You are here

  • Home
  • Archive
  • Volume 44, Issue 10
  • Risk factors for postoperative delirium in patients undergoing lower extremity joint arthroplasty: a retrospective population-based cohort study

Article Text

Article menu
Download PDFPDF
Original article
Risk factors for postoperative delirium in patients undergoing lower extremity joint arthroplasty: a retrospective population-based cohort study
  1. Stavros Memtsoudis1,2,3,
  2. Crispiana Cozowicz1,2,
  3. Nicole Zubizarreta4,5,
  4. Sarah M Weinstein1,
  5. Jiabin Liu1,
  6. David H Kim1,
  7. Lazaros Poultsides6,
  8. Marc Moritz Berger2,
  9. Madhu Mazumdar4 and
  10. Jashvant Poeran4,5
  1. 1 Department of Anesthesiology, Critical Care and Pain Management, Hospital for Special Surgery - Weill Cornell Medical College, New York, New York, USA
  2. 2 Department of Anesthesiology, Perioperative Medicine and Intensive Care Medicine, Paracelsus Medizinische Privatuniversitat, Salzburg, Austria
  3. 3 Department for Health Care Policy and Research, Weill Cornell Medical College, New York, NY, USA
  4. 4 Department of Population Health Science and Policy, Institute for Healthcare Delivery Science, Icahn School of Medicine at Mount Sinai, New York, New York, USA
  5. 5 Department of Orthopedic Surgery, Icahn School of Medicine at Mount Sinai, New York, New York, USA
  6. 6 Department of Orthopedic Surgery, NYU Langone Health, New York, New York, USA
  1. Correspondence to Dr Stavros Memtsoudis, Department of Anesthesiology, Critical Care and Pain Management, Hospital for Special Surgery - Weill Cornell Medical College, New York, NY 10021, USA; memtsoudiss{at}hss.edu

Abstract

Background With an ageing population, the demand for joint arthroplasties and the burden of postoperative delirium is likely to increase. Given the lack of large-scale data, we investigated associations between perioperative risk factors and postoperative delirium in arthroplasty surgery.

Methods This retrospective population-based cohort study, utilized national claims data from the all-payer Premier Healthcare database containing detailed billing information from >25% nationwide hospitalizations. Patients undergoing elective total hip/knee arthroplasty surgery (2006–2016) were included.

The primary outcome was postoperative delirium, while potential risk factors included age, gender, race, insurance type, and modifiable exposures including anesthesia type, opioid prescription dose (low/medium/high), benzodiazepines, meperidine, non-benzodiazepine hypnotics, ketamine, corticosteroids, and gabapentinoids.

Results Among 1 694 795 patients’ postoperative delirium was seen in 2.6% (14 785/564 226) of hip and 2.9% (32 384/1 130 569) of knee arthroplasties. Multivariable models revealed that the utilization of long acting (OR 2.10 CI 1.82 to 2.42), combined long/short acting benzodiazepines (OR 1.74 CI 1.56 to 1.94), and gabapentinoids (OR 1.26 CI 1.16 to 1.36) was associated with increased odds of postoperative delirium. Lower odds of postoperative delirium were seen for neuraxial versus general anesthesia (OR 0.81 CI 0.70 to 0.93) and with the utilization of non-steroidal anti-inflammatory drugs (OR 0.85 CI 0.79 to 0.91) as well as cyclooxygenase-2 inhibitors (OR 0.82 CI 0.77 to 0.89). Age-stratified analysis revealed lower odds with high versus low opioid dose (OR 0.86 CI 0.76 to 0.98) in patients >65 years. Findings were consistent between hip and knee arthroplasties.

Conclusions In this large national cohort, we identified various modifiable risk factors (including anesthesia type and pharmaceutical agents) for postoperative delirium, demonstrating possible prevention pathways.

  • complications
  • regional anesthesia
  • opioids
  • adverse effects

https://doi.org/10.1136/rapm-2019-100700

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

    Postoperative delirium is a severe complication and a major source of short-term and long-term morbidity, prolonged cognitive impairment, mortality risk, and increased resource utilization.1 2 As an acute alteration in mental status, delirium is characterized by a decline in cognitive function and attention.2–4 Hallmarks of this acute brain dysfunction include the inability to direct, focus, sustain, or shift attention as well as impaired memory, disorientation, or perceptual disturbances.2 3 The frequency in orthopedic surgery,5 may be driven by the elderly demographic burdened by higher comorbidity,6–8 as well as underlying surgical risk factors such as micro emboli intravasation during prosthesis implantation and the requirement for comprehensive pharmaceutical pain management interventions.9

    While these and other risk factors have been identified previously, only a small number of investigations have studied interventions that may mitigate postoperative delirium risk in elective joint arthroplasty. Moreover, data on the incidence of postoperative delirium and associated trends of risk factors in hip and knee arthroplasty patients, indirectly evaluating the overall impact of advances in care for postoperative delirium remain largely unavailable. A recent investigation by our group identified general anesthesia, intraoperative ketamine, and benzodiazepines as risk factors for postoperative delirium after joint arthroplasty.10 However, the single institutional nature of these data, emphasized the need for the utilization of a more diverse and nationally representative cohort with higher external validity.

    Therefore, we utilized a population-based data source, representing information from >600 hospitals throughout the USA in various teaching and practice settings, to investigate (1) incidence and trends of postoperative delirium among hip and knee arthroplasty recipients, (2) trends in potential perioperative risk factors, based on previous literature,2 and (3) to identify perioperative risk factors for postoperative delirium.

