Abstract
Objective:
To determine the association between preoperative benzodiazepine and nonbenzodiazepine receptor agonist (“Z-drugs”) use and adverse outcomes after surgery.
Background:
Prescriptions for benzodiazepines and Z-drugs have increased over the past decade. Despite this, the association of preoperative benzodiazepines and Z-drug receipt with adverse outcomes after surgery is unknown.
Methods:
Using the Optum Clinformatics Datamart, we performed a retrospective cohort study of adults 18 years or older who underwent any of 10 common surgical procedures between 2010 and 2015. The principal exposure was one or more filled prescriptions for a benzodiazepine or Z-drug in the 90 days before surgery. The primary outcome was any emergency department visit or hospital admission for either (1) a drug related adverse medical event or overdose or (2) a traumatic injury in the 30 days after surgery.
Results:
Of 785,346 patients meeting inclusion criteria, 94,887 (12.1%) filled a preoperative prescription for a benzodiazepine or Z-drug. From multivariable logistic regression, benzodiazepine or Z-drug use was associated with an increased odds of an adverse postoperative event [odds ratio 1.13; 95% confidence interval: 1.08–1.18). In a separate regression, coprescription of benzodiazepines or Z-drugs with opioids was associated with a 1.45 odds of an adverse postoperative event (95% confidence interval: 1.37–1.53).
Conclusions:
Preoperative benzodiazepines and Z-drug use is common and associated with increased odds of adverse outcomes after surgery, particularly when coprescribed with opioids. Counseling on appropriate benzodiazepine and Z-drug use in advance of elective surgery may potentially increase the safety of surgical care.
Keywords: benzodiazepines, opioids, perioperative outcomes, pharmacoepidemiology
Approximately 17.5 million, or 6% of adults in the United States use benzodiazepines or nonbenzodiazepine receptor agonists (“z-drug,” eg, Zolpidem).1,2 A total of 13.5 million adults fill a prescription for a benzodiazepine each year for anxiety, insomnia, and other indications,3 and an additional 4 million fill a prescription for a Z-drug for sleep disorders.4 Both benzodiazepines and Z-drugs carry risks of harm due to oversedation,5,6 particularly when taken in combination with opioids7 Notably, approximately 17% of individuals who routinely use opioids also receive prescriptions for benzodiazepines.1,7
At present, little is known about the safety implications of benzodiazepine or z-drug use among surgical patients. The perioperative period involves exposure to factors that have the potential to interact with sedative-hypnotics, including new opioid prescriptions for acute surgical pain8 and anesthetic medications.9,10 Although past research has established that preoperative opioid use is associated with an increased risk of readmission and other adverse events after surgery,11,12 few data exist to characterize patterns of benzodiazepine and Z-drug use among patients who receive opioids prior to surgery or the association of such co-prescribing with outcomes in this population. At present, guidelines recommend preoperative tapering or discontinuation of benzodiazepines among older adults as a means of reducing the risk of delirium.13,14 From the standpoint of clinical care and health policy, characterizing the safety of benzodiazepines and Z-drugs in mixed surgical populations with and without prior receipt of opioids can help to characterize the potential value of expanding recommendations for benzodiazepine and Z-drug deprescribing before surgery to additional populations and contexts.
The present study had 3 main objectives. First, we sought to characterize patterns of benzodiazepine and Z-drug prescribing before surgery among a sample of US adults who underwent any of 10 common outpatient or short-stay inpatient surgical procedures; next, we aimed to quantify the association between preoperative benzodiazepine and Z-drug use and adverse safety events in the first 30 days after a surgical procedure and compare the magnitude of this association across different types of benzodiazepines and Z-drugs. Finally, we sought to compare risks of adverse safety events associated with benzodiazepines or Z-drug receipt with and without concurrent preoperative opioid receipt to determine whether there were adverse interactions between these drugs.
METHODS
Overview and Study Sample
We conducted a retrospective cohort study to examine the risk of adverse postoperative outcomes among patients who filled a benzodiazepine or Z-drug prescription before surgery. We used data from the Optum Clinformatics Datamart, a US health insurance database with >15 million enrollees annually. The database is geographically diverse representing all US states. Compared to the typical insured US population,15 Optum has a lower proportion of older adults and is less racially and ethnically diverse, with fewer individuals of black race and Hispanic ethnicity, but has comparable annual salary and education level. This study was exempted from review by the institutional review board of the University of Pennsylvania.
The study sample included patients 18 years of age and older who underwent any of 10 common general or orthopedic surgical procedures between January 1, 2010 and September 30, 2015. Eligible procedures were identified based on Current Procedural Terminology codes in physician claims (Appendix 1, http://links.lww.com/SLA/B757). For patients with more than 1 eligible surgery during the window, we used the first procedure; patients with claims for more than 1 eligible procedure on the same day were excluded from the sample. Finally, to permit a uniform window for assessment of preoperative prescribing and comorbidities, we restricted the sample to patients who had at least 3 months of continuous enrollment before the procedure date or admission date for the hospitalization associated with the procedure, whichever came first. We chose a 3-month lookback to minimize exclusions from our sample due discontinuities in insurance program enrollment over time; however, to assess the robustness of our findings to alternate approaches for assessing comorbidities, we repeated our main analysis in a subgroup with at least 6 months of continuous enrollment before the index date.
