Delirium after elective surgery among elderly patients taking statins
=====================================================================

* Donald A. Redelmeier, MD MSHSR
* Deva Thiruchelvam, MSc
* Nick Daneman MD

*   © 2008 Canadian Medical Association

## Abstract

**Background:** Postoperative delirium after elective surgery is frequent and potentially serious. We sought to determine whether the use of statin medications was associated with a higher risk of postoperative delirium than other medications that do not alter microvascular autoregulation.

**Methods:** We conducted a retrospective cohort analysis of 284 158 consecutive patients in Ontario aged 65 years and older who were admitted for elective surgery. We identified exposure to statins from outpatient pharmacy records before admission. We identified delirium by examining hospital records after surgery.

**Results:** About 7% (*n* = 19 501) of the patients were taking statins. Overall, 3195 patients experienced postoperative delirium; the rate was significantly higher among patients taking statins (14 per 1000) than among those not taking statins (11 per 1000) (odds ratio [OR] 1.30, 95% confidence interval [CI] 1.15–1.47, *p* < 0.001). The increased risk of postoperative delirium persisted after we adjusted for multiple demographic, medical and surgical factors (OR 1.28, 95% CI 1.12–1.46) and exceeded the increased risk of delirium associated with prolonging surgery by 30 minutes (OR 1.20, 95% CI 1.19–1.21). The relative risk associated with statin use was somewhat higher among patients who had noncardiac surgery than among those who had cardiac surgery (adjusted OR 1.33, 95% CI 1.16–1.53), and extended to more complicated cases of delirium. We did not observe an increased risk of delirium with 20 other cardiac or noncardiac medications.

**Interpretation:** The use of statins is associated with an increased risk of postoperative delirium among elderly patients undergoing elective surgery.

Delirium is an acute change in mental status that is worrisome to patients and families, especially after elective surgery. This condition may contribute to delays in extubation, a prolonged need for intensive care, increased risk of nosocomial infections and about a 1-week rise in total length of stay in hospital for the average patient.1,2 Delirium also disrupts many specific aspects of care, including the administration of medications, treatment of wounds, physiotherapy, nutrition, hygiene, discharge planning and dignity.3 The management of delirium is awkward and may lead to a cascade of nonspecific testing and sedation, with an average net increase in hospital costs of $2500 per patient.4 In some cases, the delirium never completely disappears, and the patient is left with a degree of permanent disability.5

The causes of postoperative delirium are not well understood. Hypoglycemia, hypoxemia and hypotension are all possible and correctable, but they rarely have an immediate resolution.6 Medical imaging studies typically do not show specific changes; however, they may show markers of prior stroke or other lesions. One underlying factor may be cerebral ischemia secondary to inadequate perfusion. Altered cerebral perfusion may result in altered metabolism, an increased predisposition to drug toxicity or other factors during anesthesia and surgery.7 Cerebral ischemia may also explain commonly observed risk factors for postoperative delirium, including advanced age, baseline cognitive dysfunction and the failure of drug antagonists, major tranquilizers or modern volatile anesthetics to prevent postoperative delirium.8,9,10

Statins have pleiotropic properties that alter the tone of smooth muscle in small blood vessels. Experiments on endothelial cells indicate that these changes are mediated by expression of endothelial nitric oxide synthase that is unrelated to cholesterol levels or vascular disease.11 In turn, activity of endothelial nitric oxide synthase contributes to arteriolar vasodilation by relaxing the surrounding smooth-muscle cells, thereby shifting the distribution of blood flow in the microvasculature of the brain. This can compromise individual neurons even if aggregate blood flow is maintained.12 These effects can be beneficial for reducing the size of stroke or other long-term neurologic disorders; however, altered cerebral blood flow autoregulation might predispose patients to delirium after anesthesia.13–15

We sought to determine whether the use of statins was associated with postoperative delirium among elderly patients undergoing elective surgery.

## Methods

### Patient selection

Using the Canadian Institutes for Health Information database, we identified consecutive patients aged 65 years and older who underwent elective surgery in all Ontario hospitals between Apr. 1, 1992, and Apr. 1, 2002. We selected this timeframe because data for these years were available for analysis. We did not include outpatients, those who had day surgery or patients younger than 65 years because the rate of delirium in such circumstances is generally low. To reduce confounding from pre-existing illness, we initially excluded patients with major vascular disease, as evidenced by long-term use of nitrates or β-blockers; however, we relaxed this restriction in a secondary analysis.16

We received institutional review board approval from the Sunnybrook Hospital Ethics Committee and used confidentiality safeguards of the Institute for Clinical Evaluative Sciences.

