KEY POINTS
- Question: Does premedication with midazolam increase the risk of postoperative delirium in older adults?
- Findings: Premedication using midazolam was not associated with higher incidence of delirium on the first postoperative day in older patients undergoing major noncardiac surgery.
- Meaning: The pathophysiology of postoperative delirium is likely complex and not a direct result from exposure to a small dose of premedication with midazolam.
Benzodiazepines are frequently used clinically to relieve anxiety and insomnia, but may also cause delirium in susceptible population, such as older adults after major surgery. Postoperative delirium is a common geriatric syndrome that occurs in older patients after surgery.1,2 Delirium in hospitalized patients is associated with the increased length of stay and hospital costs.3,4 In a prior study of patients >50 years undergoing elective noncardiac surgeries, the postoperative use of long-acting benzodiazepine (chlordiazepoxide, diazepam, and flurazepam) was shown to be associated with higher risk of postoperative delirium than short-acting agents (oxazepam, lorazepam, triazolam, midazolam, and temazepam).5 Midazolam, a short-acting benzodiazepine with an elimination half-life of 1.5–2.5 hours, is commonly used as a premedication.6,7 However, the half-life of midazolam can be doubled in older adults.8 An important clinical question is whether premedication with midazolam is associated with postoperative delirium and whether its use should be limited in older patients at risk for postoperative delirium.
Benzodiazepines such as midazolam are commonly administered as a premedication immediately before surgery to serve as an anxiolytic. However, whether the use of a single dose of short-acting benzodiazepine is associated with delirium in older patients particularly in those aged ≥65 years is a clinical area that has not been clarified. One prior study suggested that the preoperative chronic use of benzodiazepines was associated with postoperative confusion.9 Others found that the new use of lorazepam in critically ill patients was associated with delirium.10 In mechanically ventilated patients in an intensive care unit, midazolam-treated patients experienced more delirium than patients treated with dexmedetomidine.11 The aforementioned studies referred to patients who were already at risk for delirium due to the long-term use of benzodiazepine or who were critically ill and mechanically ventilated. However, whether premedication with a short-acting benzodiazepine such as midazolam has similar adverse cognitive effects in older surgical patients who do not have these predisposing risks for postoperative delirium is unknown.
A recent meta-analysis reported that there are few high-quality studies quantifying the direct association between preoperative medication use and postoperative delirium.12 Accordingly, the aim of this study is to assess the effect of intravenous midazolam as a premedication on the incidence of postoperative delirium in older adults. We hypothesize that the use of midazolam as a premedication is not associated with postoperative delirium in older patients undergoing elective surgical procedures. We combined 3 prospective studies to investigate whether premedication with midazolam is associated with postoperative delirium and adjusted for confounding using propensity score matching.
METHODS
This is a secondary analysis of prospectively conducted studies, as we followed up all patients prospectively for the occurrence of the primary outcome. Data were from 3 prospective studies at the University of California San Francisco (UCSF) Medical Center (San Francisco, CA) between June 2001 and August 2014. These studies received approval from the institutional review board. Written informed consent was obtained from each patient for each study. All these studies were related to a general goal of investigating the pathophysiology of postoperative delirium (study 1: randomized trial of general anesthesia with or without nitrous oxide; study 2: prospective cohort of postoperative cognitive function in older surgical adults; study 3: randomized trial of perioperative administration of gabapentin or placebo).13–15 This article reports results according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.16
Criteria for Patient Selection
Patients from all 3 studies were included if they were aged from 65 to 85 years and scheduled for elective major noncardiac surgery at UCSF Medical Center. Patients were recruited if they were fluent in English and were expected to stay in the hospital for at least 48 hours after surgery. Patients were excluded if they could not provide written informed consent or were undergoing brain or cardiac surgery.
Preoperative Assessment
The preoperative assessment was performed by trained research assistants in the preoperative anesthesia clinic, typically within 1 week of surgery. Patients were interviewed, and their medical history of coexisting diseases and chronic preoperative use of benzodiazepine were recorded. The severity of preoperative coexistent conditions was determined using the Charlson comorbidity index.17 Physical status and anesthesia risk were measured by the American Society of Anesthesiologists (ASA) physical status.18 Preoperative cognitive status was measured using the Telephone Interview for Cognitive Status (TICS),19 an 11-item screening test (maximum score = 41 points), that was adapted from the Mini-Mental State Examination.
