Article Text

other Versions

Download PDFPDF
Association of perioperative midazolam use and complications: a population-based analysis
  1. Vassilis Athanassoglou1,
  2. Crispiana Cozowicz2,
  3. Haoyan Zhong3,
  4. Alex Illescas3,
  5. Jashvant Poeran4,
  6. Jiabin Liu3,5,
  7. Lazaros Poultsides6,7 and
  8. Stavros G Memtsoudis2,3,5,8
  1. 1Nuffield Department of Anaesthesia, Oxford University Hospitals NHS Foundation Trust, Oxford, UK
  2. 2Department of Anesthesiology, Perioperative Medicine and Intensive Care Medicine, Paracelsus Medical University, Salzburg, Austria
  3. 3Department of Anesthesiology, Critical Care & Pain Management, Hospital for Special Surgery, New York, New York, USA
  4. 4Institute for Healthcare Delivery Science, Department of Population Health Science & Policy / Orthopedics, Icahn School of Medicine at Mount Sinai, New York, New York, USA
  5. 5Department of Anesthesiology, Weill Cornell Medicine, New York, NY, USA
  6. 6Academic Orthopaedic Department, Papageorgiou General Hospital, Aristotle University of Thessaloniki, Thessaloniki, Greece
  7. 7Centre of Orthopaedics and Regenerative Medicine (C.O.RE.) - Centre of Interdisciplinary Research and Innovation (C.I.R.I.), Aristotle University, Thessaloniki, Greece
  8. 8Department of Health Policy and Research, Weill Cornell Medical College, New York, New York, USA
  1. Correspondence to Dr Stavros G Memtsoudis, Anesthesiology, Critical Care & Pain Management, Hospital for Special Surgery, New York, New York, USA; memtsoudiss{at}hss.edu

Abstract

Introduction The benzodiazepine midazolam is the main sedative used in the perioperative setting, resulting in anxiolysis and a reduction in anesthetic dose requirements. However, benzodiazepine use is also associated with potentially serious side effects including respiratory complications, and postoperative delirium (POD). A paucity of population level data exists on current perioperative midazolam use in adult orthopedic surgery and its effects on complications. Using a large national dataset, we aimed to determine perioperative midazolam utilization patterns and to analyze its effect on postoperative outcomes.

Methods Patients who underwent total knee and hip arthroplasty (TKA/THA) were identified from Premier database (2006–2019). Primary exposure of interest was midazolam use on the day of surgery. Multivariable logistic regression models were run to determine if midazolam was associated with postoperative cardiac and pulmonary complications, delirium, and in-hospital falls.

Results Among 2,848,897 patients, more than 75% received midazolam perioperatively. This was associated with increased adjusted odds for in-hospital falls in TKA/THA (OR 1.1, 95% CI 1.07 to 1.14)/(OR 1.1, 95% CI 1.06 to 1.16), while a decrease in the adjusted odds for cardiac complications in TKA/THA (OR 0.94, 95% CI 0.91 to 0.97)/(OR 0.93, 95% CI 0.89 to 0.97), and pulmonary complications (OR 0.92, 95% CI 0.87 to 0.96) (all p<0.001) was seen. Most notably, the concurrent use of midazolam and gabapentinoids significantly increased the adjusted odds for postoperative complications, including pulmonary complications (OR 1.22, 95% CI 1.18 to 1.27)/(OR 1.29, 95% CI 1.22 to 1.37), naloxone utilization (OR 1.56, 95% CI 1.51 to 1.60)/(OR 1.49, 95% CI 1.42 to 1.56), and POD (OR 1.45, 95% CI 1.38 to 1.52)/(OR 1.32, 95% CI 1.23 to 1.34) in THA/TKA.

Conclusion Perioperative midazolam use was associated with an increase in postoperative patient falls, and a decrease in cardiac complications. Notably, the combined use of midazolam and gabapentinoids was associated with a substantial increase in the odds for respiratory failure and delirium. Given the high prevalence of benzodiazepines perioperatively, the risk benefit profile should be more clearly established to inform perioperative decision making.

