e75640. doi: 10.1371/journal.pone.0075640.] [[Google Scholar]
- 2. Bifulco M, Pisanti S“Adiponcosis”: a new term to name the obesity and cancer link. J Clin Endocrinol Metab. 2013;98:4664–4665. doi: 10.1210/jc.2013-2645.] [[PubMed][Google Scholar]
- 3. Boeing HObesity and cancer: the update 2013. Best Pract Res Clin Endocrinol Metab. 2013;27:219–227. doi: 10.1016/j.beem.2013.04.005.] [[PubMed][Google Scholar]
- 4. Bowne WB, Antonescu CR, Leung DH, Katz SC, Hawkins WG, Woodruff JM, Brennan MF, Lewis JJDermatofibrosarcoma protuberans: a clinicopathologic analysis of patients treated and followed at a single institution. Cancer. 2000;88:2711–2720. doi: 10.1002/1097-0142(20000615)88:12<2711::AID-CNCR9>3.0.CO;2-M.] [[PubMed][Google Scholar]
- 5. Catalan V, Gomez-Ambrosi J, Rodriguez A, Fruhbeck GAdipose tissue immunity and cancer. Front Physiol. 2013;4:275. doi: 10.3389/fphys.2013.00275.] [[Google Scholar]
- 6. Chee YH, Teoh KH, Sabnis BM, Ballantyne JA, Brenkel IJTotal hip replacement in morbidly obese patients with osteoarthritis: results of a prospectively matched study. J Bone Joint Surg Br. 2010;92:1066–1071. doi: 10.1302/0301-620X.92B8.22764.] [[PubMed][Google Scholar]
- 7. Dowsey MM, Liew D, Stoney JD, Choong PFThe impact of obesity on weight change and outcomes at 12 months in patients undergoing total hip arthroplasty. Med J Aust. 2010;193:17–21.[PubMed][Google Scholar]
- 8. Edge SB AJCC Cancer Staging Manual. New York, NY: Springer; 2009. [PubMed][Google Scholar]
- 9. Elson LC, Barr CJ, Chandran SE, Hansen VJ, Malchau H, Kwon YMAre morbidly obese patients undergoing total hip arthroplasty at an increased risk for component malpositioning? J Arthroplasty. 2013;28:41–44. doi: 10.1016/j.arth.2013.05.035.] [[PubMed][Google Scholar]
- 10. Fine JP, Gray RJA proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496–509. doi: 10.1080/01621459.1999.10474144.[PubMed][Google Scholar]
- 11. Flegal KM, Carroll MD, Ogden CL, Curtin LRPrevalence and trends in obesity among US adults, 1999–2008. JAMA. 2010;303:235–241. doi: 10.1001/jama.2009.2014.] [[PubMed][Google Scholar]
- 12. Geller DS, Hornicek FJ, Mankin HJ, Raskin KASoft tissue sarcoma resection volume associated with wound-healing complications. Clin Orthop Relat Res. 2007;459:182–185. doi: 10.1097/BLO.0b013e3180514c50.] [[PubMed][Google Scholar]
- 13. GETFITTN. Get Fit Tennessee 2013. Available at: . Accessed March 8, 2014.[PubMed]
- 14. Gooley TA, Leisenring W, Crowley J, Storer BEEstimation of failure probabilities in the presence of competing risks: new representations of old estimators. Stat Med. 1999;18:695–706. doi: 10.1002/(SICI)1097-0258(19990330)18:6<695::AID-SIM60>3.0.CO;2-O.] [[PubMed][Google Scholar]
- 15. Gray RJA Class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Statist. 1988;16:1141–1154. doi: 10.1214/aos/1176350951.[PubMed][Google Scholar]
- 16. Gupta AK, Chalmers PN, Rahman Z, Bruce B, Harris JD, McCormick F, Abrams GD, Nicholson GPReverse total shoulder arthroplasty in patients of varying body mass index. J Shoulder Elbow Surg. 2014;23:35–42. doi: 10.1016/j.jse.2013.07.043.] [[PubMed][Google Scholar]
- 17. Haverkamp D, Klinkenbijl MN, Somford MP, Albers GH, van der Vis HMObesity in total hip arthroplasty: does it really matter? A meta-analysis. Acta Orthop. 2011;82:417–422. doi: 10.3109/17453674.2011.588859.] [[Google Scholar]
- 18. Jayawardena R, Byrne NM, Soares MJ, Katulanda P, Hills APPrevalence, Trends and associated socio-economic factors of obesity in South Asia. Obes Facts. 2013;6:405–414. doi: 10.1159/000355598.] [[Google Scholar]
- 19. Kaplan EL, Meier PNonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457–481. doi: 10.1080/01621459.1958.10501452.[PubMed][Google Scholar]
- 20. Krane MK, Allaix ME, Zoccali M, Umanskiy K, Rubin MA, Villa A, Hurst RD, Fichera ADoes morbid obesity change outcomes after laparoscopic surgery for inflammatory bowel disease? Review of 626 consecutive cases. J Am Coll Surg. 2013;216:986–996. doi: 10.1016/j.jamcollsurg.2013.01.053.] [[PubMed][Google Scholar]
- 21. Lawlor DA, Lean M, Sattar NABC of obesity: obesity and vascular disease. BMJ. 2006;333:1060–1063. doi: 10.1136/bmj.333.7577.1060.] [[Google Scholar]
- 22. Maradit Kremers H, Visscher SL, Kremers WK, Naessens JM, Lewallen DGObesity increases length of stay and direct medical costs in total hip arthroplasty. Clin Orthop Relat Res. 2014;472:1232–1239.