Exposure: Preoperative Benzodiazepine or Z-drug Receipt
We collected data on all filled prescriptions for a benzodiazepine or z-drug within 90 days before the day of surgery or the hospital admission for surgery consistent with time windows used for the assessment of preoperative prescribing of opioids.11 Benzodiazepines evaluated included chlordiazepoxide, diazepam, flurazepam, clorazepate, quazepam, alprazolam, clonazepam, lorazepam, oxazepam, temazepam, estazolam, midazolam, and triazolam. Z-drugs evaluated included zaleplon, zolpidem, and eszopiclone.
We defined our primary exposure as one or more filled prescription for any benzodiazepine or z-drug within 90 days before the day of surgery or hospital admission. We refrained from calculating milligram dosage equivalents for prescribed benzodiazepines and Z-drugs because available dose-conversion tables for benzodiazepines and Z-drugs are inconsistent and not comprehensive.16,17
Outcome
We aimed to capture a range of adverse events, including but not limited to documented drug overdose that we reasoned could be influenced by the sedating or respiratory effects of benzodiazepines or Z-drugs. Of specific interest were events that were significant enough to lead to an emergency department (ED) visit or hospital admission (Appendix 2, http://links.lww.com/SLA/B757). We assessed 3 categories of events: (a) documented benzodiazepine or opioid overdose or poisoning; (b) any of 19 other potentially drug-related adverse events such as aspiration, rhabdomyolysis, or drug-induced delirium (Appendix 2, http://links.lww.com/SLA/B757); and (c) traumatic injuries such as a fracture or dislocation. Each of these events were based on International Classification of Diseases 9th Revision, Clinical Modification (ICD-9-CM) diagnosis codes from either ED visit or hospitalization claims.18 We used validated algorithms for opioid poisoning and overdose (positive predictive value ~81%201385%).19 As there is no validated approach to defining sedative overdose, we used ICD-9-CM codes that specified poisoning by a benzodiazepine. To identify a set of adverse medical events, we adapted the algorithm of Green and colleagues for identifying adverse events related to opioid use.19 Finally, codes for traumatic injuries were based on established algorithms with positive predictive values of ~91–95%.20,21 All codes used in our outcome definitions are listed in Appendix 2, http://links.lww.com/SLA/B757.
Our primary outcome was a composite indicating occurrence of any of the 3 above categories of adverse events within 30 days after discharge for a surgical procedure. We further assessed each outcome category separately. Due to the low numbers of overdoses, we grouped overdoses with potential drug-related medical events for purposes of analysis. Outcomes were also assessed between 31 to 60 days and between 61 to 90 days after surgery.
Covariates
Baseline characteristics were ascertained in the 90 days prior to the index surgical procedure and included demographics, antidepressant and opioid exposure, and selected comorbidities. We defined baseline comorbidities using all ICD-9-CM inpatient and outpatient codes in the 90 days before surgery. We selected conditions that were either known to be associated with benzodiazepine (eg, depression and alcohol abuse) or that were known predictors of adverse outcomes after surgery (eg, renal and congestive heart failure).
We collected information on all filled prescriptions from pharmacy claims data for (1) antidepressant medications as listed in Appendix 3, http://links.lww.com/SLA/B757 and (2) any the following opioids: fentanyl, hydrocodone, hydromorphone, methadone, morphine, oxymorphone, oxycodone, and codeine, excluding opioid-decongestant combination products used for cough and cold treatments. We used standard conversion tables to calculate the morphine milligram equivalents (MMEs) dispensed for each filled opioid prescription and across all filled opioid prescriptions over the assessment interval.22
Statistical Analysis
We compared preoperative users and nonusers of benzodiazepines and Z-drugs with respect to patient characteristics, type of surgical procedure, and medication prescribing variables using standardized mean differences for continuous variables and standardized differences in proportion for categorical variables.
We used multivariable logistic regression with traditional covariate adjustment to control for potential confounding in the association between preoperative benzodiazepine or Z-drug receipt and the composite outcome. The dependent variable for our primary regression was an indicator variable set as 1 if the patient filled a prescription for a benzodiazepine or Z-drug in the 90 days before surgery and otherwise it was set to 0. We adjusted for covariates shown in Table 1. Preoperative opioid receipt (as a dichotomous variable), age (in 6 categories), procedure type, depression, and antidepressant use were included in the model a priori.
TABLE 1.
