Influence of a chronic beta-blocker therapy on perioperative opioid consumption – a post hoc secondary analysis

This study was designed as a post hoc analysis based on the study “Biomarker Development for postoperative Cognitive Impairment in the Elderly (BioCog)”, a multicenter, prospective clinical observational study with the aim to develop biomarkers to assess the risk and predict the occurrence of postoperative delirium (POD) and postoperative cognitive dysfunction (POCD) [12]. All procedures involving humans were in accordance with the ethical standards of the institutional research committee and with the 1964 Declaration of Helsinki and its later amendments. The study was approved by the local ethics committee at Charité – Universitätsmedizin Berlin (EA 2/092/14) on the 31.07.2014 and registered at ClinicalTrials.gov (NCT02265263) on the 15.10.2014 with the principal investigator being Univ.-Prof. Dr. med. Claudia Spies. Informed written consent was obtained from all patients. Data collection for the entire BioCog study took place between October 2014 and June 2019 at the Department of Anesthesiology and Intensive Care Medicine at Campus Charité Mitte and Campus Virchow-Klinikum, Charité – Universitätsmedizin Berlin, Germany, and the Department of Intensive Care Medicine at the University Medical Center Utrecht, Netherlands. We adhered to the CONSORT guidelines and the STROBE checklist.

The following patients were screened for inclusion in the post hoc secondary analysis to examine whether a chronic BB therapy had an influence on perioperative opioid consumption: elective non-cardiac surgery (general, thoracic, orthopedic, trauma, gynecologic, urologic, ear nose, throat, neurosurgical, dermatologic) with an anesthesia duration of ≥ 60 min, total intravenous general anesthesia without additional regional anesthesia. Only BioCog study participants from Berlin, Germany, were included in the secondary analysis. Patients, who fulfilled the inclusion criteria, were divided into two groups, whether they received a chronic BB therapy as a regular medication at the time of study inclusion or not, and compared to each other. Additionally, we exluded patients with chronic pain from our post hoc analysis to harmonise the study cohort. In the main study, patients were advised to take their BB on the morning of surgery and to resume BB treatment as quickly as possible postoperatively.

The study protocol has been previously described [12]. Preoperative data were collected from the preoperative visit as well as the individual patient file and included age, gender, weight, Body Mass Index (BMI), American Society of Anesthesiologists Physical Status (ASA PS), type of planned surgery, consumption of alcohol and/or nicotine, preexisting medical conditions and regular medication intake. Perioperative data were extracted from the electronic patient data management system (COPRA6®). This included the duration of anesthesia, administered medications, cumulative opioid application including all opioids administered from the beginning of anesthesia until the end of recovery room treatment, length of stay in the recovery room, PONV, incidence of intolerable pain and rate of admission to the intensive care unit (ICU)/postanaesthesia care unit (PACU). Postoperative data included length of hospital stay and rate of complications based on the Clavien-Dindo-Classification [13, 14]. For the ease of use, complications were divided into the following categories: 1) no complication, 2) complications (without death), 3) death. The follow-up period was from the beginning of treatment in the recovery room or ICU/PACU until discharge from hospital or death of the patient.

Postoperative pain was examined using the Numeric Rating Scale (NRS), Behavioral Pain Scale (BPS), BPS non intubated (BPS-NI), or the Critical-Care Pain Observation Tool (CPOT) [15‐17]. For an easier description of postoperative pain, multiple scores were compiled into one cumulative pain score for the BioCog study. This score only differentiates between tolerable and intolerable pain until the 7. postoperative day. The former was defined as a pain with NRS > 4 or BPS/BPS NI > 5 or COPT > 2 in one of the postoperative visits, which were undertaken twice daily in the morning and evening until the 7. postoperative day.

Opioids identified from anesthesia protocols included Fentanyl, Remifentanil, Morphine and Piritramid. Perioperative administered opioids were converted to their respective morphine equivalent dose according to their therapeutic potency [18]. For better comparability, the morphine index was also determined, defined as the cumulative perioperative morphine consumption divided by BMI. For comparison of the different BB, the value of the achieved target dose in percent was calculated, according to the guideline of the European Society of Cardiology (ESC) for the diagnosis and treatment of acute and chronic heart failure of 2016 [19].

All statistical analyses are exploratory in nature and not considered confirmatory. Statistical analysis was done using SPSS^® Statistics 25 (SPSS, Inc., Chicago, IL). Graphics were created using both SPSS and Excel. All binary and categorical variables are reported as absolute and relative frequencies, continuous non-normally distributed variables are listed as median with interquartile range (IQR). Normal distribution was tested using histograms and the Kolmogorov-Smirnov test. Testing for differences between two independent groups was done using Mann-Whitney U test for continuous non-normally distributed variables and Pearson chi-square test for categorical non-normally distributed variables.