Clinical Management
Midazolam as a premedication was given as the patient is taken to the operating room and typically within 15 minutes of anesthetic induction. The use of midazolam as a premedication and chronically preoperative use of benzodiazepines were abstracted from the anesthesia records. As this was not a randomized study, the use of midazolam was not standardized but rather determined by the treating anesthesiologists. The type and duration of surgery were recorded. As the patient recruitment dated from 2001 to 2014, the potential changing medical practice over time may affect clinical practices, and we stratified patients into 2 groups based on the decade when they were recruited.
Outcome Assessment
The primary outcome was postoperative delirium on the first postoperative day. In the secondary analysis, we also included the measurement of delirium on the second postoperative day. Structured interviews were conducted by trained research assistants on both weekdays and weekends between the hours of 9 am to 12 pm to determine the occurrence of postoperative delirium using the confusion assessment method (CAM).20 The CAM was developed as a screening instrument to detect delirium based on Diagnostic and Statistical Manual of Mental Disorders (DSM)-III-R criteria for the use of nonpsychiatric clinicians with a high sensitivity (94%–100%) and specificity (90%–95%).20 Postoperative delirium was measured on the first 2 postoperative days for 2 of the 3 studies, and the third study measured delirium on the first 3 postoperative days. We chose 2–3 days as the cutoff as the majority of our patients were discharged by the early postoperative period.
Statistical Analysis
Bivariate analyses were computed to compare the characteristics of patients who received midazolam as premedication versus those who did not receive treatment. Independent samples t tests or Mann-Whitney U tests were used to compare the 2 groups for continuous valued variables. Chi-square tests were used to compare the groups on categorical variables.
To reduce potential bias due to preoperative patient characteristics associated with being treated with midazolam, we used propensity score matching to derive 2 groups that did not differ on average in preoperative characteristics; 1 group used midazolam and the other did not. Propensity scores reflect the probability of receiving midazolam based on preoperative characteristics. We matched up to 3 midazolam patients with each no-midazolam patient. To maximize the number of matches for each midazolam patient, we used a statistical analysis system (SAS) macro that first assigned matches to those midazolam patients with the fewest possible matches, then proceeded with midazolam patients with more possible matches.21 The matching is based on a preset maximum allowable difference. We used a common value of 0.2 of the standard deviation of the logit of the propensity scores.22
We computed weighted matched standardized differences (the difference in means or proportions divided by the pooled standard deviation) and weighted matched variance ratios for each covariate to assess the performance of propensity score matching.23 The weight for the patients in the no-midazolam group was set to 1, and the weight for the patients in the midazolam group was the inverse of the number of patients in the midazolam group (1–3) that matched to patients in the no-midazolam group. The weighting ensured differences between the 2 groups were not overestimated when more than 1 patient in midazolam group was matched to a patient in the no-midazolam group. We present the difference in means between those in the no-midazolam (control) group and those in the midazolam group. We sought to ensure the weighted matched standardized difference between groups to be smaller than 0.1.24
Our study included 357 patients who received midazolam preoperatively and a weighted sample of 357 who did not receive midazolam. Our prior study found that 20% of standard care elective surgery patients had incident delirium.25 We determined that the current sample size is sufficiently large to detect a difference in incident delirium of 0.25 between those receiving midazolam versus 0.20 who did not receive midazolam or an odds ratio (OR) of 1.33. We determined that the sample allows detection of an OR of 1.33 with a power of 0.92 when α = .05.
Based on the propensity-matched groups, incidence of delirium and OR with 95% confidence intervals (CIs) were calculated for all subjects. All analyses were conducted in SAS 9.4 (SAS Institute, Inc, Cary, NC).
RESULTS
;)
Figure.: Flow chart for inclusion of patients in the analytic sample. TICS indicates Telephone Interview for Cognitive Status.