Baseline Characteristics of 785,346 Patients undergoing Common Surgical Procedures in the United States by Preoperative Benzodiazepine or Z-drug Use
| Variable, n (%) or Median (IQR) | Benzodiazepine or Z-drug Prescription (n = 94,887) | No Benzodiazepine or Z-drug Prescription (n = 690,459) | Standardized Difference |
|---|---|---|---|
| Sex | |||
| Female | 65,043 (68.5%) | 380,719 (55.1%) | 0.28 |
| Male | 29,844 (31.5%) | 309,740 (44.9%) | 0.28 |
| Race | |||
| White | 75,091 (79.1%) | 516,621 (74.8%) | 0.10 |
| Black | 7583 (8.0%) | 59,013 (8.5%) | 0.02 |
| Hispanic | 6050 (6.4%) | 58,161 (8.4%) | 0.08 |
| Asian | 1067 (1.1%) | 14,324 (2.1%) | 0.08 |
| Unknown/Missing | 5096 (5.4%) | 42,340 (6.1%) | 0.03 |
| Age, yr | |||
| 18–35 | 6839 (7.2%) | 96,005 (13.9%) | 0.22 |
| 36–45 | 11,787 (12.4%) | 85,702 (12.4%) | 0.00 |
| 46–55 | 23,244 (24.5%) | 130,093 (18.8%) | 0.14 |
| 56–65 | 28,207 (29.7%) | 147,157 (21.3%) | 0.19 |
| 66–75 | 17,088 (18.0%) | 147,839 (21.4%) | 0.09 |
| ≥76 | 7722 (8.1%) | 83,663 (12.1%) | 0.13 |
| Procedure type | |||
| Laparoscopic cholecystectomy | 21,184 (22.3%) | 148,886 (21.6%) | 0.02 |
| Open cholecystectomy | 1049 (1.1%) | 7303 (1.1%) | 0.00 |
| Laparoscopic appendectomy | 4034 (4.3%) | 45,118 (6.5%) | 0.10 |
| Open appendectomy | 484 (0.5%) | 5309 (0.8%) | 0.03 |
| Inguinal hernia repair | 4795 (5.1%) | 58,158 (8.4%) | 0.13 |
| Carpal tunnel release | 10,500 (11.1%) | 65,198 (9.4%) | 0.05 |
| Knee arthroscopy | 20,652 (21.8%) | 159,420 (23.1%) | 0.03 |
| Total knee replacement | 15,069 (15.9%) | 100,884 (14.6%) | 0.04 |
| Total hip replacement | 7734 (8.2%) | 54,921 (8.0%) | 0.01 |
| Breast excision | 9720 (10.2%) | 47,564 (6.9%) | 0.12 |
| Comorbidities | |||
| Heart failure | 3684 (3.9%) | 22,020 (3.2%) | 0.04 |
| Cardiac arrhythmia | 11,310 (11.9%) | 75,215 (10.9%) | 0.03 |
| Valvular heart disease | 4800 (5.1%) | 29,183 (4.2%) | 0.04 |
| Peripheral vascular disorders | 4006 (4.2%) | 25,876 (3.7%) | 0.02 |
| Hypertension: uncomplicated | 48,741 (51.4%) | 304,398 (44.1%) | 0.15 |
| Hypertension: complicated | 4159 (4.4%) | 27,987 (4.1%) | 0.02 |
| Neurological disorders | 3469 (3.7%) | 12,910 (1.9%) | 0.11 |
| Chronic pulmonary disease | 17,256 (18.2%) | 84,380 (12.2%) | 0.17 |
| Diabetes uncomplicated | 15,875 (16.7%) | 105,118 (15.2%) | 0.04 |
| Diabetes complicated | 3480 (3.7%) | 22,143 (3.2%) | 0.03 |
| Hypothyroidism | 15,667 (16.5%) | 83,326 (12.1%) | 0.13 |
| Renal failure | 3894 (4.1%) | 27,051 (3.9%) | 0.01 |
| Liver disease | 6983 (7.4%) | 39,877 (5.8%) | 0.06 |
| Solid tumor without metastasis | 10,644 (11.2%) | 50,893 (7.4%) | 0.13 |
| Rheumatoid arthritis/collagen disorders | 5797 (6.1%) | 23,984 (3.5%) | 0.12 |
| Obesity | 14,222 (15.0%) | 95,242 (13.8%) | 0.03 |
| Weight loss | 2245 (2.4%) | 9679 (1.4%) | 0.07 |
| Fluid and electrolyte disorders | 8245 (8.7%) | 45,695 (6.6%) | 0.08 |
| Alcohol abuse | 1646 (1.7%) | 6495 (0.9%) | 0.07 |
| Drug abuse | 2234 (2.4%) | 5122 (0.7%) | 0.13 |
| Psychoses | 1380 (1.5%) | 3059 (0.4%) | 0.10 |
| Depression | 25,742 (27.1%) | 61,284 (8.9%) | 0.49 |
| Any antidepressant prescription | 42,814 (45.1%) | 95,780 (13.9%) | 0.73 |
| Opioid prescribing before surgery | |||
| Any opioid prescription | 53,367 (56.2%) | 199,721 (28.9%) | 0.57 |
| Days of opioid therapy | 4.0 (0.0–31.0) | 0.0 (0.0–3.0) | 0.76 |
| Total morphine milligram equivalents | 675 (263–2160) | 300 (150–788) | 0.33 |
IQR indicates interquartile range.
