Influence of beta-blockers and other non-cancer drugs on outcomes after pancreatic cancer resection

Abstract

The mainstay of therapy for pancreatic cancer with curative intent remains surgical resection. The perioperative phase is characterized by physiological and psychological stress reactions mediated mainly by an excess release of catecholamines and prostaglandins, which may have protumorigenic effects1. These effects could potentially be inhibited by usual beta-blockers and cyclo-oxygenase (COX) inhibitors in terms of drug repurposing, but there is conflicting evidence for the influence of beta-blockers and other non-cancer drugs in the perioperative setting. This retrospective analysis of the institution’s prospective pancreatic database assessed the influence of beta-blockers, acetylsalicylic acid as well as their combined use, angiotensin-converting inhibitors/angiotensin II receptor type 1 antagonists, statins, and calcium channel blockers on perioperative outcome and survival of patients undergoing resection for pancreatic cancer. A logistic regression model was used to assess the influence of the different medications on postoperative morbidity and mortality. Univariable and multivariable Cox regression models were computed to evaluate the influence of individual variables on survival. Furthermore, 1 : 1 matching by the nearest-neighbour method, considering the variables age, sex, AJCC stage, ASA grade, arterial hypertension, and other cardiovascular co-morbidities, was performed to adjust for baseline imbalances. A total of 914 adults undergoing resection for pancreatic cancer were analysed, of whom 263 (28.8 per cent) were receiving concurrent beta-blocker therapy. The mean(s.d.) age of the cohort was 65.6(10.3) years and 422 of the patients (46.2 per cent) were women. Patients under beta-blocker therapy were older, and more likely to have an ASA fitness grade of III or higher and co-morbidities, especially hypertension and other cardiovascular co-morbidities. The medications had no influence on postoperative morbidity and mortality, in particular no effect on pancreas-specific complications such as pancreatic fistula, postpancreatectomy haemorrhage, delayed gastric emptying, and intra-abdominal abscess. Patients using beta-blockers (HR for death 1.19, 95 per cent c.i. 1.00 to 1.41; P = 0.047) and those taking acetylsalicylic acid (HR 1.24, 1.01 to 1.54; P = 0.044) showed impaired overall survival in the unadjusted survival analyses, whereas all other medications investigated had no impact on survival. However, the negative impact on survival diminished in the multivariable analysis, and also in the propensity score-matched cohort after adjusting for baseline imbalances in age and co-morbidities. In multivariable analysis, negative prognostic factors were age at least 67 years, neoadjuvant therapy, ASA grade III or higher, AJCC stage III, and R1 resection (Fig. 1). The only positive prognostic factor was concurrent statin use (HR 0.76, 95 per cent c.i. 0.58 to 0.99). Forest plot of multivariate cox regression analysis of overall survival The analysis included 914 patients. HRs are shown with 95% confidence intervals. ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker. Several studies have reported a positive effect of beta-blocker therapy with regard to short-term2 and long-term3 outcomes after cancer surgery. These effects could not be reproduced in the present study. The survival difference in the unadjusted analyses, which was not present in the adjusted and matched analyses, can be explained for the main part by the greater co-morbidity in patients in the beta-blocker group and their older age. The only medication with a potentially positive effect on survival was statin use, which is in line with a recent population-based study from Norway4. Nevertheless, this finding needs to be further explored in future studies alongside other innovative approaches to harnessing the perioperative phase for cancer-directed therapy. One such approach is currently being investigated in the randomized controlled PROSPER trial, which is assessing whether the combination of propranolol, a non-selective beta-blocker, and etodolac, a COX-2-inhibitor, may have a positive effect in beta-blocker-naive patients undergoing pancreatic cancer surgery5. The authors have no funding to declare. The authors declare no conflict of interest.