Obesity and Pancreatic Cancer: A Matched-Pair Survival Analysis

Background: Morbid obesity is a risk factor for pancreatic ductal adenocarcinoma (PDAC). However, the impact of obesity on postoperative outcomes and overall survival in patients with PDAC remains a controversial topic. Methods: Patients who underwent pancreatic surgery for PDAC between 1997 and 2018 were included in this study. Matched pairs (1:1) were generated according to age, gender and American Society of Anesthesiologists status. Obesity was defined according to the WHO definition as BMI ≥ 30 kg/m2. The primary endpoint was the difference in overall survival between patients with and without obesity. Results: Out of 553 patients, a total of 76 fully matched pairs were generated. Obese patients had a mean BMI-level of 33 compared to 25 kg/m2 in patients without obesity (p = 0.001). The frequency of arterial hypertension (p = 0.002), intraoperative blood loss (p = 0.039), and perineural invasion (p = 0.033) were also higher in obese patients. Clinically relevant postoperative complications (p = 0.163) and overall survival rates (p = 0.885) were comparable in both study groups. Grade II and III obesity resulted in an impaired overall survival, although this was not statistically significant. Subgroup survival analyses revealed no significant differences for completion of adjuvant chemotherapy and curative-intent surgery. Conclusions: Obesity did not affect overall survival and postoperative complications in these patients with PDAC. Therefore, pancreatic surgery should not be withheld from obese patients.


Introduction
Obesity has become a pandemic affecting more than 400 million people worldwide with the incidence rising [1][2][3][4]. Major complications of obesity are cardiovascular diseases, diabetes mellitus type 2 and the development of several cancers, including pancreatic ductal adenocarcinoma (PDAC) [5][6][7]. PDAC is a devastating malignancy with a rate of overall survival (OS) of about 5%. The aging and growing population is leading to an increase in the rate of pancreatic cancer diagnoses. Surgery and multimodal treatment remain the only curative therapy, but unfortunately, only 20% of patients are candidates for surgery at the time of diagnosis. Despite advances in surgical techniques, perioperative morbidity and mortality rates remain high in pancreatic surgery [8,9]. In obese patients, surgical therapy and perioperative management are even more challenging due to the coexistence of several comorbidities and excessive fat tissue [10]. In the past, several studies have indicated that pancreatectomy for PDAC in obese and overweight patients is associated with high postoperative morbidity and mortality rates [11][12][13][14]. However, the evidence is conflicting as other studies have

Definitions and Outcomes
Demographic and clinical characteristics included age, sex, BMI, and preoperative status of patients according to the ASA status classification and comorbidities listed for cardiac history (coronary artery disease, congestive heart failure, history of myocardial infarction, artificial valves), pulmonary diseases (chronic obstructive pulmonary disease, pulmonary hypertension, bronchial asthma), arterial hypertension, diabetes mellitus, history of chronic pancreatitis, and preoperative biliary stenting [34]. Histopathological data were analyzed by the Department of Pathology, Universitätsmedizin Mannheim, Mannheim, Germany according to the 6th-8th versions of the TNM classification. The primary endpoint was defined as the OS difference in patients with histologically confirmed PDAC between those with obesity and those without. Secondary endpoints included intraoperative outcome, postoperative morbidity, pathological characteristics, postoperative length of hospital stay, and adjuvant chemotherapy. Obesity was defined according to the WHO criteria as BMI ≥ 30 kg/m 2 [35,36]. Preoperative laboratory values were assessed as following: albumin (normal: ≥35 g/L), bilirubin (normal: ≤1.2 mg/dL), creatinine (normal: ≤1.4 mg/dL), c-reactive protein (CRP) (normal: ≤2,9 mg/dL), hemoglobin (normal: ≥13 g/dL), platelets (normal: 145-348 × 10 9 /L), and international normalized ratio (INR) (normal: 0.9-1.15). Operative characteristics included operating time (min), total blood loss (mL), surgical details, and perioperative transfusion within 48 h. Postoperative complications were graded in line with the Clavien-Dindo classification if they occurred within 90 days of index operation [37]. In addition, specific complications such as postoperative pancreatic hemorrhage (PPH), postoperative pancreatic fistula (POPF), and postoperative delayed gastric emptying (DGE) were assessed according to the criteria of the International Study Group of Pancreatic Surgery (ISGPS) [38][39][40][41]. Long-term survival was recorded by telephone follow-up in December, 2019, and the survival period or the time of death was noted. Patients lost to follow up were censored.