We performed further analysis to test the robustness of the association between benzodiazepine or Z-drug receipt and postoperative adverse events. As the half-life of benzodiazepines may correlate with risk,23 we categorized usage based on the duration of action of the benzodiazepine or Z-drug (very short-acting, ie, Z-drugs, short-or intermediate-acting and long-acting, Appendix 4, http://links.lww.com/SLA/B757) and performed an additional logistic regression model with this categorical variable as the exposure of interest. To determine whether the duration of benzodiazepine or Z-drug therapy was associated with increased risk, we repeated the logistic regression model with the continuous variable corresponding to the total days of supply of benzodiazepine or Z-drug prescribed in the 90 days before surgery in place of our binary exposure variable. Finally, as older adults are both more likely to use benzodiazepines and Z-drugs24 and be at greater risk of adverse outcomes,25 we tested for the potential of effect modification by age categories (<65 years and ≥65 years).
As concurrent use of benzodiazepines and opioids have been shown to increase the risk of fatal overdoses compared to each drug individually,26 we also tested for a multiplicative statistical interaction between benzodiazepine or Z-drug and opioid exposure.27,28 Specifically, we first performed logistic regression to estimate the separate and joint effects of these exposures. We created an indicator variable with a reference category of patients that did not use a benzodiazepine or Z-drug or opioid, and then other categories for those who used only sedatives, only opioids, or both. We then evaluated for interaction on the multiplicative scale by including an interaction term between preoperative benzodiazepine or Z-drug and opioid use.
Sensitivity Analysis
As some prescriptions filled before surgery may only be intended for postoperative use, we redefined preoperative benzodiazepine or Z-drug use as having a filled prescription more than 7 days before surgery and then repeated our main regression. We also repeated our primary analysis with opioid prescription as a continuous variable of the total MME prescribed in the 90 days before surgery modeled as a linear and quadratic term.
To determine the robustness of our primary association to unmeasured or uncontrolled confounding, we calculated an E value for the association between benzodiazepine or Z-drug use and adverse events after surgery.29 The E value is the minimum strength of association, on the relative scale, that a confounder would need to have with both the exposure and the outcome to fully explain away the association between benzodiazepine or Z-drug use and adverse events after surgery.
Analyses were conducted with STATA statistical software version 15 (StataCorp, College Station, TX).
RESULTS
The study sample included 785,346 patients; 75.3% were white, 56.8% were women, and the median age was 57 (interquartile range 45–69). The most common surgical procedures were knee arthroscopy (22.9%) and laparoscopic cholecystectomy (21.7%). Of the cohort, 94,887 (12.1%) filled a prescription for a benzodiazepine or Z-drug in the 90 days before surgery, the majority for either an intermediate-acting benzodiazepine (55.5%) or Z-drug (35.9%). Of these, 11.3% of patients filled a prescription for more than 1 type of sedative in the 90 days before surgery.
Selected baseline characteristics of the study cohort by preoperative benzodiazepine or Z-drug prescription are presented in Table 1. A concurrent preoperative prescription for an opioid was more likely in patients who were on a benzodiazepine or Z-drug (56.2%) compared to patients not on a benzodiazepine or Z-drug (28.9%); P < 0.001. The amount of opioid in MME over the 90 days before surgery was also significantly higher in those receiving benzodiazepines or Z-drugs before surgery compared to those not receiving such medications (675 vs 300mg, P < 0.001).
Association Between Preoperative Benzodiazepine or Z-drug Prescription and Adverse Postoperative Outcomes
In the primary analysis, a total of 18,485 patients (2.35%) experienced the composite outcome of an overdose, potential drug-related medical event, or traumatic injury in the 30 days after surgery. Of note, the overall incidence of this composite outcome in our sample decreased with increasing time from surgery, and was 1.38% from 31 to 60 days after surgery and 1.27% from 61 to 90 days after surgery.