The flow chart for inclusion of patients in the analytic sample is shown in the Figure. Of 1479 patients with data, 127 were excluded due to incomplete postoperative delirium assessment. The reasons for the missing delirium data included patients’ refusal to assessment, stupor or nonresponsive, or postoperative tracheal intubation. An additional 56 patients were excluded due to missing TICS scores, and another subject was excluded due to missing midazolam data. One subject had missing data for the Charlson comorbidity score, and 28 subjects were <65 years; these patients were also excluded. The final analytic sample included 1266 patients.
Sensitivity Analysis
Comparison of the patients with missing delirium data versus those with data showed that there was no difference in their mean age and other covariates of postoperative delirium, such as gender, ASA physical status, surgical site, Charlson comorbidity score, dates of surgery, or preoperative use of benzodiazepines, surgical site, or year of surgery (Supplemental Digital Content, Table, https://links.lww.com/AA/D427).
Patient Characteristics
Table 1. -
Preoperative Characteristics of All Patients
|
Use of midazolam |
No midazolam |
P value |
Standardized differencea |
| n (%) |
909 (71.8) |
357 (28.2) |
|
|
| Age (y), mean ± SD |
72.2 ± 5.6 |
74.5 ± 6.6 |
<.01 |
0.38 |
| Female, n (%) |
474 (52.2) |
169(47.3) |
.12 |
–0.10 |
| Charlson, mean ± SD |
0.9 ± 1.5 |
1.2 ± 1.4 |
<.01 |
0.17 |
| ASA physical status, n (%) |
| I–II |
497 (54.7) |
191 (53.5) |
.71 |
–0.02 |
| III–V |
412 (45.3) |
166 (46.5) |
|
|
| TICS score, mean ± SD |
33.8 ± 3.5 |
32.7 ± 4.2 |
<.01 |
–0.28 |
| Preoperative use of benzodiazepine, n (%) |
112 (12.3) |
30 (8.4) |
.05 |
0.13 |
| Surgical site, n (%) |
| Hip |
221 (24.3) |
77 (21.6) |
<.01 |
0.20 |
| Knee |
204 (22.4) |
55 (15.4) |
|
|
| Spine |
324 (35.6) |
137 (36.4) |
|
|
| Other |
160 (17.6) |
88 (24.7) |
|
|
| Year, n (%) |
|
|
|
|
| <2010 |
449 (66.9) |
222 (33.1) |
<.01 |
0.26 |
| ≥2010 |
460 (77.3) |
135 (22.7) |
|
|
Abbreviations: ASA, American Society of Anesthesiologists; SD, standard deviation; TICS, Telephone Interview for Cognitive Status.
aStandardized difference: the difference in means or proportions divided by the pooled standard deviation.
Clinical characteristics of patients with or without the use of preoperative midazolam as a premedication are presented in Table 1. Patients who were administered midazolam were younger, had lower Charlson comorbidity scores and higher preoperative TICS total scores, and were more likely to have used benzodiazepines before the scheduled surgery. There was also a difference in the administration of midazolam between the different surgical types and the dates of surgery. The dose of midazolam that was given as a premedication was 1.99 ± 1.03 (mean ± standard deviation) mg.
Propensity Score Matching
Table 2. -
Preoperative Characteristics for the Propensity Score–Matched Cohorts
|
Use of midazolam (weighted) |
No midazolam |
P value |
Standardized differencea |
| N |
357 |
357 |
|
|
| Age (y), mean ± SD |
74.1 ± 4.2 |
74.5 ± 6.6 |
.35 |
0.06 |
| Female, n (%) |
182.2 (51.0) |
169 (47.3) |
.36 |
–0.07 |
| Charlson, mean ± SD |
1.2 ± 1.2 |
1.2 ± 1.4 |
.63 |
–0.03 |
| ASA physical status, n (%) |
| I–II |
190.5 (53.4) |
191 (53.5) |
1.00 |
0.00 |
| III–V |
166.5 (46.6) |
166 (46.5) |
|
|
| TICS score, mean ± SD |
32.9 ± 2.6 |
32.7 ± 4.2 |
.52 |
–0.04 |
| Preoperative use of benzodiazepine, n (%) |
30.8 (8.6) |
30 (8.4) |
1.00 |
0.01 |
| Surgical site, n (%) |
| Hip |
62.3 (17.5) |
77 (21.6) |
.48 |
–0.03 |
| Knee |
65 (18.2) |
55 (15.4) |
|
|
| Spine |
137.7 (38.6) |
137 (36.4) |
|
|
| Other |
92 (25.8) |
88 (24.7) |
|
|
| Year, n (%) |
|
|
|
|
| <2010 |
223.3 (62.6) |
222 (62.2) |
.98 |
–0.01 |
| ≥2010 |
132.7 (37.4) |
135 (37.8) |
|
|
Abbreviations: ASA, American Society of Anesthesiologists; SD, standard deviation; TICS, Telephone Interview for Cognitive Status.
