A nationwide observational cohort study of the relationship between beta-blockade and survival after hip fracture surgery

This observational cohort study adheres to the principles of the Declaration of Helsinki. Ethical approval was obtained from the National Ethics Review Board (reference number 2019-02094). The national and prospectively collected patient register for proximal femur fractures “Rikshoft” was utilized to identify all adults aged \(\ge 18\text{ years}\) undergoing primary surgery for a proximal femur fracture between January 2008 and December 2017 in Sweden. Patients with pathological fractures from primary or secondary bone tumors and those who had a re-operation were excluded. The Rikshoft register was created in the 1980s. It collects and maintains data on all proximal femur fractures suffered in Sweden [16]. Patient data obtained from the register included the date of hospital admission, age, sex, fracture type, American Society of Anesthesiologist (ASA) classification, Body Mass Index (BMI), date of surgery, surgical method, and hospital discharge date.

Data on postoperative mortality, comorbidity and beta-blocker use were obtained from national registers maintained by the National Board of Health and Welfare. Mortality data \(\le 90\text{ days}\) postoperatively were collected from the Cause of Death register that records all deaths of Swedish residents and cause(s) of death stated on the death certificate. Comorbidity data were collected from the National Patient Register of both primary and secondary healthcare providers. This information was used to calculate a Charlson Comorbidity Index (CCI) for each patient. Finally, data on beta-blocker prescriptions (ACT codes C07AA, C07AB, C07AG) were attained from the national drug register. The national drug register is a population-based database that records all drug prescriptions issued by physicians in Sweden. Beta-blocker data were obtained within 1 year before and after surgery. This was chosen since beta-blockers rarely are discontinued once initiated and, therefore, commonly issued on a long-term basis covering up to a one-year period with a single prescription. Patients were included as beta-blocker users (BB⁺) if they were issued and collected a beta-blocker prescription. All other patients were considered beta-blocker non-users (BB⁻). Patients on beta-blocker therapy were assumed to continue beta-blocker intake during their hospital admission. Due to the lack of access to nationwide hospital records, this assumption was tested by assessing the electronic medical records of all patients included from one selected county during a specified time-period.

The primary outcome of interest was 30-day all-cause mortality following surgery for proximal femur fractures. The secondary outcome of interest was cause-specific mortality. Patient demographics and clinical outcomes were compared between BB⁺ and BB⁻ groups. Statistical tests included Chi-square test for categorical variables and Student’s t test (or Mann–Whitney \(U\) test where appropriate) for continuous variables. The relationship between beta-blocker exposure and 30-day all-cause and cause-specific mortality was evaluated using Poisson regression with robust standard errors of variance [17]. Potential confounding was accommodated in the models by including age, sex, ASA, CCI, type of fracture and surgical method. The time from surgery to death or to the end of 30-day follow-up was used as an offset in the Poisson regression models. The outcome measure was reported as Incidence Rate Ratio (IRR) with 95% confidence interval (CI). Multiple imputations were used to handle missing values [18]. Results were considered statistically significant at a two-sided p value of < 0.05.

Cardiovascular deaths are of specific relevance in the context of adrenergic overload as a result of tissue injury and subsequent major surgery [12]. Hip surgery has been especially associated with increased risk of major cardiac events in the postoperative period occurring in six percent of patients [14]. Adverse cardiac events during the immediate postoperative period does not only affect in-hospital survival but has a major influence on the continued mortality risk for several months after noncardiac surgery [23]. Therefore, to achieve reductions in postoperative mortality, focus must be directed towards the optimization of this high-risk group. Raw data in this study portray cardiovascular events as a crucial cause of death with a total of 4182 cardiovascular fatalities, corresponding to 41.0% of deaths in all patients. Beta-blocker use caused a 76% risk reduction in the incidence of cardiovascular deaths (IRR 0.24, 95% CI 0.22–0.26, p < 0.001). Similarly, the PeriOperative ISchemic Evaluation (POISE) trial published in 2008 showed a significant reduction in adverse cardiac events in noncardiac surgical patients treated with metoprolol [24]. The degree of benefit appears to increase with increasing cardiac risk index [25].