Statistical Analysis
Categorical data were presented by absolute and relative frequencies (percentage) and compared using Pearson's χ 2 test or Fisher's exact test. Quantitative data were summarized as the mean (standard deviation) or median (interquartile range or 95% confidence interval (95% CI)) and compared depending on the pattern of distribution using the unpaired 2-tailed t-test or Wilcoxon test. Survival analysis was done by using the Kaplan-Meier method with the log-rank test. Bonferonni correction was applied for multiple testing. p-values < 0.05 were defined as statistically significant. R (version 3.6.1) was used for all statistical analyses.

Patients Characteristics
Out of 553 patients who underwent surgery for PDAC, a total of 76 fully matched pairs were generated according to the predefined criteria ( Figure 1). J. Clin. Med. 2020, 9, x FOR PEER REVIEW 3 of 12 mean (standard deviation) or median (interquartile range or 95% confidence interval (95% CI)) and compared depending on the pattern of distribution using the unpaired 2-tailed t-test or Wilcoxon test. Survival analysis was done by using the Kaplan-Meier method with the log-rank test.
Bonferonni correction was applied for multiple testing. p-values < 0.05 were defined as statistically significant. R (version 3.6.1) was used for all statistical analyses.

Patients Characteristics
Out of 553 patients who underwent surgery for PDAC, a total of 76 fully matched pairs were generated according to the predefined criteria ( Figure 1). Patient characteristics according to the study groups are outlined in Table 1. Obese patients had a mean BMI of 33 kg/m 2 compared to 25 kg/m 2 in patients without obesity (p = 0.001). The majority of obese patients had grade 1 obesity. There were no patients with ASA status higher than grade III. Patients with obesity had a higher frequency of arterial hypertension than patients without obesity (n = 56 (74%) vs. n = 36 (47%), p = 0.002, respectively). Other baseline characteristics including the values from preoperative laboratory tests were well-balanced between the groups. A total of 46 obese patients (61%) received adjuvant chemotherapy compared to 44 patients (58%) without obesity (p = 0.180). As expected, gemcitabine-based chemotherapy was the most commonly employed chemotherapeutic agent. Of note is that five obese patients (n = 3 obesity grade 1, n = 2 obesity grade 3) discontinued gemcitabine-based chemotherapy due to increased toxicity compared to only one patient in the control group (p = 0.102). Patient characteristics according to the study groups are outlined in Table 1. Obese patients had a mean BMI of 33 kg/m 2 compared to 25 kg/m 2 in patients without obesity (p = 0.001). The majority of obese patients had grade 1 obesity. There were no patients with ASA status higher than grade III. Patients with obesity had a higher frequency of arterial hypertension than patients without obesity (n = 56 (74%) vs. n = 36 (47%), p = 0.002, respectively). Other baseline characteristics including the values from preoperative laboratory tests were well-balanced between the groups. A total of 46 obese patients (61%) received adjuvant chemotherapy compared to 44 patients (58%) without obesity (p = 0.180). As expected, gemcitabine-based chemotherapy was the most commonly employed chemotherapeutic agent. Of note is that five obese patients (n = 3 obesity grade 1, n = 2 obesity grade 3) discontinued gemcitabine-based chemotherapy due to increased toxicity compared to only one patient in the control group (p = 0.102).

Intraoperative Outcome and Pathological Characteristics
The intraoperative and pathological characteristics are summarized in Table 2. Blood loss was higher in obese than in non-obese patients (1050 ± 760 mL vs. 809 ± 607 mL, respectively, p = 0.039). Pancreaticoduodenectomies were the most frequently performed procedures in both study groups (n = 57 (75%) vs. n = 56 (74%), respectively). Six obese patients (8%) had undergone palliative bypass compared to five patients in the control group (7%). Pathological characteristics revealed a higher frequency of perineural invasion in obese than in non-obese patients (p = 0.033). No significant differences in tumor size, histologic grading, lymphovascular invasion, and metastasis were observed between obese and non-obese patients.

