Publications

INTRODUCTION: The large number of heterogeneous instruments in active use for identification of delirium prevents direct comparison of studies and the ability to combine results. In a recent systematic review we performed, we recommended four commonly used and well-validated instruments and subsequently harmonized them using advanced psychometric methods to develop an item bank, the Delirium Item Bank (DEL-IB). The goal of the present study was to find optimal cut-points on four existing instruments and to demonstrate use of the DEL-IB to create new instruments.

METHODS: We used a secondary analysis and simulation study based on data from three previous studies of hospitalized older adults (age 65+ years) in the USA, Ireland, and Belgium. The combined dataset included 600 participants, contributing 1,623 delirium assessments, and an overall incidence of delirium of about 22%. The measurements included the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition diagnostic criteria for delirium, Confusion Assessment Method (long form and short form), Delirium Observation Screening Scale, Delirium Rating Scale-Revised-98 (total and severity scores), and Memorial Delirium Assessment Scale (MDAS).

RESULTS: We identified different cut-points for each existing instrument to optimize sensitivity or specificity, and compared instrument performance at each cut-point to the author-defined cut-point. For instance, the cut-point on the MDAS that maximizes both sensitivity and specificity was at a sum score of 6 yielding 89% sensitivity and 79% specificity. We then created four new example instruments (two short forms and two long forms) and evaluated their performance characteristics. In the first example short form instrument, the cut-point that maximizes sensitivity and specificity was at a sum score of 3 yielding 90% sensitivity, 81% specificity, 30% positive predictive value, and 99% negative predictive value.

DISCUSSION/CONCLUSION: We used the DEL-IB to better understand the psychometric performance of widely used delirium identification instruments and scorings, and also demonstrated its use to create new instruments. Ultimately, we hope that the DEL-IB might be used to create optimized delirium identification instruments and to spur the development of a unified approach to identify delirium.

The identification of delirium in electronic health records (EHRs) remains difficult due to inadequate assessment or under-documentation. The purpose of this research is to present a classification model that identifies delirium using retrospective EHR data. Delirium was confirmed with the Confusion Assessment Method for the Intensive Care Unit. Age, sex, Elixhauser comorbidity index, drug exposures, and diagnoses were used as features. The model was developed based on the Columbia University Irving Medical Center EHR data and further validated with the Medical Information Mart for Intensive Care III dataset. Seventy-six patients from Surgical/Cardiothoracic ICU were included in the model. The logistic regression model achieved the best performance in identifying delirium; mean AUC of 0.874 ± 0.033. The mean positive predictive value of the logistic regression model was 0.80. The model promises to identify delirium cases with EHR data, thereby enable a sustainable infrastructure to build a retrospective cohort of delirium.

BACKGROUND: The REGAIN (Regional versus General Anesthesia for Promoting Independence after Hip Fracture) trial found similar ambulation and survival at 60 days with spinal versus general anesthesia for hip fracture surgery. Trial outcomes evaluating pain, prescription analgesic use, and patient satisfaction have not yet been reported.

OBJECTIVE: To compare pain, analgesic use, and satisfaction after hip fracture surgery with spinal versus general anesthesia.

DESIGN: Preplanned secondary analysis of a pragmatic randomized trial. (ClinicalTrials.gov: NCT02507505).

SETTING: 46 U.S. and Canadian hospitals.

PARTICIPANTS: Patients aged 50 years or older undergoing hip fracture surgery.

INTERVENTION: Spinal or general anesthesia.

MEASUREMENTS: Pain on postoperative days 1 through 3; 60-, 180-, and 365-day pain and prescription analgesic use; and satisfaction with care.

RESULTS: A total of 1600 patients were enrolled. The average age was 78 years, and 77% were women. A total of 73.5% (1050 of 1428) of patients reported severe pain during the first 24 hours after surgery. Worst pain over the first 24 hours after surgery was greater with spinal anesthesia (rated from 0 [no pain] to 10 [worst pain imaginable]; mean difference, 0.40 [95% CI, 0.12 to 0.68]). Pain did not differ across groups at other time points. Prescription analgesic use at 60 days occurred in 25% (141 of 563) and 18.8% (108 of 574) of patients assigned to spinal and general anesthesia, respectively (relative risk, 1.33 [CI, 1.06 to 1.65]). Satisfaction was similar across groups.

LIMITATION: Missing outcome data and multiple outcomes assessed.

CONCLUSION: Severe pain is common after hip fracture. Spinal anesthesia was associated with more pain in the first 24 hours after surgery and more prescription analgesic use at 60 days compared with general anesthesia.

PRIMARY FUNDING SOURCE: Patient-Centered Outcomes Research Institute.