aStandardized difference: the difference in means or proportions divided by the pooled standard deviation.
Before propensity score matching, standardized differences in preoperative characteristics ranged between –0.28 and 0.38. The largest standardized differences were for age (0.38), Charlson comorbidity score (0.17), preoperative cognitive score (–0.28), surgical site (0.20), and year of surgery (0.26). After propensity score matching, all standardized differences were less than |0.10|, indicating that the groups were well matched (Table 2). Specifically, the standardized differences between groups ranged from –0.07 to 0.06.
Incidence of Postoperative Delirium
The incidence of postoperative delirium did not differ significantly between the propensity score–matched groups: 23.30% in the midazolam group developed delirium the first day after surgery compared to 24.93% in the group that did not receive midazolam (OR = 0.91; 95% CI, 0.65-1.29; P = .67). The result was similar when including delirium assessment on the second postoperative day or between hypoactive or hyperactive delirium subtypes. In additional secondary data analysis, when excluding patients who used benzodiazepines on a chronic basis, the incidence of delirium on the first postoperative day also did not differ between the groups who were given midazolam (22.9%) versus those who did not receive the drug (25.4%; OR = 0.88 [95% CI, 0.61–1.25]).
DISCUSSION
In this study, we found that intravenous midazolam given as a premedication immediately before surgery was not associated with higher risk of postoperative delirium.
Comparison With Previous Studies
No study has examined the use of premedication of midazolam and its association with postoperative delirium in older surgical patients. Studies that have examined the use of benzodiazepines have mainly focused on new use in the intensive care unit or postoperative exposure.5,10 In the study by Pandharipande et al,10 the new use of lorazepam was associated with delirium in the critical care unit. In the study by Marcantonio et al,5 postoperative exposure to meperidine and benzodiazepines was found to be associated with postoperative delirium. In the study by Chaiwat et al,26 perioperative exposure to benzodiazepine was reported to be a risk factor for postoperative delirium in patients who were admitted to surgical intensive care unit within 7 days after surgery. However, they did not report on the types of benzodiazepine used or whether they were used before or after surgery. All of the aforementioned studies did not evaluate the association between midazolam used as a premedication and postoperative delirium. One study by Kudoh et al9 reported that “postoperative confusion” increased in older patients who had preoperative use of benzodiazepine. In 57 patients (17%) who reported taking benzodiazepines, most of them used triazolam (53%) or ethyl loflazepate (26%). No one had preoperative exposure to midazolam. Of importance, only long-term users (daily use for more than 1 year) were found to have higher incidence of postoperative confusion. In our cohort, the preoperative use of benzodiazepine was used in the computation of the propensity scores and consequently did not differ between the propensity-matched groups.
Patients who received midazolam as premedication in our study were younger, had fewer medical comorbidities, and had more depressive symptoms preoperatively. This potential bias of not administering midazolam based on age is likely secondary to the anticipated negative effects of benzodiazepine as a premedication on cognition in older patients.27,28
Potential Limitations
There are several potential limitations in our study. First, the use of premedication of benzodiazepine or the administration of postoperative benzodiazepine was not randomized. We adjusted for this potential confounding factor using the propensity score–matching method. However, despite this method, there may be other factors that cannot be controlled adequately, such as the depth of anesthesia during surgery, which may affect the incidence of postoperative delirium. Second, we focused on measuring postoperative delirium once on the first few postoperative days. As a result, we may have missed patients who developed postoperative delirium, which was not assessed during the interview because of the fluctuating nature of delirium or because delirium occurred in the later postoperative period. However, given the short distribution half-life of midazolam, it is unlikely that the drug will have a substantial effect on cognition beyond the first day after surgery. Finally, our studies focused on patients who underwent elective noncardiac and noncranial surgery, and the results may not be directly generalized to patients undergoing emergency or other types of surgery.