Postoperative Morbidity
There were no statistically significant differences regarding clinically relevant postoperative complications (Clavien-Dindo ≥III) between the obese and non-obese patients (n = 20 (26%) vs. n = 12 (16%), respectively, p = 0.163) ( Table 3). The 30-day mortality rates were 5% in both study groups. Complications such as burst abdomen, radiological or endoscopic interventions, and pancreatic surgery-specific complications (DGE, POPF, and PPH) were equally distributed between the groups. One patient in the non-obese group died of an intraoperative cardiac arrest, whereas the other patients had a multiorgan dysfunction syndrome.

Discussion
Surgery and adjuvant chemotherapy constitute the standard of care for curative therapy for PDAC [42]. In the last three decades, advances in oncological treatment strategies have markedly improved OS in PDAC. Still, there is a lack of data regarding predictors for long-term survival and risk stratification of patients suffering from PDAC as the survival rates are highly variable in certain patient populations [43][44][45][46][47]. Obesity represents a major risk factor for developing pancreatic cancer [6,48]. Due to a rising incidence of obesity and PDAC, the surgical treatment of obese patients with pancreatic cancer is expected to increase further [4,49]. However, inconsistent postoperative outcomes and oncological results have been outlined in previous studies that included obese patients with PDAC following oncological therapy. This was mainly due to the use of heterogeneous definitions of obesity and inhomogeneous patient cohorts as well as variable surgical techniques [11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26]50]. Some authors detected comparable OS rates in obese and non-obese patients (19.8 months vs. 23.5 months, respectively; p = 0.46) [11], while others observed a worse survival in patients with BMI > 35 kg/m 2 than in those with BMI < 23 kg/m 2 (13.2 months vs. 27.4 months, respectively; p = 0.02) [18]. Tsai et al. even reported improved OS in patients with a BMI > 25 than in patients with a normal weight following pancreaticoduodenectomy (20.3 and 20.1 months vs. 14.6 months, respectively) [17]. Therefore, we addressed this lack of evidence by analyzing a fully matched pair analysis of 152 patients with PDAC and surgical treatment in a tertiary referral center. We matched the groups according to age, gender, ASA status, and BMI-level and detected a comparable long-term survival in patients with BMI ≥ 30 kg/m 2 and in non-obese patients.
Age, gender, chronic pancreatitis, and modifiable factors such as obesity and metabolic syndrome-related diabetes are well-known risk factors for PDAC. Chronic pancreatitis and obesity are associated with desmoplasia in pancreatic cancer. However, the effect of desmoplasia is not fully understood and remains a matter of controversy [51,52]. Desmoplasia represents a two-edged sword. On the one hand, it has been reported to prevent local tumor invasion through encapsulation of the tumor by fibrous tissue, but on the other hand, desmoplasia might cause a poor response to Next, we assessed the impact of chemotherapy on OS in patients by excluding patients who did not receive or complete adjuvant or palliative chemotherapy. Six obese patients discontinued chemotherapy due to increased toxicity (n = 1 FOLFIRNOX, n = 5 gemcitabine-based chemotherapy), compared to one patient in the control group (p = 0.056). We detected a lower median survival rate in obese patients than in the control group (19 months (95% CI: 11-43 months) vs. 42 months (95% CI: 21-84 months, respectively)), however, this did not reach statistical significance (p = 0.133) ( Figure S3).