Furthermore, two meta-analyses and a Cochrane review published in the last 15 years on beta-blocker therapy and noncardiac surgery have outlined protective effects on adverse cardiac events [11, 26, 27]. Consequently, international guidelines from the American Heart Association, American College of Cardiology (AHA/ACC) and the European Society of Cardiology (ESC) both advise against discontinuing regular beta-blockers prior to or following any surgery [28, 29]. The results of this nationwide study are in line with these recommendations and support the use of beta-blockers in orthopedic patients undergoing emergency surgery for hip fractures to prevent postoperative cardiac deaths. Interestingly, 55.3% (n = 45,601) of BB⁻ patients had a diagnosis of cardiovascular disease prior to admission. In particular, congestive heart failure was found in 10.0% (n = 8,252) and arrythmia in 10.3% (n = 8,479) of patients in the BB⁻ group, suggesting that there may be a substantial underuse of beta-blockade in this patient group.

The Agency for Healthcare Research and Quality’s evidence report from 2001 promotes the use of perioperative beta-blocker therapy as the second-highest-rated safety practice protecting against surgical mortality [30]. In 2013, an updated version of the report amended these recommendations, stating that perioperative beta-blockers should not be considered a safe practice for all surgical patients [31]. The revision was to a large extent based on the observation of an increase in all-cause mortality and non-fatal strokes from recent randomized trials [24, 26]. In contrast, the results of the current study detect a significant reduction in 30- and 90-day all-cause mortality and a survival benefit that was doubled when comparing beta-blocker users to non-users irrespective of which cause of death analyzed. Although limited by the retrospective nature, the authors hypothesize that this observed pattern could be the result of a widespread benefit of a general antagonistic action to catecholamine hyperactivation in a patient group with high general frailty whose physiology is inadequate to safely deal with uncontrolled adrenergic storms.

There are other methodological differences which can explain this difference. Most scientific work on beta-blockers in noncardiac surgery is based on a mixture of noncardiac surgery. For instance, the POISE trial included vascular, orthopedic and abdominal surgery. While this may lead to increased generalizability, it increases the risk of including undetected biases. To the best knowledge of the authors, our group is the first to evaluate beta-blockers in fracture surgery, specifically. Furthermore, the POISE trial administered a high dose of extended-release metoprolol (200 mg) to beta-blocker naive patients with drug administration occurring only hours prior to surgery. These decisions reflect a major difference to the current study which evaluates a cohort with a narrow age span, undergoing surgery for the same indication, and with most patients being prescribed 50 mg of extended-release metoprolol. The authors hypothesize that these methodological modifications may have led to fewer bradycardic/hypotensive episodes and thereby avoided the mechanisms believed to underlie previously observed increased stroke risks. Our results demonstrate a 70% risk reduction in fatal strokes for beta-blocker users (IRR 0.30, 95% CI 0.20–0.44, p < 0.001). Similar results have been reported by Obeid et al. who demonstrated a 34% reduction in stroke/death risk in beta-blocker treated patients undergoing vascular surgery [32].

The literature on beta-blockers in noncardiac surgery is growing and there is increasing belief that catecholamines play an imperative role in the initiation of inflammatory responses and subsequent complications and resulting multiorgan failure [33]. Regression analysis in the current study demonstrate a significant reduction in deaths attributable to respiratory causes, sepsis and multiorgan failure (Table 5). The stress response from tissue injury results in the release of cytokines interleukin (IL)-1, IL-6 and tumor necrosis factor-α [22]. Modulating immune responses by decreasing cytokine release through beta-blockers has been reported in animal models of traumatic injury [34, 35]. Cytokine downregulation is one proposed mechanism explaining reports of survival advantages for beta-blocker exposed patients treated for severe sepsis and burns [36, 37]. In animal models of pulmonary dysfunction, beta-blockade has demonstrated attenuation of lung injury and acute respiratory distress syndromes [38]. These mechanisms could explain the reduction in the deaths due to sepsis, pulmonary events and multiorgan failure seen in beta-blocker users in the current study. Regrettably, due to a high proportion of missing data in postoperative complications in the Rikshoft register, we are unable to account for the incidence of non-fatal complications. This is a recognized limitation and we would encourage this to be addressed in a randomized trial.

Furthermore, due to the use of national quality registers we cannot with certainty conclude that all BB⁺ patients received continued beta-blocker therapy after hospital admission. However, a review of 2443 patients from three hospitals in Örebro County demonstrates that discontinuation of beta-blocker treatment does not normally occur. The continuation of a regular beta-blocker whilst in hospital is also clearly enforced by both American and European cardiology guidelines [28, 29]. In contrast to major abdominal surgery, orthopedic patients do not generally have any restrictions regarding enteral nutrition postoperatively, which further reduces the risk of in-hospital discontinued drug administrations.