    Such data are of high importance as they would allow for the establishment of a benchmark incidence, enable the evaluation of the impact of cumulative efforts to reduce the occurrence of postoperative delirium and finally facilitate the identification of potentially modifiable perioperative risk factors such as those related to pharmacological and anesthetic interventions.

    We hypothesized that the incidence of postoperative delirium and its risk factors among joint arthroplasty patients would be reduced over time and that modifiable risk factors could be identified.

    Methods

    This is a retrospective cross-sectional cohort study of patients undergoing primary hip or knee arthroplasty. Data were extracted from the nationwide all-payer Premier Healthcare database (Premier Inc., Charlotte, North Carolina, USA), which contains detailed billing information on >25% of hospitalizations in the USA.11 12 Patient records with International Classification of Diseases, Ninth Revision (ICD-9) procedure codes for primary hip (81.51) or knee (81.54) arthroplasty from 2006 to 2016 were included, yielding a total of 1 814 389 patient records. Exclusion criteria were: patients undergoing both procedures (n=341), non-elective procedures (n=1 10 464), unknown sex (n=294), unknown discharge status (n=814), outpatient procedures (n=7341), and surgery at a hospital that performed less than 30 procedures per year (to ensure sufficient sample size per hospital; n=340).13

    The study was exempt from requirements for consent by the Institutional Review Boards of the Hospital for Special Surgery (# 2012–050) and the Mount Sinai Hospital (# 14–00647) due to the de-identified nature of the data and compliance with the Health Insurance Portability and Accountability Act.

    Analysis plan

    An analysis plan was developed a priori to define the population, exposures, baseline variables, and the primary outcome of interest.

    Primary outcome

    The primary outcome was postoperative delirium which was identified by ICD-9 codes and/or billing for antipsychotics (haloperidol, olanzapine, and quetiapine) as recently recommended for studies utilizing claims-based data14 (online supplementary file 1).

    Supplemental material

    Cohort characteristics

    According to the a priori established analysis plan, a number of baseline variables were investigated as potential modifiers of outcome. This included patient related (age, gender, race), healthcare related (insurance type, hospital-location, hospital-size, hospital-teaching status, hospital-specific number of annual hip/knee arthroplasties), and procedure related variables (year of procedure, anesthesia type, utilization of a peripheral nerve block, acetaminophen/paracetamol, non-steroidal anti-inflammatory drugs (NSAIDs), or cyclooxygenase-2 (COX-2) inhibitors) on the day of surgery (D0) or the day after (D1). Furthermore, comorbidity burden was assessed using the Quan adaptation of the Deyo-Charlson comorbidity index which includes the presence of dementia as a separate variable.15 The presence of substance use/abuse, chronic pain, and psychiatric conditions were investigated and adjusted for due to their potential impact on the occurrence of postoperative delirium either through their association with pain or opioid utilization.

    Exposure variables

    The following exposures were selected as potential (modifiable) predictors of postoperative delirium according to the Postoperative Delirium Best Practice Statement by the American Geriatrics Society2:

    Opioid utilization represented by total opioid prescription dose on D0 and D1, applied anesthesia technique (general, neuraxial, combined general/neuraxial), and benzodiazepine utilization stratified by use of long-acting and short-acting drug agents or the combination thereof. Further exposures of interest included the use meperidine, sedative hypnotics, ketamine, corticosteroids, gabapentinoids (gabapentin, pregabalin), NSAIDs, and COX-2 inhibitors on D0 or D1. These were expressed as binary variables. Opioid prescription was expressed in oral morphine equivalents (calculated by using the Lexicomp “opioid agonist conversion” and the GlobalRPH “opioid analgesic converter” calculator).16 17 Furthermore, opioid prescription was broken down into low, medium and high dose, using percentiles (<25th, 25th–75th, and >75th percentile).

    Statistical analysis

    First, trends in the incidence of delirium and associated exposure variables were cross-tabulated and graphed (figures 1 and 2). These results were utilized to aid in the interpretation of findings from multivariable models.

    Figure 1
    Figure 1

    Trends in the incidence of postoperative delirium after total hip and knee arthroplasty.

    Figure 2
    Figure 2

    Trends in the utilization of exposures of interest.

    Analyses were performed separately for hip and knee arthroplasties. 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 of 0.1 (or 10%) has been suggested to indicate a meaningful difference between groups.18 Multilevel, multivariable regression models measured associations between the aforementioned variables and the outcome of postoperative delirium. Multilevel (or mixed-effects) models account for the correlation of patients within hospitals and fit separate regression lines by hospital.19 This is necessary as patients within the same hospital may be correlated, given the likely similar treatment and care. Multivariable models were adjusted for: patient age, gender, race/ethnicity, insurance type, hospital location, hospital size, hospital teaching status, annual procedure volume, year, anesthesia type, use of benzodiazepine, opioids, meperidine, hypnotics, ketamine, corticosteroids, gabapentinoids, COX-2 inhibitors, NSAIDs, acetaminophen/paracetamol, patient controlled analgesia (PCA), and peripheral nerve blocks as well as the Deyo-Charlson comorbidity index, and drug use/abuse, chronic pain, and psychiatric conditions. In addition to assessing effects in the full cohort where age was treated as a continuous variable, we introduced interaction terms in our multivariable model in a second analysis step to assess potential effects of the established risk factors separately for those aged <65 and ≥65 years. Adjusted ORs and Bonferroni-adjusted 95% CIs are reported taking into account the number of hypotheses tested for in the main analyses (112 hypotheses assessing associations between exposures and incidence of postoperative delirium). It must be noted that while this step may reduce the risk of type I errors the likelihood of type II errors may be increased.20 For all models the PROC GLIMMIX procedure in SAS V.9.4 statistical software was used.