Discussion
Surgery and adjuvant chemotherapy constitute the standard of care for curative therapy for PDAC [42]. In the last three decades, advances in oncological treatment strategies have markedly improved OS in PDAC. Still, there is a lack of data regarding predictors for long-term survival and risk stratification of patients suffering from PDAC as the survival rates are highly variable in certain patient populations [43][44][45][46][47]. Obesity represents a major risk factor for developing pancreatic cancer [6,48]. Due to a rising incidence of obesity and PDAC, the surgical treatment of obese patients with pancreatic cancer is expected to increase further [4,49]. However, inconsistent postoperative outcomes and oncological results have been outlined in previous studies that included obese patients with PDAC following oncological therapy. This was mainly due to the use of heterogeneous definitions of obesity and inhomogeneous patient cohorts as well as variable surgical techniques [11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26]50]. Some authors detected comparable OS rates in obese and non-obese patients (19.8 months vs. 23.5 months, respectively; p = 0.46) [11], while others observed a worse survival in patients with BMI > 35 kg/m 2 than in those with BMI < 23 kg/m 2 (13.2 months vs. 27.4 months, respectively; p = 0.02) [18]. Tsai et al. even reported improved OS in patients with a BMI > 25 than in patients with a normal weight following pancreaticoduodenectomy (20.3 and 20.1 months vs. 14.6 months, respectively) [17]. Therefore, we addressed this lack of evidence by analyzing a fully matched pair analysis of 152 patients with PDAC and surgical treatment in a tertiary referral center. We matched the groups according to age, gender, ASA status, and BMI-level and detected a comparable long-term survival in patients with BMI ≥ 30 kg/m 2 and in non-obese patients.
Age, gender, chronic pancreatitis, and modifiable factors such as obesity and metabolic syndrome-related diabetes are well-known risk factors for PDAC. Chronic pancreatitis and obesity are associated with desmoplasia in pancreatic cancer. However, the effect of desmoplasia is not fully understood and remains a matter of controversy [51,52]. Desmoplasia represents a two-edged sword. On the one hand, it has been reported to prevent local tumor invasion through encapsulation of the tumor by fibrous tissue, but on the other hand, desmoplasia might cause a poor response to chemotherapy due to reduced tumor vascularization resulting in defective drug delivery to the tumor [51,52]. In a previous study, we showed that desmoplastic reaction in patients with chronic pancreatitis and pancreatic cancer had no impact on survival after resection with curative intent [27]. As obese patients are accompanied by severe comorbidities, adjuvant therapy for them can be more challenging and result in dose reductions or even discontinuation of therapy [53]. In the present study, although the risk factors posed by diabetes and the history of chronic pancreatitis were not matched, we detected no difference among the distribution in these groups. Moreover, adjuvant therapy was well balanced between the groups, although a higher proportion of obese patients discontinued chemotherapy following surgery. This was reflected by the slightly worse survival rate in obese patients detected in the subgroup analysis of patients who received adjuvant chemotherapy although this did not reach statistical significance.
Obesity is characterized by changes in cancer-associated adipokine production [54]. Moreover, DNA damage pathways, adipokines, and proinflammatory environment related to obesity contribute to angiogenesis and metastasis with advanced tumor size, residual tumor and nodal involvement [17,18]. In line with these findings, obese patients exhibited a higher frequency of perineural invasion in the present study, indicating more aggressive tumor biology. We also observed a higher frequency of nodal involvement in obese patients although this did not reach statistical significance.
Furthermore, obesity is known to be associated with an increased risk of cardiovascular diseases, diabetes, and premature death [55]. Previous studies have found that surgical therapy of obese patients with PDAC is associated with high morbidity rates (e.g., intraabdominal abscess-and/or fluid collections, wound infection rates, and POPF) of up to 68% [11][12][13][14]. The present study demonstrates that obesity is not associated with increased postoperative overall complications, pancreatic surgery-specific complications, length of stay, and 90-day mortality rates as described in the literature [16,[56][57][58].
Moreover, while the operating time was comparable for both study groups, intraoperative blood loss was significantly increased in obese patients, reflecting a more demanding resection in obese patients, which is in line with other studies [16,17]. Still, one has to consider the intraoperative difficulties arising in obese patients from a higher amount of visceral fat tissue, which demand special surgical expertise and equipment (e.g., retractors, positioning). Therefore, oncological treatment of obese patients with PDAC should be restricted to high-volume pancreatic centers with experience in handling obese patients [59,60].
There are some limitations to the present study. First, this is a single-center study with a retrospective data study design. Therefore, there will be some selection and reporting bias. Second, there are some missing values for pathological characteristics in this prospectively recorded data base, which is mainly due to the non-standardized assessment of perineural or lymphatic invasion before 2009. Third, the patient cohort consists of mainly patients with grade I obesity. Although we detected impaired survival in patients with grade II and III obesity, this was not statistically significant. This was probably due to the small sample size of patients with high-grade obesity. Fourth, the follow-up time was rather short in the present analysis. This was mainly due to the number of censored patients and the high number of patients who died during the follow-up period (up to 55%).

Conclusions
Within the above-mentioned limitations, obesity was not associated with worse postoperative outcomes and long-term survival despite a predominance of perineural invasion and higher intraoperative blood loss. However, the impact of high-grade obesity still remains unclear.
These findings should therefore be considered when assessing patients for operation and when counseling patients about their operative risk. Nevertheless, pancreatic surgery should not be withheld for obese PDAC patients when it is oncologically appropriate. There is urgent need for multicenter-or register-based studies to validate the present results before definitive conclusions can be drawn.