  • anesthesia
  • conduction
  • analgesia
  • epidemiology

Data availability statement

Data may be obtained from a third party and are not publicly available. Data obtained from a third party, Premier (Premier Healthcare Solutions, Charlotte, North Carolina, USA), and are not publicly available.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Preoperative anxiety is associated with adverse clinical outcomes such as increased postoperative pain and thus increased opiate consumption, as well as emergence delirium, and postoperative sleep disturbances.1 Therefore, many anesthesiologists use midazolam preoperatively with the goal to provide anxiolysis. In this context, anterograde amnesia is a welcome side-effect of benzodiazepines as decreased recall enhances patient tolerance for surgical and diagnostic procedures.2 In particular, short-acting benzodiazepines have been shown to allay anxiety, reduce anesthetic dose requirements, and reduce the incidence of postoperative nausea and vomiting without prolonging recovery time.3 4

Importantly, despite the reported popularity of perioperative benzodiazepine use, several investigations have questioned their anxiolytic benefits.5 6 Benzodiazepines are associated with serious side effects such as airway collapse, apnea, bradypnea, paradoxical reactions, and postoperative delirium (POD).7 In addition, not all patients require anxiolysis; therefore, putting into question on the reasoning for routine perioperative administration of benzodiazepines. Our group previously reported a possible synergistic effect of benzodiazepines and gabapentinoids for postoperative complications in arthroplasty surgery.8 The US Food and Drug Administration has recently cautioned against simultaneous prescribing of benzodiazepines and gabapentinoids9 suggesting a potentiating effect in terms of risk for respiratory depression and sedation, similar to coprescribing gabapentinoids with an opioid.10

Despite these concerns, a paucity of population level data exists on current benzodiazepine usage in the setting of adult orthopedic surgery and their effects on patient outcomes. We therefore used a large national dataset to (1) determine the perioperative utilization of midazolam in the setting of total hip and knee arthroplasty (THA and TKA) and (2) analyze the effect on postoperative outcomes and complications. We hypothesized that a majority of patients would receive midazolam perioperatively and that its use would be associated with increased odds for complications, especially when concurrently administered with gabapentinoids.

Methods

Study design and cohort

Patients undergoing inpatient elective TKA or THA from 2006 to 2019 were identified from the Premier Healthcare database (Premier Healthcare Solutions, Charlotte, North Carolina, USA).11 TKA was defined based on International Classification of Diseases 9th Revision (ICD-9) procedure code 81.54 or 10th Revision (ICD-10) procedure codes 0SRC0xx, 0SRD0xx. THA was defined based on ICD-9 procedures codes 81.51 or ICD-10 procedure codes 0SR90xx, 0SRB0xx. Exclusion criteria were unknown sex (n=322), unknown discharge status (n=1493), surgery at a hospital that performed fewer than 30 procedures (n=621), and age <18 (n=570).

Study variables

The primary exposure of interest was billing for midazolam use on the day of surgery. To reflect the current standard for perioperative anesthesia practice with the intention of routine anxiolysis, sedation, or hypnosis, benzodiazepine use was restricted to midazolam (which accounted for 85.9% of overall benzodiazepine use). In contrast, long-acting benzodiazepine utilization, is not common practice for the scope of perioperative anesthesia but may rather reflect interventions outside the anesthesiologist’s area of care such as treatment of other serious comorbidities, including anxiety disorders, seizures, insomnia, alcohol withdrawal syndrome, or other acute health conditions. Therefore, to eliminate any related potential source of bias, long-acting benzodiazepines were excluded from the current analysis which would require a framework with more longitudinal patient information.

Patients who received midazolam were identified by using billing descriptions.12

Outcomes of interests included cardiac complications, pulmonary complications, delirium, and in-hospital falls. All complications were defined using ICD-9 and ICD-10 diagnosis codes based on previous literature (online supplemental appendix 1).13–15

Supplemental material

According to the a priori established analysis plan, a number of baseline variables were investigated as potential modifiers of outcome, including patient-related, procedure-related and healthcare-related characteristics. Patient characteristics included age, sex, race (black, white, or other), Charlson-Deyo comorbidity index16 (categorized as 0, 1, 2, 3, and insurance provider (commercial, Medicaid, Medicare, uninsured, unknown). Procedure-related variables included type of anesthesia (general, neuraxial, or other/unknown) and year of surgery (2006–2019). Healthcare characteristics analyzed included location (urban vs rural), bed size (<300 beds, 300–500 beds, >500 beds), teaching status, and region (Midwest, Northeast, South, West).