    Results

    The final study sample consisted of a total of 1 694 795 cases of patients undergoing elective total hip/knee arthroplasty surgery (564 226/1 130 569). The incidence of postoperative delirium was 2.6% (14 785) in the hip and 2.9% (32 384) in the knee cohort. A continuous decrease in the incidence of postoperative delirium was observed particularly after 2011 (figure 1).

    Trends in the utilization of exposures of interest are presented in figure 2. Most notably, a significant increase in the use of gabapentinoids from 5.9% in 2006 to 40.7% in 2016 was observed, while opioid prescription decreased from 305 mg to 190 mg in hip and from 343 mg to 227 mg oral morphine equivalents in knee arthroplasty patients.

    Table 1 presents covariate distributions by postoperative delirium, separately for hip and knee arthroplasty. In hip arthroplasty patients, a higher incidence of postoperative delirium was observed in patients of increased age (median 69 vs 65), on Medicaid (4.7%) and Medicare insurance (3.5%), with utilization of long-acting benzodiazepines (7.0%), with a higher Charlson-Deyo comorbidity score (4.3%–6.6%) as well as in patients with a history of substance use/abuse (6.5%), chronic pain (3.4%) or a psychiatric condition (6.0%); all with standardized differences >0.1. These observations were similar in knee arthroplasty patients.

    View this table:
    Table 1

    Baseline variables of patients receiving total hip and knee arthroplasty

    Multilevel logistic regression

    Table 2 shows effect estimates for the odds of postoperative delirium, after controlling for covariates. Age was associated with increased odds for postoperative delirium in both patient cohorts, OR 1.03 (CI 1.03 to 1.03) and OR 1.03 (CI 1.02 to 1.03), respectively in hip and knee patients.

    View this table:
    Table 2

    Multilevel regression models predicting the odds for postoperative delirium

    Compared with patients on commercial insurance, those covered by government programs had significantly higher odds for delirium; this was OR 2.20 (CI 1.91 to 2.54) in hip and OR 2.21 (CI 2.00 to 2.45) in knee patients on Medicaid insurance. In patients with Medicare insurance estimates were OR 1.72 (CI 1.57 to 1.89) in hip and OR 1.68 (CI 1.58 to 1.78) in knee arthroplasty.

    With regard to perioperative exposures, the odds for postoperative delirium were reduced with neuraxial versus general anesthesia, OR 0.81 (CI 0.70 to 0.93) in the hip and OR 0.83 (CI 0.76 to 0.92) in knee cohort. The use of long acting benzodiazepines was associated with increased odds for delirium, OR 2.10 (CI 1.82 to 2.42) and OR 2.24 (CI 2.01 to 2.49), respectively for hip and knee patients. Similarly, the risk for delirium was increased when short and long acting benzodiazepines were combined, OR 1.74 (CI 1.56 to 1.94) and OR 1.78 (CI 1.64 to 1.92) in hip/knee patients, respectively. However, lower odds for delirium were observed for short acting benzodiazepines OR 0.79 (CI 0.72 to 0.87) and OR (CI 0.82 (CI 0.77 to 0.88). The odds for postoperative delirium were significantly increased with the use of gabapentinoids, OR 1.26 (CI 1.16 to 1.36) and OR 1.29 (CI 1.22 to 1.36) in hip and knee arthroplasty, respectively. Finally, in both patient cohorts (hip/knee), the utilization of NSAIDs and selective COX-2 inhibitors was associated with significantly decreased odds for delirium OR 0.85 (CI 0.79 to 0.91)/OR 0.84 (CI 0.80 to 0.88) and OR 0.82 (0.77–0.89)/OR 0.83 (0.79–0.88), respectively.

    Model adjusted for: gender, race/ethnicity, insurance type, hospital location, size, teaching status and annual procedure volume, year of procedure, admission type, surgery type, day of hospitalization on which surgery was performed, use of non-opioid analgesics, Charlson-Deyo comorbidities, and history of substance use/abuse, chronic pain or psychiatric conditions.

    Multilevel logistic regression after age stratification

    Given the importance of age in this context, we stratified patients using a cut-off at 65 years. (table 3). Here, 47.6% of hip arthroplasty patients were aged <65, while this proportion was 42.3% for hip arthroplasties. Results by age group were in line with findings from the non-stratified models regarding anesthesia type, benzodiazepines, and gabapentinoids. Moreover, age-stratified analysis showed that in patients aged ≥65, high and medium versus low opioid use was associated with reduced odds for delirium in both patient cohorts; the ORs for high versus low opioid dose were OR 0.86 (CI 0.76 to 0.98) in hip and OR 0.82 (CI 0.77 to 0.88) in knee arthroplasty patients.

    View this table:
    Table 3

    Age stratified multilevel regression models predicting the odds for postoperative delirium

    Model adjusted for: gender, race/ethnicity, insurance type, hospital location, size, teaching status and annual procedure volume, year of procedure, admission type, surgery type, day of hospitalization on which surgery was performed, use of non-opioid analgesics, Charlson-Deyo comorbidities, and history of substance use/abuse, chronic pain or psychiatric conditions.