Table 2 shows the association between preoperative benzodiazepine or Z-drug receipt and the study outcomes. The absolute risk increase for benzodiazepine or Z-drug receipt on the composite outcome was 0.93% with a number needed to harm of 107.5. After adjustment, any preoperative benzodiazepine or Z-drug receipt was associated with increased odds of the composite outcome [odds ratio (OR): 1.13; 95% confidence interval (CI): 1.08, 1.18]. The full multivariable regression model is shown in Appendix 5, http://links.lww.com/SLA/B757. A benzodiazepine or Z-drug prescription before surgery was also associated with an increased odds of potential drug-related medical events including overdoses (OR 1.18; 95% CI: 1.12–1.24) and traumatic injury (OR 1.08; 95% CI: 1.01–1.16). As there were only 61 cases of opioid overdoses and 32 cases of benzodiazepine overdoses, we did not perform an analysis of this outcome alone. Our findings were robust to multiple sensitivity analyses. We observed similar findings when we repeated our analysis using a continuous exposure variable indicating total days of sedative hypnotic therapy; with adjustment for the total MME of opioid received in the 90 days before surgery (either as a linear or quadratic term); with adjustment only for opioids prescriptions filled before 7 days preoperatively; and with using a 6-month look back for comorbidities (data not shown). The E value for the primary analysis of the composite outcome was 1.49 (lower CI, 1.36). Lastly, there was no difference in the primary association stratified by lower extremity orthopedic surgeries (total knee/hip replacement, knee arthroscopy) versus all other surgical procedures.
TABLE 2.
Postoperative Adverse Events in the 30 Days after Surgery of 785,346 Patients Undergoing Common Surgical Procedures in the United States by Preoperative Benzodiazepine or Z-drug Prescription
| Outcome, n (%) | Benzodiazepine or Z-drug Prescription (n = 94,887) | No Benzodiazepine or Z-drug Prescription (n = 690,459) | Unadjusted OR (95% CI) | Adjusted OR (95% CI) |
|---|---|---|---|---|
| Composite outcome | 3011 (3.17%) | 15,474 (2.24%) | 1.43 (1.37–1.49) | 1.13 (1.08–1.18) |
| Opioid/sedative adverse event or overdose | 2126 (2.24%) | 10,605 (1.54%) | 1.47 (1.40–1.54) | 1.18 (1.12–1.24) |
| Traumatic injury | 1152 (1.21%) | 5943 (0.86%) | 1.42 (1.33–1.51) | 1.08 (1.01–1.16) |
ORs adjusted for age, sex, race/ethnicity, surgical procedure, comorbidities, opioid prescription, antidepressant prescription.
Full regression model shown in Appendix 5, http://links.lww.com/SLA/B757.
The odds of the composite outcome varied by type of benzodiazepine or Z-drug (Table 3). Compared to patients who did not fill a prescription for a benzodiazepine or Z-drug, a filled prescription for a Z-drug was associated with an increased adjusted odds of the composite outcome (OR 1.18; 95% CI: 1.11–1.26). With regard to benzodiazepines, both short- or intermediate-acting (OR 1.11; 95% CI: 1.95–1.17) and long-acting (OR 1.05; 95% CI: 0.91–1.20) drugs had increased adjusted odds of the composite outcome, although the result was not statistically significant for long-acting benzodiazepines. Compared to patients filling no benzodiazepine or Z-drug prescriptions before surgery, the adjusted OR for the composite outcome among patients filling prescriptions for more than 1 type of benzodiazepine or Z-drug was 1.24 (95% CI: 1.11–1.38) and 1.12 (95% CI: 1.07–1.17) among those who filled a prescription for only 1 type of benzodiazepine or Z-drug.
TABLE 3.
Odds Ratio for Preoperative Benzodiazepine or Z-drug Prescription on Composite Outcome Adjusted for Covariates
| Total N | N (%) experiencing composite outcome | Unadjusted OR (95% CI) | Adjusted OR (95% CI) | |
|---|---|---|---|---|
| Multivariable regression for benzodiazepine or Z-drug class | ||||
| No prescription | 690,459 | 15,474 (2.24) | Reference | |
| Z-drug | 35,324 | 1180 (3.34) | 1.51 (1.42–1.60) | 1.18 (1.11–1.26) |
| Short-/intermediate-acting benzodiazepine | 49,961 | 1597 (3.20) | 1.44 (1.37–1.52) | 1.11 (1.05–1.17) |
| Long-acting benzodiazepine | 9602 | 234 (2.44) | 1.09 (0.96–1.24) | 1.05 (0.91–1.20) |
| Multivariable regression for number of concurrent benzodiazepine or Z-drug prescriptions | ||||
| No prescription | 690,459 | 15,474 (2.24) | Reference | |
| Prescription for only 1 class | 84,042 | 2636 (3.14) | 1.41 (1.35–1.47) | 1.12 (1.07–1.17) |
| Prescription for >1 class | 10,845 | 375 (3.46) | 1.56 (1.41–1.73) | 1.24 (1.11–1.38) |
ORs adjusted for age, sex, race/ethnicity, surgical procedure, comorbidities, opioid prescription, antidepressant prescription.
Finally, we found that the association of receipt of a preoperative benzodiazepine or Z-drug with the composite outcome differed by patient age. On the relative scale, there was a greater magnitude of association in patients aged less than 65 years (OR 1.19; 95% CI: 1.12–1.26) compared to patients 65 years or older (OR 1.08, 95% CI 1.02–1.14; P value for interaction 0.02; Appendix 6, http://links.lww.com/SLA/B757).