Gabapentinoids

An additional exposure of interest was the concurrent perioperative use of gabapentinoids (including gabapentin, pregabalin) with midazolam. The risk of cardiac complications, pulmonary complications, delirium, and the use of naloxone, as a surrogate marker for respiratory failure was compared between patients with combined midazolam and gabapentinoid use and those with midazolam use only.

Statistical analysis

All statistical analyses were performed using SAS V.9.4 (SAS Institute). Analyses were performed separately for TKA and THA. Categorical variables were reported as absolute numbers and frequencies; continuous variables were reported as median and IQR. Given our large sample size, univariable differences between groups easily reach statistical significance; therefore, we evaluated univariable associations using standardized differences instead of p values. A standardized difference greater than 0.1 (or 10%) has been suggested to indicate a meaningful difference between groups.17

Multivariable logistic regression models were fitted to determine if midazolam use was significantly associated with the aforementioned outcomes after adjusting for all other covariates. ORs and 95% CIs were reported. Bonferroni adjustment was used to account for multiple hypotheses tested (16 hypotheses) for primary outcomes. Results with a p value less than 0.003 were considered statistically significant. For subgroup analyses, similar multivariable logistic regression models—adding the use of gabapentinoids—were employed to determine if combining gabapentinoids and midazolam was significantly associated with the studied complications of interest. A p value less than 0.005 was considered statistically significant (10 hypotheses).

Results

Patient characteristics

In the TKA cohort, patients who received midazolam on the day of surgery were somewhat younger, more frequently white, covered by commercial insurance, received neuraxial anesthesia more often, and were more frequently treated at smaller hospitals (>500 beds), and in the Midwest. Similar observations were made in the THA cohort (table 1).

Table 1

Summary of patient characteristics stratified by use of benzodiazepine on the day of surgery

Midazolam use in TKA patients

Among 1,863,841 TKA patients, 78.7% received midazolam on the day of surgery (table 1). After adjustment for available covariates (age, gender, race, Charlson-Deyo comorbidity index, insurance provider, type of anesthesia, year of surgery, hospital location, hospital teaching status, hospital size, and region), midazolam use was associated with a significant increase in the odds for in-hospital falls (adjusted OR 1.11, 95% CI 1.07 to 1.14, p<0.001), while a decrease in cardiac complications (adjusted OR 0.94, 95% CI 0.91 to 0.97, p<0.001) was seen. Interestingly the occurrence of POD did not differ by midazolam use (table 2).

Table 2

Risk of using benzodiazepine for complications following total joint arthroplasty surgery

Midazolam use in THA patients

Among 985,056 THA patients, 74.2% received midazolam on the day of surgery (table 1). After adjustment for available covariates, a significant increase in the odds of in-hospital falls was found with midazolam use (adjusted OR 1.11, 95% CI 1.06 to 1.16, p<0.001), while the odds decreased for cardiac (adjusted OR 0.93, 95% CI 0.89 to 0.97, p<0.001), and pulmonary complications (adjusted OR 0.92, 95% CI 0.87 to 0.96, p<0.001). Consistent with the TKA cohort, no effect was observed for the occurrence of delirium with the use of midazolam (table 2).

Subgroup analysis

A total of 489,487 (26.3%) TKA patients received midazolam concurrently with gabapentinoid medication. After adjustment for all covariates, combined use of gabapentinoids and midazolam as opposed to midazolam alone, the former was associated with a significant increase in pulmonary complications (adjusted OR 1.22, 95% CI 1.18 to 1.27), and the use of naloxone (adjusted OR 1.56, 95% CI 1.51 to 1.60). Notably, the concurrent use of gabapentinoids and midazolam also substantially increased the odds for POD (adjusted OR 0.1.45, 95% CI 1.38 to 1.52). A consistent pattern was observed in THA patients where n=255,921 (26.0%) received both drugs concurrently. Here, the odds for pulmonary complications (adjusted OR 1.29, 95% CI 1.22 to 1.37), naloxone administration (adjusted OR 1.49, 95% CI 1.42 to 1.56) as well as POD (adjusted OR 1.32, 95% CI 1.23 to 1.43) were also substantially increased (table 3).