    Discussion

    In this cohort of approximately 1.7 million cases of total hip and knee arthroplasties performed in the USA between 2006 and 2016 the overall incidence of postoperative delirium was 2.6% and 2.9%, respectively, with a decreasing trend. Adjusted analysis revealed significant differences in delirium incidence by insurance status. The use of general anesthesia, long acting benzodiazepines and gabapentinoids were identified as independent, modifiable risk factors for delirium. Conversely, the use of NSAIDs and selective COX-2 inhibitors appeared protective given the significantly reduced odds for delirium. Age-stratified analysis revealed that high versus low opioid prescription dose was associated with decreased odds for postoperative delirium, potentially reflecting implications of more aggressive pain control measures. A significant increase in the use of gabapentinoids concurrent with a decrease in opioid utilization was observed.

    Postoperative delirium is a severe complication associated with a cascade of adverse consequences, ranging from increased perioperative morbidity and resource utilization to the loss of functional independence, reduced cognitive function and higher mortality risk.2 This warrants the investigation of modifiable risk factors for the development of preventive measures. Perioperative anesthesia related factors appear to be of major interest in this context, given the association with multiple interventions and drugs affecting cognition and pain.

    Trends in the incidence of delirium and exposure variables

    In line with recently published institutional data, we identified a decreasing trend in postoperative delirium incidence.10 This trend may in part reflect deliberate attempts to modify risk by implementing interventions such as the identification of patients at risk and avoidance of inciting factors including sleep deprivation and polypharmacy. Measures of sleep hygiene and early recovery after surgery protocols, designed to hasten rehabilitation of patients by maximizing interventions targeted to facilitate mobility and return to function, have been increasingly used in practice and are both factors known to mitigate delirium risk.21 22 Nevertheless, we identified a trend of potential concern in the increasing use of gabapentinoids. Having identified their use as a risk factor for the development of postoperative delirium potential causality needs to be explored further.

    Patient and healthcare related factors

    A number of demographic risk factors for postoperative delirium as previously described, were confirmed in our data, including age and insurance type. Indeed, older patients are conceivably at higher risk based on higher comorbidity burden, age-related changes in organ and brain composition, renal function, pharmacodynamics and metabolism.23

    Patients with Medicaid and Medicare insurance had a significantly higher risk for delirium compared with commercial insurance. While the association with Medicare is likely related to older age and subsequent higher delirium risk, these findings are consistent with Xu et al,24 who demonstrated that Medicaid and Medicare payer status was associated with a higher risk for postoperative complications, readmissions, and mortality in hip replacement surgery. Primary payer status appears to be a surrogate marker of socioeconomic differences and disparities in the healthcare system,25 by affecting variables such as race, ethnicity, education, language barriers, access to care and its quality, as well as comorbidities and their treatment. Therefore, underlying factors that confer insurance type as a risk factor for poor postoperative outcome require further evaluation to abolish disparities in care.24 26

    Anesthesia technique

    Neuraxial anesthesia was associated with reduced odds for postoperative delirium compared with general anesthesia. This is consistent with most previous studies on delirium in joint replacement surgery.8 10 27 Recent findings by our group using data from a specialized high volume orthopedic institution corroborate the notion of neuraxial anesthesia mitigating delirium risk, showing a 41%–45% decrease in the odds when compared with general anesthesia.10 However, the exact contribution of anesthesia technique related factors remains controversial. Potential effects of neuraxial anesthesia in reducing delirium risk may be related to the avoidance or reduction of systemic anesthetic drugs that interact with the central nervous system.28 Further, potential implications of general anesthesia that can be avoided through regional anesthesia in this context include postoperative hypoxemia, cerebral blood flow changes (due to reduced cardiac output from positive pressure ventilation) or potential hyperventilation with hypocapnia causing subsequent cerebral vasoconstriction.29 Indirect effects of neuraxial anesthesia on delirium risk may include moderation of the neuroendocrine surgical stress response, pain intensity, blood loss and thromboembolic complications.29 30 This further highlights the importance of regional anesthesia as a potential modifier, possibly mitigating delirium risk, as also supported by the American Geriatric Society.2

    Benzodiazepines

    We found that particularly the use of long-acting benzodiazepines was associated with increased odds for postoperative delirium. This is in line with Marcantonio et al 31 who showed that exposure to long-acting benzodiazepines had a stronger association with delirium than short-acting agents (OR 5.4 CI 1.0 to 29.2 vs 2.6 CI 1.1 to 6.5; respectively).