Concurrent Benzodiazepine or Z-drug and Opioid Prescriptions and the Risk of Adverse Postoperative Outcomes
Baseline characteristics of patients with individual and concurrent prescriptions for benzodiazepines or Z-drugs and opioids are shown in Appendix 7, http://links.lww.com/SLA/B757. Table 4 shows the results of the joint analysis of the effects of having a preoperative prescription for a benzodiazepine or Z-drug and opioid. A prescription for a benzodiazepine or Z-drug without a concurrent opioid prescription had an OR of 1.07 (95% CI: 0.99–1.14). A prescription for an opioid without a concurrent benzodiazepine or Z-drug prescription had an OR of 1.23 (95% CI: 1.19–1.28). Having a prescription for both a benzodiazepine or Z-drug and opioid had an OR of 1.45 (95% CI: 1.37–1.53). There was a positive interaction on the multiplicative scale [OR 1.1 (95% CI 1.01–1.20) P = 0.02], indicating that coreceipt of benzodiazepines or Z-drugs and opioids was associated with a modest increase in the adjusted odds of an adverse outcome above and beyond the association attributable to either medication alone.
TABLE 4.
Interaction Between Benzodiazepine or Z-drug and Opioid Prescription on the Risk of an Adverse Postoperative Outcome
| Benzodiazepine or Z-drug Prescription in the 90 Days Before Surgery | Opioid Prescription in the 90 Days Before Surgery | Total N | N (%) Meeting Composite Outcome | Unadjusted OR (95% CI) | Adjusted OR (95% CI) |
|---|---|---|---|---|---|
| No | No | 490,738 | 9933 (2.02%) | Reference | |
| Yes | No | 41,520 | 981 (2.36%) | 1.17 (1.10–1.25) | 1.07 (0.99–1.14) |
| No | Yes | 199,721 | 5541 (2.77%) | 1.38 (1.34–1.43) | 1.23 (1.19–1.28) |
| Yes | Yes | 53,367 | 2030 (3.80%) | 1.91 (1.82–2.01) | 1.45 (1.37–1.53) |
Measure of interaction on the multiplicative scale: OR 1.1 (95% CI 1.01–1.20) P= 0.02.
ORs adjusted for age, sex, race/ethnicity, surgical procedure, comorbidities, opioid prescription, antidepressant prescription.
DISCUSSION
In this sample of 785,346 US adults who had 1 of 10 common surgical procedures between 2010 and 2015, 1 in 8 patients filled a prescription for a benzodiazepine or a Z-drug in the 90 days before surgery, and 1 in 15 patients filled a prescription for both a benzodiazepine or Z-drug and an opioid. Patients who filled any prescription for a benzodiazepine or Z-drug in the 90 days before surgery had a 13% greater odds of an ED visit or hospitalization for an overdose, drug-related medical event, or traumatic injury within the first 30 days after discharge. We found both benzodiazepines and z-drugs to be associated with adverse outcomes after surgery and observed the association between preoperative benzodiazepine or Z-drug use and adverse postoperative outcomes to be greater among patients younger than 65 years and among those with concurrent preoperative opioid prescriptions.
To our knowledge, this is the first national analysis of either the extent of benzodiazepine or Z-drug use before common general and orthopedic surgical procedures, or the association between preoperative benzodiazepine or Z-drug receipt and adverse postoperative outcomes. Current recommendations regarding preoperative benzodiazepine or Z-drug use focus primarily on prevention of delirium after surgery among older adults13,14; our work expands on this literature to highlight potential risks associated with the use of such medications in both older and younger populations across a broad range of outcomes not limited to delirium. In addition, our finding of significant associations between Z-drug receipts and adverse events after surgery emphasizes that risks associated with sedative-hypnotic use before surgery may not be restricted to benzodiazepines alone.
Of particular note is our finding of a significant interaction between preoperative benzodiazepine or Z-drug receipt and preoperative opioid receipt with regard to their association with adverse postoperative outcomes. More than half of all of the patients in our sample who received benzodiazepine or Z-drug medications before surgery also filled a preoperative opioid prescription, and the rate of preoperative opioid receipt was nearly twice as high among patients receiving benzodiazepines or Z-drugs compared to other patients. These findings mirror prior research indicating substantially higher rates of prevalent benzodiazepine and Z-drug use among chronic opioid users compared to the rest of the population.1,7 Our finding that concurrent prescribing of benzodiazepines and Z-drugs with opioids is associated with increased risks of postoperative adverse events compared to receipt of either type of medication alone highlights such coprescribing as a potential contributor to previously observed associations between preoperative opioid use and adverse outcomes,11,12 and a target for future efforts to improve care. Moreover, the relatively low adjusted OR for benzodiazepines and Z-drugs in isolation suggest that interventions to reduce the harm of these drugs should be focused on patients with concomitant opioid use rather than on patients who only use benzodiazepines or Z-drugs.