Currently, a pragmatic, multicenter, cluster crossover trial is ongoing to evaluate the association between intraoperative benzodiazepine and postoperative delirium in cardiac surgical patients (ClinicalTrials.gov Identifier: NCT03928236). A recently completed multicenter, randomized trial investigated the impact of premedication with midazolam on outcome of older surgical patients scheduled for noncardiac surgery (ClinicalTrials.gov Identifier: NCT03052660).29 We need to wait for the current prospective studies to make a broader generalization about whether benzodiazepines can be used without concern for postoperative delirium.
CONCLUSIONS
In summary, our study showed that premedication using midazolam was not associated with higher incidence of delirium in the early postoperative period. These results suggest that the pathophysiology of postoperative delirium is likely complex and not as a direct result from exposure to a small dose of premedication with midazolam.
CONTRIBUTORS
Members of the Perioperative Medicine Research group are as follows: Principal investigator: Jacqueline M. Leung, MD, MPH. Research associates: Christopher Tang, BA, Devon Pleasants, BS, Sanam Tabatabai, BS, Danielle Tran, BS, Stacey Chang, BA, Gabriela Meckler, BA, Stacey Newman, BA, Tiffany Tsai, MD, Vanessa Voss, MD, and Emily Youngblom, BA.
ACKNOWLEDGMENTS
We acknowledge the Perioperative Medicine Research Group in their assistance in patient recruitment, interviews, and data collection.
DISCLOSURES
Name: Man-Ling Wang, MD.
Contribution: This author helped with conception and design of study, acquisition of data, interpretation of data, drafting and revising the manuscript, and approval of the version of the manuscript to be published.
Name: Jie Min, MS.
Contribution: This author helped with the analysis of data and approval of the version of the manuscript to be published.
Name: Laura P. Sands, PhD.
Contribution: This author helped with the analysis of data, interpretation of data, revising the manuscript critically for important intellectual content, and approval of the version of the manuscript to be published.
Name: Jacqueline M. Leung, MD, MPH.
Contribution: This author helped with the conception and design of study, acquisition of data, interpretation of data, drafting the manuscript, revising the manuscript critically for important intellectual content, and approval of the version of the manuscript to be published.
This manuscript was handled by: Robert A. Whittington, MD.
REFERENCES
1. Santana Santos F, Wahlund LO, Varli F, Tadeu Velasco I, Eriksdotter Jonhagen M. Incidence, clinical features and subtypes of delirium in elderly patients treated for hip fractures. Dement Geriatr Cogn Disord. 2005;20:231–237.
2. Kalisvaart KJ, Vreeswijk R, de Jonghe JF, van der Ploeg T, van Gool WA, Eikelenboom P. Risk factors and prediction of postoperative delirium in elderly hip-surgery patients: implementation and validation of a medical risk factor model. J Am Geriatr Soc. 2006;54:817–822.
3. Lipowski ZJ. Delirium in the elderly patient. N Engl J Med. 1989;320:578–582.
4. Thomas RI, Cameron DJ, Fahs MC. A prospective study of delirium and prolonged hospital stay. Exploratory study. Arch Gen Psychiatry. 1988;45:937–940.
5. Marcantonio ER, Juarez G, Goldman L, et al. The relationship of postoperative delirium with psychoactive medications. JAMA. 1994;272:1518–1522.
6. Kanto JH. Midazolam: the first water-soluble benzodiazepine. Pharmacology, pharmacokinetics and efficacy in insomnia and anesthesia. Pharmacotherapy. 1985;5:138–155.
7. Kain ZN, Mayes LC, Bell C, Weisman S, Hofstadter MB, Rimar S. Premedication in the United States: a status report. Anesth Analg. 1997;84:427–432.