Despite its devastating clinical and societal impact, approaches to treat delirium in older adults remain elusive, making it important to identify factors that may confer resilience to this syndrome. Here, we investigated a cohort of 93 cognitively normal older patients undergoing elective surgery recruited as part of the Successful Aging after Elective Surgery study. Each participant was classified either as a SuperAger (n = 19) or typically aging older adult (n = 74) based on neuropsychological criteria, where the former was defined as those older adults whose memory function rivals that of young adults. We compared these subgroups to examine the role of preoperative memory function in the incidence and severity of postoperative delirium. We additionally investigated the association between indices of postoperative delirium symptoms and cortical thickness in functional networks implicated in SuperAging based on structural magnetic resonance imaging data that were collected preoperatively. We found that SuperAging confers the real-world benefit of resilience to delirium, as shown by lower (i.e. zero) incidence of postoperative delirium and decreased severity scores compared with typical older adults. Furthermore, greater baseline cortical thickness of the anterior mid-cingulate cortex-a key node of the brain's salience network that is also consistently implicated in SuperAging-predicted lower postoperative delirium severity scores in all patients. Taken together, these findings suggest that baseline memory function in older adults may be a useful predictor of postoperative delirium risk and severity and that superior memory function may contribute to resilience to delirium. In particular, the integrity of the anterior mid-cingulate cortex may be a potential biomarker of resilience to delirium, pointing to this region as a potential target for preventive or therapeutic interventions designed to mitigate the risk or consequences of developing this prevalent clinical syndrome.

IMPORTANCE: Gabapentin has been increasingly used as part of a multimodal analgesia regimen to reduce opioid use in perioperative pain management. However, the safety of perioperative gabapentin use among older patients remains uncertain.

OBJECTIVE: To examine in-hospital adverse clinical events associated with perioperative gabapentin use among older patients undergoing major surgery.

DESIGN, SETTING, AND PARTICIPANTS: This retrospective cohort study using data from the Premier Healthcare Database included patients aged 65 years or older who underwent major surgery at US hospitals within 7 days of hospital admission from January 1, 2009, to March 31, 2018, and did not use gabapentin before surgery. Data were analyzed from June 14, 2021, to May 23, 2022.

EXPOSURES: Gabapentin use within 2 days after surgery.

MAIN OUTCOMES AND MEASURES: The primary outcome was delirium, identified using diagnosis codes, and secondary outcomes were new antipsychotic use, pneumonia, and in-hospital death between postoperative day 3 and hospital discharge. To reduce confounding, 1:1 propensity score matching was performed. Risk ratios (RRs) and risk differences (RDs) with 95% CIs were estimated.

RESULTS: Among 967 547 patients before propensity score matching (mean [SD] age, 76.2 [7.4] years; 59.6% female), the rate of perioperative gabapentin use was 12.3% (119 087 patients). After propensity score matching, 237 872 (118 936 pairs) gabapentin users and nonusers (mean [SD] age, 74.5 [6.7] years; 62.7% female) were identified. Compared with nonusers, gabapentin users had increased risk of delirium (4040 [3.4%] vs 3148 [2.6%]; RR, 1.28 [95% CI, 1.23-1.34]; RD, 0.75 [95% CI, 0.75 [0.61-0.89] per 100 persons), new antipsychotic use (944 [0.8%] vs 805 [0.7%]; RR, 1.17 [95% CI, 1.07-1.29]; RD, 0.12 [95% CI, 0.05-0.19] per 100 persons), and pneumonia (1521 [1.3%] vs 1368 [1.2%]; RR, 1.11 [95% CI, 1.03-1.20]; RD, 0.13 [95% CI, 0.04-0.22] per 100 persons), but there was no difference in in-hospital death (362 [0.3%] vs 354 [0.2%]; RR, 1.02 [95% CI, 0.88-1.18]; RD, 0.00 [95% CI, -0.04 to 0.05] per 100 persons). Risk of delirium among gabapentin users was greater in subgroups with high comorbidity burden than in those with low comorbidity burden (combined comorbidity index <4 vs ≥4: RR, 1.20 [95% CI, 1.13-1.27] vs 1.40 [95% CI, 1.30-1.51]; RD, 0.41 [95% CI, 0.28-0.53] vs 2.66 [95% CI, 2.08-3.24] per 100 persons) and chronic kidney disease (absence vs presence: RR, 1.26 [95% CI, 1.19-1.33] vs 1.38 [95% CI, 1.27-1.49]; RD, 0.56 [95% CI, 0.42-0.69] vs 1.97 [95% CI, 1.49-2.46] per 100 persons).

CONCLUSION AND RELEVANCE: In this cohort study, perioperative gabapentin use was associated with increased risk of delirium, new antipsychotic use, and pneumonia among older patients after major surgery. These results suggest careful risk-benefit assessment before prescribing gabapentin for perioperative pain management.

BACKGROUND: Homocysteine and C-reactive protein (CRP) may serve as biomarkers of postoperative delirium. We set out to compare the role of blood concentration of homocysteine versus CRP in predicting postoperative delirium in patients.

MATERIALS AND METHODS: In this prospective observational cohort study, the plasma concentration of preoperative homocysteine and postoperative CRP was measured. Delirium incidence and severity within 3 days postoperatively were determined using the Confusion Assessment Method and Confusion Assessment Method-Severity algorithm.