A simple linear regression analysis was used to test whether beta-blockade or the dose of BB therapy had an effect on perioperative opioid consumption. Multiple regression analysis was used to identify risk factors for predicting an increased or decreased morphine index. Subsequently, a normal distribution of the data used was obtained by regressing the morphine index, which allowed further precision of the confounder adjustment. The individual significant predictors were then graphically represented using bivariate regression analysis and grouped boxplots.

In total, 747 patients were recruited for inclusion in the BioCog study in Berlin. Thereof, 247 patients were exluded for post hoc analysis, as their respective anesthesia protocol was recorded with an older version of the patient data management system where an exact calculation of the perfusor-controlled opioid administration, as is the case with Remifentanil, was not possible. Of the remaining 439 analyzed protocols, 106 patients fulfilled the inclusion criteria of which 37 patients had a long-term BB therapy (Fig. 1). Baseline characteristics were similar between groups (Table 1).

Preexisting medical conditions are displayed in Table 2. Patients with chronic BB therapy presented with significantly more arterial hypertension, chronic renal failure and hyperlipoproteinemia.

Data of the perioperative phase and incidence of complications are shown in Tables 3 and 4, respectively. Patients on chronic BB therapy did not differ significantly from patients without BB therapy regarding the cumulative total opioid dose during surgery. Similarly, median anesthesia times were similarly distributed. There were no differences in the amount of administered blood products, catecholamines, antihypertensive agents and atropine. Postoperatively, no differences concerning intolerable pain, PONV as well as recovery room length of stay between both groups were detected. Admission to ICU or PACU as well as length of stay in the ICU or PACU (BB: 0 days (0–0.1); no BB: 0 days, p = 0.608) in median and hospital length of stay (BB: 6 (3.5–9) days; no BB: 6 (3.5–9) days, p = 0.828) in median did not differ between both groups. The incidence of postoperative complications are given in Table 4, revealing no significant differences between groups.

There was no correlation between doses of chronic BB therapy and increasing morphine equivalents (Supplementary Fig. 1).

Using multiple regression analyses independent predictors for the amount of perioperative opioid consumption, expressed by the morphine index, were examined (Table 5). Chronic BB therapy had no influence on the opioid consumption, whereas age, gender, anesthesia duration as well as type of surgery had. Their respective influence on the morphine index is given by the regression coefficient. Regarding age, each year yielded an increase of the morphine index by 0.547 mg/BMI, whereas female gender leads to a reduction of the morphine index by 9.104 mg/BMI (Table 5). Intracranial surgery as well as thoracic, abdominal and pelvic surgery were associated with an increased morphine index compared to surgery involving the extremities.

There are several limitations applying to our study. Concerning its post hoc design, an adequate power analysis and sample size calculation was not performed. Secondly, chronic beta-blockade was only defined by its preoperative existence, not by its duration before surgery. We also did not record the total duration of the chronic BB treatment. Compared to other studies, different pharmacokinetics between oral intake in chronic blockade versus intravenous application in acute beta-blockade may result in varying bioavailability. Aggregation of perioperatively administered opioids into a morphine equivalent dose without taking into consideration that pharmacokinetic differences exist between different opioids must be considered the greatest limitation. We also exluded patients who received regional anesthesia and did not analyse alternate multimodal strategies for opioid reduction in our study, as these were too heterogenous between the different types of surgery and the two campuses of which data were analysed. We might also missed other confounding factors regarding analgesic therapy, e.g. influence of non-steroidal anti-inflammatory drugs, which might have been not used in patients with vascular disease, leading to a higher use in patients with BB obscuring the effect of the latter. There was no study specific protocol for intra- and postoperative pain management due to the retrospective nature of the post hoc analysis. Due to the limited sample size, it could be possible that an effect of a chronic BB therapy on opioid consumption might not have been detected in our cohort. Therefore, prospective trials with larger sample sizes are needed to investigate this question further. Finally, there was a great amount of heterogeneity regarding types of surgery included resulting in a heterogenous study group as well as different BB that were taken in the chronic BB group. We did not collect information about the exact indication for chronic BB intake as well as about the proportion of patients who complied to the recommendations concerning perioperative BB intake. We did not analyze the effect of individual risk factors on outcome as sample size was too small. Also, we cannot for sure certify the comparability of similar procedures, i.e. procedure per specialism.