- 23. Motaghedi R, Bae JJ, Memtsoudis SG, Kim DH, Beathe JC, Paroli L, Yadeau JT, Gordon MA, Maalouf DB, Lin Y, Ma Y, Cunningham-Rundles S, Liu SSAssociation of obesity with inflammation and pain after total hip arthroplasty. Clin Orthop Relat Res. 2014;472:142–1448. doi: 10.1007/s11999-013-3282-2.] [[Google Scholar]
- 24. Odum SM, Springer BD, Dennos AC, Fehring TKNational obesity trends in total knee arthroplasty. J Arthroplasty. 2013;28:148–151. doi: 10.1016/j.arth.2013.02.036.] [[PubMed][Google Scholar]
- 25. Park JW, Lim SW, Choi HS, Jeong SY, Oh JH, Lim SBThe impact of obesity on outcomes of laparoscopic surgery for colorectal cancer in Asians. Surg Endosc. 2010;24:1679–1685. doi: 10.1007/s00464-009-0829-0.] [[PubMed][Google Scholar]
- 26. Protani M, Coory M, Martin JHEffect of obesity on survival of women with breast cancer: systematic review and meta-analysis. Breast Cancer Res Treat. 2010;123:627–635. doi: 10.1007/s10549-010-0990-0.] [[PubMed][Google Scholar]
- 27. Rajgopal R, Martin R, Howard JL, Somerville L, MacDonald SJ, Bourne ROutcomes and complications of total hip replacement in super-obese patients. Bone Joint J. 2013;95:758–763. doi: 10.1302/0301-620X.95B6.31438.] [[PubMed][Google Scholar]
- 28. Ramos-Nino METhe role of chronic inflammation in obesity-associated cancers. ISRN Oncol. 2013;2013:697521.[Google Scholar]
- 29. Sinicrope FA, Foster NR, Sargent DJ, O’Connell MJ, Rankin CObesity is an independent prognostic variable in colon cancer survivors. Clin Cancer Res. 2010;16:1884–1893. doi: 10.1158/1078-0432.CCR-09-2636.] [[Google Scholar]
- 30. Tavani A, Soler M, La Vecchia C, Negri E, Gallus S, Franceschi SBody weight and risk of soft-tissue sarcoma. Br J Cancer. 1999;81:890–892. doi: 10.1038/sj.bjc.6690781.] [[Google Scholar]
- 31. Wang Z, Yang L, Jiang Y, Ling ZQ, Li Z, Cheng Y, Huang H, Wang L, Pan Y, Wang Z, Yan X, Chen YHigh fat diet induces formation of spontaneous liposarcoma in mouse adipose tissue with overexpression of interleukin 22. PloS One. 2011;6:e23737. doi: 10.1371/journal.pone.0023737.] [[Google Scholar]
- 32. Watson SE, Li Z, Tu W, Jalou H, Brubaker JL, Gupta S, Huber JN, Carroll A, Hannon TSObstructive sleep apnoea in obese adolescents and cardiometabolic risk markers. Pediatr Obes. 2013 Sept 17. [Epub ahead of print].
- 33. Williams SCLink between obesity and cancer. Proc Natl Acad Sci USA. 2013;110:8753–8754. doi: 10.1073/pnas.1308182110.] [[Google Scholar]