    Notably, in our previously published institutional data we found only postoperatively but not intraoperatively administered benzodiazepines to be associated with increased odds for delirium.10 Given that for intraoperative administration short-acting benzodiazepines such as midazolam are common, while postoperative formulations more frequently include long-acting agents, these findings may find their explanation in our presented observations. Interestingly, short-acting benzodiazepines were associated with reduced odds for delirium, demonstrating the need for future studies to determine mechanisms by which categories of benzodiazepines may or may not impact on delirium. Certainly, the issue of indication or selection bias cannot be ruled out in our data reflecting everyday practice, as we cannot identify reasons for specific drug administration. However, consistent with previous findings, our data indicate that long acting benzodiazepines should be avoided in the perioperative setting, particularly as there is little rationale for their use in acute care.31

    Gabapentinoids

    The odds for delirium were increased by about 30% in patients who received perioperative gabapentinoids. This in particular may have important implications, because in line with recommendations by the American Society of Anesthesiologists, pregabalin and gabapentin have seen an increase in perioperative utilization over the past decade as components of multimodal analgesia and enhanced recovery programs.32 33 According to a recent European survey, 70% of anesthesiologists appear to prescribe perioperative gabapentinoids, expecting to reduce acute pain, opioid consumption, opioid related side effects, and possibly postoperative chronic neuropathic pain.34 Despite their analgesic benefits however,35–37 gabapentinoids have also been associated with cognition-related adverse side effects, including sedation, confusion, dizziness, light-headedness, and visual disturbances, particularly when used simultaneously with opioids and other sedatives.37–42 Nevertheless, awareness for these complications among clinicians appears to be limited.34 Given our results, identifying gabapentinoids as potential drivers postoperative delirium, crucial follow-up studies should focus on understanding the therapeutic efficacy of routine perioperative gabapentinoid regimens relative to their potential harm.33 35 43

    Opioids

    Opioid utilization at higher dose levels was associated with significantly reduced odds for postoperative delirium, particularly in patients aged 65 or older, that is, those with higher baseline risks for delirium. This observation may potentially reflect differences in levels of postoperative pain control, given that inadequate pain management is known to precipitate postoperative delirium.44 45

    Indeed, our findings are in line with Morrison et al 46 who demonstrated that hip fracture patients receiving lower doses of parenteral morphine, were more likely to develop delirium compared with those with more potent analgesia. Likewise, Robinson et al 45 found an association between low dose analgesia and delirium, supporting the notion that adequate opioid analgesia may not increase the risk for delirium,31 but rather the underuse of analgesic medication.47 48 Nevertheless, opioids are frequently avoided for fear of potential detrimental impact on cognition. Our data appear to highlight the importance of adequate pain control, which is known to be critical in minimizing the risk for postoperative delirium.44 45 49

    In contrast to previous evidence,31 our analysis did not show an association between meperidine use and delirium. Given the novelty of our data, this could indicate a more cautious and selective utilization of meperidine that has emerged based on concerns raised in older studies.

    NSAIDs and selective COX-2 inhibitors

    In patients utilizing NSAIDs and selective COX-2 inhibitors the odds for postoperative delirium were reduced by more than 15%. Although causality cannot be inferred by our data, potential mechanisms could involve opioid sparing. Further, inflammatory mediators, such as those encountered in the perioperative period, are presumed to be among the most common triggers of delirium.50 As key enzymes in the conversion of arachidonic acid to prostaglandins and other lipid mediators, COX-1 and COX-2 appear to play a major role in neuroinflammation.50 Griffin et al described the role of COX-1 generated prostaglandins as a critical part of the inflammatory cascade leading to delirium, particularly in patients with a history of dementia.50 Moreover, the authors suggested the further investigation of the widely used NSAID ibuprofen as a potential prophylactic treatment or intervention for delirium of inflammatory etiology in the setting of surgery. In the context of inflammatory signaling, emerging evidence supports the notion of protection against cognitive deficits with COX-1 and COX-2 inhibitors, thus, supporting further investigation of these agents as potential therapeutic approaches in neurodegenerative diseases with an inflammatory component.50 51

    Our study is subject to a number of limitations, primarily based on the retrospective nature of the utilized data. Despite efforts to rigorously adjust for baseline differences, the lack of randomization and detailed clinical information (eg, delirium type and severity, pain intensity, sleep deprivation) limits our ability to provide causal conclusions and the potential for residual confounding must be acknowledged. Thus, associations require cautious interpretation within the context of physiological plausibility, logical, and current evidence. From a clinical perspective, the identification of (particularly hypoactive) delirium requires adequate screening and detection tools that are currently not widely established and thus delirium burden is commonly underestimated. Therefore, the potential underestimation of delirium burden by our definition reflects clinical practice. Moreover, we utilized an algorithm recently recommended specifically for the current database.14 Strengths of this observational study lie in the large sample size and the representation of delirium in orthopedic surgery on a national level, enabling investigation of real-world clinical practice. Given the potential for substantial harm and the increasing demand for joint arthroplasty surgery worldwide, the identification of modifiable risk factors is crucial in an attempt increase perioperative patient safety.52

    Conclusions

    In conclusion, among 1.7 million cases of major orthopedic surgery, the incidence of postoperative delirium was 2%–3%. While the incidence was decreasing, the absolute burden is expected to increase with an ageing population and an increasing demand for joint arthroplasty surgery. Importantly, we identified modifiable drivers of postoperative delirium including the utilization of general versus regional anesthesia, long acting benzodiazepines, gabapentinoids, and insufficient use of analgesia. Given the association of postoperative delirium with substantial harm, these results may prove valuable in targeting interventions to prevent postoperative delirium. Crucial follow-up studies may evaluate the benefits of interventions based on results from the current study.