RESULTS: Of 143 participants [69% female, median (interquartile range, 25th-75th) age of 71 (67-76) years] who had knee or hip surgery under general anesthesia, 44 (31%) participants developed postoperative delirium. Postoperative plasma concentration of CRP was associated with postoperative delirium incidence [adjusted odds ratio (OR) per one standard deviation change in CRP: 1.51; 95% Confidence Interval (CI): 1.05, 2.16; P = 0.026], and severity [in which each one standard deviation increase in postoperative CRP was associated with a 0.47 point (95% CI: 0.18-0.76) increase in the severity of delirium, P = 0.002] after adjusting age, sex, preoperative Mini-Mental State Examination score and the days when postoperative CRP was measured. A statistically significant interaction (adjusted P = 0.044) was also observed, in which the association between postoperative plasma concentration of CRP and postoperative delirium incidence was stronger in the participants with lower preoperative plasma concentrations of homocysteine compared to those with higher preoperative levels.

CONCLUSION: Pending validation studies, these data suggest that preoperative plasma concentration of homocysteine modifies the established association between postoperative plasma concentration of CRP and postoperative delirium incidence.

OBJECTIVES: To characterize the proteomic signature of surgery in older adults and association with postoperative outcomes.

SUMMARY OF BACKGROUND DATA: Circulating plasma proteins can reflect the physiological response to and clinical outcomes after surgery.

METHODS: Blood plasma from older adults undergoing elective surgery was analyzed for 1305 proteins using SOMAscan. Surgery-associated proteins underwent Ingenuity Pathways Analysis. Selected surgery-associated proteins were independently validated using Luminex or enzyme-linked immunosorbent assay methods. Generalized linear models estimated correlations with postoperative outcomes.

RESULTS: Plasma from a subcohort (n = 36) of the Successful Aging after Elective Surgery (SAGES) study was used for SOMAscan. Systems biology analysis of 110 proteins with Benjamini-Hochberg (BH) corrected P value ≤0.01 and an absolute foldchange (|FC|) ≥1.5 between postoperative day 2 (POD2) and preoperative (PREOP) identified functional pathways with major effects on pro-inflammatory proteins. Chitinase-3-like protein 1 (CHI3L1), C-reactive protein (CRP), and interleukin-6 (IL-6) were independently validated in separate validation cohorts from SAGES (n = 150 for CRP, IL-6; n = 126 for CHI3L1). Foldchange CHI3L1 and IL-6 were associated with increased postoperative complications [relative risk (RR) 1.50, 95% confidence interval (95% CI) 1.21-1.85 and RR 1.63, 95% CI 1.18-2.26, respectively], length of stay (RR 1.35, 95% CI 0.77-1.92 and RR 0.98, 95% CI 0.52-1.45), and risk of discharge to postacute facility (RR 1.15, 95% CI 1.04-1.26 and RR 1.11, 95% CI 1.04-1.18); POD2 and PREOP CRP difference was associated with discharge to postacute facility (RR 1.14, 95% CI 1.04-1.25).

CONCLUSION: SOMAscan can identify novel and clinically relevant surgery-induced protein changes. Ultimately, proteomics may provide insights about pathways by which surgical stress contributes to postoperative outcomes.

BACKGROUND: Our objective was to assess the performance of machine learning methods to predict post-operative delirium using a prospective clinical cohort.

METHODS: We analyzed data from an observational cohort study of 560 older adults (≥ 70 years) without dementia undergoing major elective non-cardiac surgery. Post-operative delirium was determined by the Confusion Assessment Method supplemented by a medical chart review (N = 134, 24%). Five machine learning algorithms and a standard stepwise logistic regression model were developed in a training sample (80% of participants) and evaluated in the remaining hold-out testing sample. We evaluated three overlapping feature sets, restricted to variables that are readily available or minimally burdensome to collect in clinical settings, including interview and medical record data. A large feature set included 71 potential predictors. A smaller set of 18 features was selected by an expert panel using a consensus process, and this smaller feature set was considered with and without a measure of pre-operative mental status.

RESULTS: The area under the receiver operating characteristic curve (AUC) was higher in the large feature set conditions (range of AUC, 0.62-0.71 across algorithms) versus the selected feature set conditions (AUC range, 0.53-0.57). The restricted feature set with mental status had intermediate AUC values (range, 0.53-0.68). In the full feature set condition, algorithms such as gradient boosting, cross-validated logistic regression, and neural network (AUC = 0.71, 95% CI 0.58-0.83) were comparable with a model developed using traditional stepwise logistic regression (AUC = 0.69, 95% CI 0.57-0.82). Calibration for all models and feature sets was poor.

CONCLUSIONS: We developed machine learning prediction models for post-operative delirium that performed better than chance and are comparable with traditional stepwise logistic regression. Delirium proved to be a phenotype that was difficult to predict with appreciable accuracy.