Our study has limitations. First, although our dataset captured several important potential confounders, we cannot rule out the possibility that our findings may be explained by differences in unmeasured baseline characteristics such as socioeconomic status and the level of social support between patients who did and did not receive benzodiazepines or Z-drugs before surgery. Based on our sensitivity analysis, a potential unmeasured confounder would have to be associated with benzodiazepine or Z-drug use and the composite outcome by an OR of greater than 1.49 to explain away our primary association. And although our dataset does not contain reliable measures of socioeconomic status such as income, the association between income and benzodiazepine and Z-drug use is weak and unlikely to have changed the point estimate of our primary association.30 We took additional effort to control for the indications of benzodiazepine or Z-drug use by adjusting for both psychiatric diagnoses such as depression and psychoses and preoperative use of antidepressants. Moreover, as our data only captured filled prescriptions, we were unable to formally quantify the amount or frequency of benzodiazepine, Z-drug, or opioid use in our population. Lastly, our findings may not be generalizable to major surgery and nonelective procedures.
Despite this, our results have important implications for perioperative clinical practice and policy. For patients with chronic opioid use, national and state clinical practice guidelines state that opioids should not be prescribed in combination with benzodiazepines due to the significant increase in risk of harm, including death, when these 2 drugs are used in combination. Our findings highlight potential risks associated with receipt of benzodiazepines and other sedative-hypnotics in the perioperative period, particularly in combination with preoperative opioid prescriptions. As such, it provides novel information that may be used to inform discussions with individual patients and offer counseling regarding the risks and benefits of continued sedative-hypnotic use prior to surgery. In addition, available recommendations related to sedative-hypnotic medication management before surgery now focus primarily on benzodiazepines.14 Our finding of significant associations between receipt of Z-drugs with adverse outcomes highlights needs for further research on the potential risks associated with use of these medications in the perioperative period to inform clinical and policy interventions to improve postoperative outcomes.
CONCLUSIONS
Among adults who undergo common elective surgical procedures in the United States, preoperative use of a benzodiazepine or Z-drug was associated with an increased the risk of postoperative adverse events, especially with concurrent use of opioids.
Supplementary Material
Acknowledgments
This study was supported by a grant from the National Institute on Drug Abuse (1R01DA042299-01A1).
C.E.L. serves on the Executive Committee of the University of Pennsylvania’s Center for Pharmacoepidemiology Research and Training and the center receives funding from Pfizer and Sanofi, and his spouse is employed by a for-profit health technology company that receives funding from AbbVie, Adamas, Celgene, Lilly, Lundbeck, Novartis, and Sunovion. B.T.B. is an investigator on grants to his institution from Baxalta, Lilly, GSK, Pfizer, and Pacira for unrelated studies and is a consultant to the Alosa Foundation and Aetion, Inc. S.H. and has consulted for Braeburn Pharmaceuticals, Inc; Daiichi Sankyo, Inc; Egalet Corporation; Indivior, Inc; Inspiron Delivery Sciences, LLC; Nektar Therapeutics Inc; and Purdue Pharma, LP. The remaining authors report no conflicts of interest.
Footnotes
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.annalsofsurgery.com).
REFERENCES
- 1.Lembke A, Papac J, Humphreys K. Our other prescription drug problem. N Engl J Med. 2018;378:693–695. [DOI] [PubMed] [Google Scholar]
- 2.Kaufmann CN, Spira AP, Alexander GC, et al. Trends in prescribing ofsedative-hypnotic medications in the USA: 1993–2010. Pharmacoepidemiol Drug Saf. 2016;25:637–645. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bachhuber MA, Hennessy S, Cunningham CO, et al. Increasing benzodiazepine prescriptions and overdose mortality in the United States, 1996–2013. Am J Public Health. 2016;106:686–688. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Bertisch SM, Herzig SJ, Winkelman JW, et al. National use of prescription medications for insomnia: NHANES 1999–2010. Sleep. 2014;37:343–349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Dassanayake T, Michie P, Carter G, et al. Effects of benzodiazepines, antidepressants and opioids on driving: a systematic review and meta-analysis of epidemiological and experimental evidence. Drug Saf. 2011;34: 125–156. [DOI] [PubMed] [Google Scholar]