    References

    1. ↵
      2. Gleason LJ ,
      3. Schmitt EM ,
      4. Kosar CM , et al
      . Effect of delirium and other major complications on outcomes after elective surgery in older adults. JAMA Surg 2015;150:1134–40.doi:10.1001/jamasurg.2015.2606
      OpenUrl
    2. ↵
      1. American Geriatrics Society Expert Panel on Postoperative Delirium in Older Adults
      . Postoperative delirium in older adults: best practice statement from the American geriatrics Society. J Am Coll Surg 2015;220:136–48.doi:10.1016/j.jamcollsurg.2014.10.019
      OpenUrlCrossRefPubMed
    3. ↵
      2. Rengel KF ,
      3. Pandharipande PP ,
      4. Hughes CG
      . Postoperative delirium. Presse Med 2018;47:e53–64.doi:10.1016/j.lpm.2018.03.012
      OpenUrl
    4. ↵
      2. Maclullich AMJ ,
      3. Ferguson KJ ,
      4. Miller T , et al
      . Unravelling the pathophysiology of delirium: a focus on the role of aberrant stress responses. J Psychosom Res 2008;65:229–38.doi:10.1016/j.jpsychores.2008.05.019
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      2. Papaioannou A ,
      3. Fraidakis O ,
      4. Michaloudis D , et al
      . The impact of the type of anaesthesia on cognitive status and delirium during the first postoperative days in elderly patients. Eur J Anaesthesiol 2005;22:492–9.doi:10.1017/S0265021505000840
      OpenUrlCrossRefPubMed
    6. ↵
      2. Chen W ,
      3. Ke X ,
      4. Wang X , et al
      . Prevalence and risk factors for postoperative delirium in total joint arthroplasty patients: a prospective study. Gen Hosp Psychiatry 2017;46:55–61.doi:10.1016/j.genhosppsych.2017.03.008
      OpenUrl
    7. ↵
      2. Bruce AJ ,
      3. Ritchie CW ,
      4. Blizard R , et al
      . The incidence of delirium associated with orthopedic surgery: a meta-analytic review. Int Psychogeriatr 2007;19:197–214.doi:10.1017/S104161020600425X
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      2. Scott JE ,
      3. Mathias JL ,
      4. Kneebone AC
      . Incidence of delirium following total joint replacement in older adults: a meta-analysis. Gen Hosp Psychiatry 2015;37:223–9.doi:10.1016/j.genhosppsych.2015.02.004
      OpenUrlCrossRefPubMed
    9. ↵
      2. Cox G ,
      3. Tzioupis C ,
      4. Calori GM , et al
      . Cerebral fat emboli: a trigger of post-operative delirium. Injury 2011;42 Suppl 4(Suppl 4):S6–S10.doi:10.1016/S0020-1383(11)70005-5
      OpenUrl
    10. ↵
      2. Weinstein SM ,
      3. Poultsides L ,
      4. Baaklini LR , et al
      . Postoperative delirium in total knee and hip arthroplasty patients: a study of perioperative modifiable risk factors. Br J Anaesth 2018;120:999–1008.doi:10.1016/j.bja.2017.12.046
      OpenUrl
    11. ↵
      2. Makadia R ,
      3. Ryan PB
      . Transforming the premier perspective Hospital database into the observational medical outcomes partnership (OMOP) common data model. EGEMS 2014;2.doi:10.13063/2327-9214.1110
    12. ↵
    13. ↵
      2. Moineddin R ,
      3. Matheson FI ,
      4. Glazier RH
      . A simulation study of sample size for multilevel logistic regression models. BMC Med Res Methodol 2007;7:34.doi:10.1186/1471-2288-7-34
    14. ↵
      2. Kim DH ,
      3. Lee J ,
      4. Kim CA , et al
      . Evaluation of algorithms to identify delirium in administrative claims and drug utilization database. Pharmacoepidemiol Drug Saf 2017;26:945–53.doi:10.1002/pds.4226
      OpenUrlCrossRefPubMed
    15. ↵
      2. Quan H ,
      3. Sundararajan V ,
      4. Halfon P , et al
      . Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005;43:1130–9.doi:10.1097/01.mlr.0000182534.19832.83
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      Lexicomp® online opioid agonist conversion calculator. Available: Http://online.Lexi.Com/lco/action/calc/calculator/70050.
    17. ↵
      Globalrph opioid analgesic converter. Available: Http://www.Globalrph.Com/narcoticonv.Htm
    18. ↵
      2. Yang D
      . A unified approach to measuring the effect size between two groups using sas®. In: dongsheng Yang and jarrod E. dalton departments of quantitative health sciences and outcomes research Cleveland clinic SAS global forum 2012 statistics and data analysis. Sas gloabl forum 2012, 2012: 335.
    19. ↵
      2. Witte JS ,
      3. Greenland S ,
      4. Kim LL , et al
      . Multilevel modeling in epidemiology with GLIMMIX. Epidemiology 2000;11:684–8.doi:10.1097/00001648-200011000-00012
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      2. Perneger TV
      . What's wrong with Bonferroni adjustments. BMJ 1998;316:1236–8.doi:10.1136/bmj.316.7139.1236
      OpenUrlFREE Full Text
    21. ↵
      2. Petersen PB ,
      3. Jørgensen CC ,
      4. Kehlet H , et al
      . Delirium after fast-track hip and knee arthroplasty - a cohort study of 6331 elderly patients. Acta Anaesthesiol Scand 2017;61:767–72.doi:10.1111/aas.12932
      OpenUrl
    22. ↵
      2. Aldecoa C ,
      3. Bettelli G ,
      4. Bilotta F , et al