8. Albrecht S, Ihmsen H, Hering W, et al. The effect of age on the pharmacokinetics and pharmacodynamics of midazolam. Clin Pharmacol Ther. 1999;65:630–639.
9. Kudoh A, Takase H, Takahira Y, Takazawa T. Postoperative confusion increases in elderly long-term benzodiazepine users. Anesth Analg. 2004;99:1674–1678.
10. Pandharipande P, Shintani A, Peterson J, et al. Lorazepam is an independent risk factor for transitioning to delirium in intensive care unit patients. Anesthesiology. 2006;104:21–26.
11. Riker RR, Shehabi Y, Bokesch PM, et al.; SEDCOM (Safety and Efficacy of Dexmedetomidine Compared With Midazolam) Study Group. Dexmedetomidine vs midazolam for sedation of critically ill patients: a randomized trial. JAMA. 2009;301:489–499.
12. Kassie GM, Nguyen TA, Kalisch Ellett LM, Pratt NL, Roughead EE. Preoperative medication use and postoperative delirium: a systematic review. BMC Geriatr. 2017;17:298.
13. Leung JM, Sands LP, Vaurio LE, Wang Y. Nitrous oxide does not change the incidence of postoperative delirium or cognitive decline in elderly surgical patients. Br J Anaesth. 2006;96:754–760.
14. Leung JM, Sands LP, Lim E, Tsai TL, Kinjo S. Does preoperative risk for delirium moderate the effects of postoperative pain and opiate use on postoperative delirium? Am J Geriatr Psychiatry. 2013;21:946–956.
15. Leung JM, Sands LP, Chen N, et al.; Perioperative Medicine Research Group. Perioperative gabapentin does not reduce postoperative delirium in older surgical patients: a randomized clinical trial. Anesthesiology. 2017;127:633–644.
16. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ. 2007;335:806–808.
17. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–383.
18. Wolters U, Wolf T, Stützer H, Schröder T. ASA classification and perioperative variables as predictors of postoperative outcome. Br J Anaesth. 1996;77:217–222.
19. Brandt J, Spencer M, Folstein M. The telephone interview for cognitive status. Neuropsychiatry Neuropsychol Behav Neurol. 1988;1:111–117.
20. Inouye SK, van Dyck CH, Alessi CA, Balkin S, Siegal AP, Horwitz RI. Clarifying confusion: the confusion assessment method. A new method for detection of delirium. Ann Intern Med. 1990;113:941–948.
21. Murphy B, Fraeman KH, Walham E. SAS
® Macro to implement optimal N:1 propensity score matching with a maximum radius. Paper 812-2017. Accessed March 7, 2021.
https://support.sas.com/resources/papers/proceedings17/0812-2017.pdf.
22. Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. 2011;10:150–161.
23. Schulte PJ, Mascha EJ. Propensity score methods: theory and practice for anesthesia research. Anesth Analg. 2018;127:1074–1084.
24. Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med. 2009;28:3083–3107.
25. Tang CJ, Jin Z, Sands LP, et al. ADAPT-2: a randomized clinical trial to reduce intraoperative EEG suppression in older surgical patients undergoing major noncardiac surgery. Anesth Analg. 2020;131:1228–1236.
26. Chaiwat O, Chanidnuan M, Pancharoen W, et al. Postoperative delirium in critically ill surgical patients: incidence, risk factors, and predictive scores. BMC Anesthesiol. 2019;19:39.
27. Sun GC, Hsu MC, Chia YY, Chen PY, Shaw FZ. Effects of age and gender on intravenous midazolam premedication: a randomized double-blind study. Br J Anaesth. 2008;101:632–639.
28. Pisani MA, Murphy TE, Araujo KL, Slattum P, Van Ness PH, Inouye SK. Benzodiazepine and opioid use and the duration of intensive care unit delirium in an older population. Crit Care Med. 2009;37:177–183.
29. Kowark A, Rossaint R, Keszei AP, et al.; I-PROMOTE Study Group. Impact of PReOperative Midazolam on OuTcome of Elderly patients (I-PROMOTE): study protocol for a multicentre randomised controlled trial. Trials. 2019;20:430.