Table 3

Risk of combining use of benzodiazepine and gabapentinoids for complications following total joint arthroplasty surgery

Discussion

National data indicates that three out of four patients undergoing joint arthroplasty surgery of the lower limb, receive midazolam perioperatively. The postoperative impact of benzodiazepine use, however, may significantly differ based on concurrent drug use and pharmacokinetics. While midazolam use was associated with a higher incidence of falls, no effect was found on the occurrence of POD. Moreover, the odds for cardiac and pulmonary complications were reduced among patients with midazolam use. In contrast, concurrent administration of midazolam and gabapentionoid medication, as commonly deployed in enhanced recovery after surgery protocols, was associated with a substantial increase in the odds for POD and respiratory failure. The high frequency of perioperative benzodiazepine use underlines the importance of considering benefit and harm for individual patients .

The effects of benzodiazepines are due to the potentiation of central neural inhibition mediated by gamma-aminobutyric acid. Main effects include sedation, hypnosis, decreased anxiety, anterograde amnesia, centrally mediated muscle relaxation, and anticonvulsant activity. Moreover, benzodiazepines appear to exhibit a dose-dependent ventilatory depressant effect and can also cause a decrease in systemic vascular resistance and arterial blood pressure, all of which are important in possibly explaining some postoperative outcomes. Midazolam is the most common benzodiazepine in clinical anesthesia, exhibiting a fast onset and offset profile. In addition to pharmacokinetic interactions, benzodiazepines are known to have synergistic effects with other hypnotics and opioids.18

The current population-based analysis demonstrates that in the United States, more than 75% of the patients undergoing TKA and THA, receive midazolam perioperatively. Thus, only minority of major orthopedic procedures are managed without benzodiazepines. Literature, however, suggests that preoperative benzodiazepine administration can depresses the central respiratory drive and cause hypoventilation,19 and pulmonary complications. The propensity to contribute to the risk for respiratory compromise also extends to short acting benzodiazepines, such as midazolam, based on the relatively long t1/2keo.7 As such, Gonzalez Castro et al19 demonstrated a dose dependent decrease in both minute ventilation and tidal volumes within 15 min of midazolam administration. Even with weight-based dosing, differences in sex, age, volume of distribution, fitness, and metabolic state patients appear to have differing effects from the same dose, placing some at a higher risk for adverse events and making it difficult for the clinician to predict who will develop them.20

In contrast, the current analysis indicates that the utilization of short acting-benzodiazepines may be associated with reduced odds for cardiac and pulmonary complications. Moreover, while the incidence of falls increased, the odds for POD was unaltered with the use of midazolam. Thus, these results add to the existing body of evidence that POD may not necessarily be associated with short-acting benzodiazepine use per se, but could involve additional factors.19 Indeed, Wang et al, also recently demonstrated that perioperative midazolam use was not associated with a higher incidence of POD in older patients.21

The American Geriatrics Society and American College of Critical Care Medicine both maintain the recommendation of minimizing perioperative benzodiazepine use to possibly prevent POD.22 23 Notably, the use of benzodiazepines has directly been implicated in POD after lower limb total joint arthroplasty.14 The current analysis, however, could not confirm a direct association with midazolam. Therefore, other augmenting drivers may need to be considered to establish an accurate risk–benefit ratio for perioperative midazolam use. Modifying factors may include drug interactions, differences in pain management and opioid dosing, and pharmacokinetic features of individual benzodiazepine formulas.8 For instance long-acting benzodiazepines may be administered perioperatively for purposes beyond perioperative anesthesia.8 Irrespective of the specific etiology, POD has the potential to distinctly affect the overall outcome and prognosis of patients, as well as significantly increase healthcare costs. This group has previously identified various modifiable risk factors for POD including the utilization of long-acting benzodiazepines, opioid dosing, and general anesthesia.8

We found that the perioperative utilization rate of midazolam significantly differed by racial background, socioeconomic status, and insurance type. This is in accordance with previous studies on perioperative anxiety24 25 that identified the above as risk factors of perioperative anxiety, potentially explaining the higher midazolam use in the current study cohort. In addition, the use of midazolam was significantly different between hospitals of different sizes, an observation whose reasons must remain speculative. Not surprisingly, there was a significant difference between the type of anesthesia and the use of midazolam. This is important as presumably undergoing an invasive procedure while awake is more anxiety generating than undergoing surgery under general anesthesia. However, this is in contrast with studies on perioperative anxiety, in which patients undergoing procedures under regional techniques reported decreased anxiety levels.24 26