      . European Society of Anaesthesiology evidence-based and consensus-based guideline on postoperative delirium. Eur J Anaesthesiol 2017;34:192–214.doi:10.1097/EJA.0000000000000594
      OpenUrl
    23. ↵
      2. Rothberg MB ,
      3. Herzig SJ ,
      4. Pekow PS , et al
      . Association between sedating medications and delirium in older inpatients. J Am Geriatr Soc 2013;61:923–30.doi:10.1111/jgs.12253
      OpenUrlCrossRefPubMed
    24. ↵
      2. Xu HF ,
      3. White RS ,
      4. Sastow DL , et al
      . Medicaid insurance as primary payer predicts increased mortality after total hip replacement in the state inpatient databases of California, Florida and new York. J Clin Anesth 2017;43:24–32.doi:10.1016/j.jclinane.2017.09.008
      OpenUrl
    25. ↵
      2. Memtsoudis SG ,
      3. Poeran J ,
      4. Zubizarreta N , et al
      . Anesthetic care for orthopedic patients: is there a potential for differences in care? Anesthesiology 2016;124:608–23.doi:10.1097/ALN.0000000000001004
      OpenUrl
    26. ↵
      2. Browne JA ,
      3. Novicoff WM ,
      4. D’Apuzzo MR
      . Medicaid payer status is associated with in-hospital morbidity and resource utilization following primary total joint arthroplasty. J Bone Joint Surg Am 2014;96:e180:1–1-6.doi:10.2106/JBJS.N.00133
      OpenUrlAbstract/FREE Full Text
    27. ↵
      2. Davis N ,
      3. Lee M ,
      4. Lin AY , et al
      . Postoperative cognitive function following general versus regional anesthesia: a systematic review. J Neurosurg Anesthesiol 2014;26:369–76.doi:10.1097/ANA.0000000000000120
      OpenUrlCrossRefPubMed
    28. ↵
      2. Praticò C ,
      3. Quattrone D ,
      4. Lucanto T , et al
      . Drugs of anesthesia acting on central cholinergic system may cause post-operative cognitive dysfunction and delirium. Med Hypotheses 2005;65:972–82.doi:10.1016/j.mehy.2005.05.037
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      2. Hole A ,
      3. Terjesen T ,
      4. Breivik H
      . Epidural versus general anaesthesia for total hip arthroplasty in elderly patients. Acta Anaesthesiol Scand 1980;24:279–87.doi:10.1111/j.1399-6576.1980.tb01549.x
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      2. Rasmussen LS ,
      3. Johnson T ,
      4. Kuipers HM , et al
      . Does anaesthesia cause postoperative cognitive dysfunction? a randomised study of regional versus general anaesthesia in 438 elderly patients. Acta Anaesthesiol Scand 2003;47:260–6.doi:10.1034/j.1399-6576.2003.00057.x
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      2. Marcantonio ER et al
      . The relationship of postoperative delirium with psychoactive medications. JAMA 1994;272:1518–22.doi:10.1001/jama.1994.03520190064036
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. American Society of Anesthesiologists Task Force on Acute Pain Management
      . Practice guidelines for acute pain management in the perioperative setting: an updated report by the American Society of Anesthesiologists Task force on acute pain management. Anesthesiology 2012;116:248–73.doi:10.1097/ALN.0b013e31823c1030
      OpenUrlCrossRefPubMedWeb of Science
    33. ↵
      2. Hamilton TW ,
      3. Strickland LH ,
      4. Pandit HG
      . A meta-analysis on the use of gabapentinoids for the treatment of acute postoperative pain following total knee arthroplasty. J Bone Joint Surg Am 2016;98:1340–50.doi:10.2106/JBJS.15.01202
      OpenUrlAbstract/FREE Full Text
    34. ↵
      2. Martinez V ,
      3. Carles M ,
      4. Marret E , et al
      . Perioperative use of gabapentinoids in France. mismatch between clinical practice and scientific evidence. Anaesth Crit Care Pain Med 2018;37:43–7.doi:10.1016/j.accpm.2017.01.010
      OpenUrl
    35. ↵
      2. Dong J ,
      3. Li W ,
      4. Wang Y
      . The effect of pregabalin on acute postoperative pain in patients undergoing total knee arthroplasty: a meta-analysis. Int J Surg 2016;34:148–60.doi:10.1016/j.ijsu.2016.08.521
      OpenUrl
    36. ↵
      2. Clarke H ,
      3. Bonin RP ,
      4. Orser BA , et al
      . The prevention of chronic postsurgical pain using gabapentin and pregabalin: a combined systematic review and meta-analysis. Anesth Analg 2012;115:428–42.doi:10.1213/ANE.0b013e318249d36e
      OpenUrlCrossRefPubMedWeb of Science
    37. ↵
      2. Buvanendran A ,
      3. Kroin JS ,
      4. Della Valle CJ , et al
      . Perioperative oral pregabalin reduces chronic pain after total knee arthroplasty: a prospective, randomized, controlled trial. Anesth Analg 2010;110:199–207.doi:10.1213/ANE.0b013e3181c4273a
      OpenUrlCrossRefPubMedWeb of Science
    38. ↵
      2. Lunn TH ,
      3. Husted H ,
      4. Laursen MB , et al
      . Analgesic and sedative effects of perioperative gabapentin in total knee arthroplasty: a randomized, double-blind, placebo-controlled dose-finding study. Pain 2015;156:2438–48.doi:10.1097/j.pain.0000000000000309
      OpenUrlCrossRefPubMed
    39. ↵
      2. Mishriky BM ,
      3. Waldron NH ,
      4. Habib AS