- 6.Gunja N In the Zzz zone: the effects of Z-drugs on human performance and driving. J Med Toxicol. 2013;9:163–171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Sun EC, Dixit A, Humphreys K, et al. Association between concurrent use of prescription opioids and benzodiazepines and overdose: retrospective analysis. BMJ. 2018;356:j760. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wunsch H, Wijeysundera DN, Passarella MA, et al. Opioids prescribed after low-risk surgical procedures in the United States, 2004–2012. JAMA. 2016;315:1654–1657. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sieber FE, Zakriya KJ, Gottschalk A, et al. Sedation depth during spinal anesthesia and the development of postoperative delirium in elderly patients undergoing hip fracture repair. Mayo Clin Proc. 2010;85:18–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Eger EI 2nd, Bowland T, Ionescu P, et al. Recovery and kinetic characteristics of desflurane and sevoflurane in volunteers after 8-h exposure, including kinetics of degradation products. Anesthesiology. 1997;87:517–526. [DOI] [PubMed] [Google Scholar]
- 11.Waljee JF, Cron DC, Steiger RM, et al. Effect of preoperative opioid exposure on healthcare utilization and expenditures following elective abdominal surgery. Ann Surg. 2017;265:715–721. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Cron DC, Englesbe MJ, Bolton CJ, et al. Preoperative opioid use is independently associated with increased costs and worse outcomes after major abdominal surgery. Ann Surg. 2017;265:695–701. [DOI] [PubMed] [Google Scholar]
- 13.Chow WB, Rosenthal RA, Merkow RP, et al. Optimal preoperative assessment of the geriatric surgical patient: a best practices guideline from the American College of Surgeons National Surgical Quality Improvement Program and the American Geriatrics Society. J Am Coll Surg. 2012;215:453–466. [DOI] [PubMed] [Google Scholar]
- 14.American Geriatrics Society Expert Panel on Postoperative Delirium in Older Adults. American Geriatrics Society abstracted clinical practice guideline for postoperative delirium in older adults. J Am Geriatr Soc. 2015;63:142–150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Barnett JC, Vornovitskey MS. Health insurance coverage in the United States: 2015 Current population reports United States Census Bureau. Government Report; September 2016:1–35. Available at: https://census.gov/content/dam/Census/library/publications/2016/demo/p60-257.pdf. Accessed January 14, 2009. [Google Scholar]
- 16.Government of South Africa. Benzodiazepine equivalents. Available at: https://www.sahealth.sa.gov.au/wps/wcm/connect/83838b80407711959274ba222b2948cf/Benzodiazepine+equivalentsDASSA-August2014.pdf?MOD=AJPERES&CACHEID=ROOTWORKSPACE-83838b80407711959274ba222b2948cf-m08ay9k. Published August 2014. Accessed January 14, 2019.
- 17.NHS Foundation Trust. Calculating equivalent doses of oral benzodiazepines. Available at: http://www.sussexpartnership.nhs.uk/sites/default/files/documents/bdz_equivalent_doses_spt_guidance_update_-_0714.pdf. Updated July 2014. Accessed January 14, 2019.
- 18.Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43:1130–1139. [DOI] [PubMed] [Google Scholar]
- 19.Green CA, Perrin NA, Janoff SL, et al. Assessing the accuracy of opioid overdose and poisoning codes in diagnostic information from electronic health records, claims data, and death records. Pharmacoepidemiol Drug Saf. 2017;26:509–517. [DOI] [PubMed] [Google Scholar]
- 20.Langley J, Stephenson S, Thorpe C, et al. Accuracy of injury coding under ICD-9 for New Zealand public hospital discharges. Inj Prev. 2006;12:58–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sund R Quality of the Finnish Hospital Discharge Register: a systematic review. Scand J Public Health. 2012;40:505–515. [DOI] [PubMed] [Google Scholar]
- 22.Centers for Medicare and Medicaid. Opioid Oral Morphine Milligram Equivalent (MME) Conversion Factors. Available at: https://www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovContra/Downloads/Opioid-Morphine-EQ-Conversion-Factors-April-2017.pdf. Updated April 2017. Accessed January 14, 2019.
- 23.Ray WA, Thapa PB, Gideon P. Benzodiazepines and the risk of falls in nursing home residents. J Am Geriatr Soc. 2000;48:682–685. [DOI] [PubMed] [Google Scholar]
- 24.Olfson M, King M, Schoenbaum M. Benzodiazepine use in the United States. JAMA Psychiatry. 2015;72:136–142. [DOI] [PubMed] [Google Scholar]
- 25.Campanelli CM. American Geriatrics Society updated beers criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2012;60:616–631. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Hernandez I, He M, Brooks MM, et al. Exposure-response association between concurrent opioid and benzodiazepine use and risk of opioid-related overdose in medicare part D beneficiaries. JAMA Ne tw Open. 2018;1:e180919. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Knol MJ, VanderWeele TJ. Recommendations for presenting analyses of effect modification and interaction. Int J Epidemiol. 2012;41:514–520. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Vandenbroucke JP, von Elm E, Altman DG, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. Epidemiology. 2007;18:805–835. [DOI] [PubMed] [Google Scholar]
- 29.VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the E-value. Ann Intern Med. 2017;167:268–274. [DOI] [PubMed] [Google Scholar]
- 30.Cunningham CM, Hanley GE, Morgan S. Patterns in the use of benzodiazepines in British Columbia: examining the impact of increasing research and guideline cautions against long-term use. Health Policy. 2010;97:122–129. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.