The use of midazolam significantly increased the risk of in-hospital falls after TKA and THA. This is in keeping with other studies that identified an increase in odds for falls by around 1.5.27 Benzodiazepine use is associated with dizziness, mobility impairment, cognitive impairment, and visual impairment all of which could be reasons for benzodiazepines increasing the risk of falls.28

Importantly, the concurrent use of midazolam with gabapentinoid drugs was associated with a higher incidence of pulmonary complications, and naloxone utilization, as well as a substantial increase in the occurrence of POD. This is supported by previous evidence29 as both classes of medicines work through the inhibition of the same central neuronal pathway. It appears therefore that their effects may be additive, augmenting the risk for respiratory failure and delirium. Moreover, patients exposed to benzodiazepines and sedative/hypnotics are known to have a prolonged opioid requirement after surgery.30–32

This could explain the increased use of naloxone postoperatively in these patients.

The risk profile of perioperative midazolam may therefore vary based on concurrent conditions and modifying factors. These drivers of outcome should be identified to inform clinical anesthesia practice on the balance between benefit and harm of benzodiazepines.

The current study has a number of limitations. Our data were obtained retrospectively from a healthcare database; therefore, our analysis is limited by the data available and recorded on the database. Moreover, the dose of and the timing of administration of midazolam, as well as whether there were repeated doses administered and at what time intervals, were not recorded, both of which could have an impact on the postoperative result and the appearance of complications. In addition, the database does not follow patients longitudinally; therefore, we did not have any data about whether patients were taking benzodiazepines, gabapentinoids, or opiates preoperatively. Finally, we do not know at what point postoperatively the complications occurred. Nevertheless, given the lack of data, our aim was to investigate the perioperative use of midazolam and their impact on a population-based level. Without suggesting causal conclusions, particularly the concurrent use of midazolam and gabapentinoids was associated with serious postoperative complications in both patient cohorts. This warrants further investigation, especially as the combined use of these drug agents remains common.

In conclusion, midazolam use is highly prevalent in the perioperative setting of major orthopedic surgery. The impact of midazolam, however, requires a distinctive view in terms of drug interactions and pharmacokinetics. While midazolam was associated with an increase in patient falls, the odds for cardiac complications was decreased. Moreover, the risk for POD appeared to be unaltered with midazolam use. Notably, however, the concurrent use of benzodiazepines and gabapentinoids was associated with a substantial increase in the odds for delirium as well as for serious respiratory complications. The perioperative risk profile of midazolam may therefore significantly differ based modifying factors, such as the concurrent use of gabapentinoids. Such potential modifiers should be identified to inform clinical anesthesia on the balance between benefit and harm of midazolam.

Data availability statement

Data may be obtained from a third party and are not publicly available. Data obtained from a third party, Premier (Premier Healthcare Solutions, Charlotte, North Carolina, USA), and are not publicly available.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and was approved by This study was approved by the Institutional Review Board at the Hospital for Special Surgery (IRB#2012-050)The requirement for written informed consent was waived given the de-identified nature of the data.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • VA and CC are joint first authors.

  • Twitter @jashvant_p, @jbLiujb, @sgmemtsoudis

  • Contributors VA: this author helped in study design/planning, interpretation of results, manuscript draft, and review. CC: this author helped in study design/planning, interpretation of results, manuscript preparation, and review. HZ: this author helped in study design/planning, data analysis, interpretation of results, manuscript preparation, and review. JP: this author helped in study design/planning, interpretation of results, and manuscript review. JB: this author helped in study design/planning, interpretation of results, and manuscript review. AI: this author helped in study design/planning, interpretation of results, and manuscript review. LP: this author helped in study design/planning, interpretation of results, and manuscript review. SGM: this author helped in study design/planning, interpretation of results, manuscript preparation, and review. SGM is the guarantor of the study and affirms that the manuscript is an honest, accurate and transparent account of the current data analysis; that no important aspects of the study have been omitted, and that any discrepancies from the study as originally planned have been explained.

  • Funding This work was supported by the Research and Education Fund, Department of Anesthesiology, Critical Care & Pain, Hospital for Special Surgery.

  • Competing interests SGM is a one-time consultant for Sandoz Inc and Teikoku. He is the owner of a US Patent for a Multicatheter Infusion System. US-2017-0361063. He is the owner of SGM Consulting, LLC and co-owner of FC Monmouth, LLC. None of the above relations influenced the conduct of the present study. All other authors declare no competing interest.

  • Provenance and peer review Not commissioned; externally peer reviewed.