      . Impact of pregabalin on acute and persistent postoperative pain: a systematic review and meta-analysis. Br J Anaesth 2015;114:10–31.doi:10.1093/bja/aeu293
      OpenUrlCrossRefPubMed
    40. ↵
      2. Eipe N ,
      3. Penning J ,
      4. Yazdi F , et al
      . Perioperative use of pregabalin for acute pain-a systematic review and meta-analysis. Pain 2015;156:1284–300.doi:10.1097/j.pain.0000000000000173
      OpenUrlCrossRefPubMed
    41. ↵
      2. Mathiesen O ,
      3. Wetterslev J ,
      4. Kontinen VK , et al
      . Adverse effects of perioperative paracetamol, NSAIDs, glucocorticoids, gabapentinoids and their combinations: a topical review. Acta Anaesthesiol Scand 2014;58:1182–98.doi:10.1111/aas.12380
      OpenUrlCrossRefPubMed
    42. ↵
      2. Myhre M ,
      3. Diep LM ,
      4. Stubhaug A
      . Pregabalin has analgesic, ventilatory, and cognitive effects in combination with remifentanil. Anesthesiology 2016;124:141–9.doi:10.1097/ALN.0000000000000913
      OpenUrl
    43. ↵
      2. White PF ,
      3. Tufanogullari B ,
      4. Taylor J , et al
      . The effect of pregabalin on preoperative anxiety and sedation levels: a dose-ranging study. Anesth Analg 2009;108:1140–5.doi:10.1213/ane.0b013e31818d40ce
      OpenUrlCrossRefPubMedWeb of Science
    44. ↵
      2. Vaurio LE ,
      3. Sands LP ,
      4. Wang Y , et al
      . Postoperative delirium: the importance of pain and pain management. Anesth Analg 2006;102:1267–73.doi:10.1213/01.ane.0000199156.59226.af
      OpenUrlCrossRefPubMedWeb of Science
    45. ↵
      2. Robinson S ,
      3. Vollmer C
      . Undermedication for pain and precipitation of delirium. Medsurg Nurs 2010;19:79–83. quiz 84.
      OpenUrlPubMed
    46. ↵
      2. Morrison RS ,
      3. Magaziner J ,
      4. Gilbert M , et al
      . Relationship between pain and opioid analgesics on the development of delirium following hip fracture. J Gerontol A Biol Sci Med Sci 2003;58:M76–M81.doi:10.1093/gerona/58.1.M76
      OpenUrlCrossRefPubMed
    47. ↵
      2. Rao SS ,
      3. Cherukuri M
      . Management of hip fracture: the family physician's role. Am Fam Physician 2006;73:2195–200.
      OpenUrlPubMedWeb of Science
    48. ↵
      2. Agnoletti V ,
      3. Ansaloni L ,
      4. Catena F , et al
      . Postoperative delirium after elective and emergency surgery: analysis and checking of risk factors. A study protocol. BMC Surg 2005;5:12.doi:10.1186/1471-2482-5-12
    49. ↵
      2. Lynch EP ,
      3. Lazor MA ,
      4. Gellis JE , et al
      . The impact of postoperative pain on the development of postoperative delirium. Anesth Analg 1998;86:781–5.doi:10.1213/00000539-199804000-00019
      OpenUrlCrossRefPubMedWeb of Science
    50. ↵
      2. Griffin Éadaoin W ,
      3. Skelly DT ,
      4. Murray CL , et al
      . Cyclooxygenase-1-Dependent prostaglandins mediate susceptibility to systemic inflammation-induced acute cognitive dysfunction. J Neurosci 2013;33:15248–58.doi:10.1523/JNEUROSCI.6361-11.2013
      OpenUrlAbstract/FREE Full Text
    51. ↵
      2. Beloosesky Y ,
      3. Hendel D ,
      4. Weiss A , et al
      . Cytokines and C-reactive protein production in hip-fracture-operated elderly patients. J Gerontol A Biol Sci Med Sci 2007;62:420–6.doi:10.1093/gerona/62.4.420
      OpenUrlCrossRefPubMedWeb of Science
    52. ↵
      2. Deljou A ,
      3. Hedrick SJ ,
      4. Portner ER , et al
      . Pattern of perioperative gabapentinoid use and risk for postoperative naloxone administration. Br J Anaesth 2018;120:798–806.doi:10.1016/j.bja.2017.11.113
      OpenUrl

    Footnotes

    • Contributors SM helped design the study, conduct the study, analyze the data, write the manuscript and has approved the final manuscript. CC helped design the study, conduct the study, analyze the data, and write the manuscript and has approved the final manuscript. NZ helped conduct the study, analyze the data, and write the manuscript and has approved the final manuscript. SMW helped conduct the study, analyze the data and write the manuscript and has approved the final manuscript. JL helped conduct the study, analyze the data and write the manuscript and has approved the final manuscript. DHK helped conduct the study, analyze the data and write the manuscript and has approved the final manuscript. LP helped conduct the study, analyze the data, and write the manuscript and has approved the final manuscript. MMB helped, conduct the study, analyze the data, and write the manuscript and has approved the final manuscript. MM helped conduct the study, analyze the data, and write the manuscript and has approved the final manuscript. JP helped design the study, conduct the study, analyze the data, and write the manuscript and has approved the final manuscript.

    • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

    • Competing interests None declared.

    • Patient consent for publication Not required.

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

    • Data availability statement Data may be obtained from a third party and are not publicly available.