### Hospital records

We gathered population-based data that counted each patient only once. We analyzed only the first admission for patients who had more than 1 elective surgical procedure during the study interval (analyses based on separate admissions yielded more extreme results and are not reported). In cases where patients were transferred to a different hospital, we counted outcomes according to the original hospital admission. All databases have been used extensively in past research.17,18

### Statin prescriptions

For each patient, we searched prescription records from the Ontario Drug Benefit database for the year before admission, reasoning that statins would customarily be continued perioperatively. We classified people who received 2 or more prescriptions for a statin in the year before surgery (including at least 1 prescription in the 90 days before surgery) as receiving this medication on an ongoing basis. Otherwise, we classified the patient as not receiving a statin. This strategy assured that outcome ascertainment was blind to exposure status, free of reverse-causality bias and conservative in design.19 The specific statins were atorvastatin, simvastatin, pravastatin, lovastatin, fluvastatin, rosuvastatin and cerivastatin. A previous validation study indicated that the Ontario Drug Benefit database had an accuracy rate of 99% using pharmacy records as the reference standard.20

### Risk factors for delirium

We gathered data on patient and surgical factors that could potentially contribute to delirium, since established risk factors are controversial. We derived the patients' age, sex and income quintile at the time of admission from the Ontario Registered Persons Database. We identified patients receiving drug therapy for dementia at baseline by assessing whether they were prescribed a cholinesterase inhibitor (e.g., donepezil) in the year before admission. We determined the use of other important baseline neuropsychiatric treatments with prescriptions for antipsychotics (e.g., risperidone), antidepressants (e.g., citalopram) and benzodiazepines (e.g., lorazepam) in the year before admission. We derived information on additional medical treatments from corresponding long-term medications, in accordance with prior research. We determined operative procedures based on the primary surgical procedures, and we classified them as either cardiac or noncardiac. We further distinguished noncardiac surgeries by anatomic site.

### Duration of surgery

An innovative aspect of our study was to develop a method for determining the duration of surgery for each patient. This was necessary because prior research indicated that surgical duration was a risk factor for delirium and could vary substantially among patients undergoing the same primary procedure. Our method relied on anesthesiology billing fees linked to individual patients and reimbursed in 15-minute intervals. This method was analogous to estimating anesthesia and surgical times from Medicare claims (Part B data) in the United States. We examined the reliability and validity of the method and found high correspondence to chart review (*R*2 = 88%).21

### Other medications

We examined other medications prescribed for cardiovascular disease to check whether findings with statins differed from those with other long-term drug therapies that are not known to have pleiotropic properties on the microvasculature. These included nonstatin lipid-lowering medications (e.g., fenofibrate), antihypertensive medications (e.g., hydrochlorothiazide), loop diuretics (e.g., furosemide) and miscellaneous cardiovascular agents (e.g., digoxin). Another set of analyses focused on medications that have indirect effects on the vascular system, including antiplatelet agents (e.g., clopidogrel), orally administered anticoagulants (e.g., coumadin) and other vascular agents (e.g., pentoxifylline). The final set of medications included common drugs that have no major effect on the cardiovascular system (e.g., omeprazole).

### Outcome assessment

The primary outcome was the development of delirium during the patient's stay in hospital, based on the International Classification of Diseases codes 293.0–293.9. Surgical complications such as delirium are not always identified by clinicians, recorded in the patient's chart or entered into databases. Hence, the codes are specific (about 98%), but not sensitive (about 35%).22 A separate chart review of the codes showed a positive predictive value of 100% when compared with doctors' and nurses' progress notes (95% confidence interval [CI] 96%–100%). To check the robustness of our results and to examine the more severe spectrum of postoperative delirium, we considered 7 complex combinations of outcomes: we examined data for patients who experienced postoperative delirium and who also received a computed tomography scan while in hospital, had a wound infection develop during their stay in hospital, experienced a myocardial infarction while in hospital, needed home nursing care after discharge, received ongoing sedatives after discharge, required hospital readmission or died in hospital after surgery.

### Statistical analysis

We used the χ2 test to assess the proportion of patients who experienced delirium, comparing those patients taking statins with those who were not taking statins. We used logistic regression to adjust this comparison for patient characteristics, since the time of onset of delirium was not recorded (estimates computed with 95% CIs). We tested generalizability by repeating analyses in 3 more patient groups: those excluded because data on the duration of anesthesia were unavailable, those excluded because of pre-existing major vascular disease and those admitted for emergency surgery. We also tested for selection bias through a secondary analysis of patients who had received a statin in the past but not in the 90 days before surgery. In addition, we examined a cohort matched by propensity score to explore whether the main finding was due to hidden confounding.23

## Results

A total of 541 827 elective surgeries were performed on 454 084 patients during the study interval. The patients were dispersed across 246 hospitals. We observed no major trends over the years. Overall, we excluded 100 832 patients because anesthesiology records were unavailable. We excluded a further 69 094 patients because they had major vascular disease. This left 284 158 patients for analysis. The typical patient took multiple medications on an outpatient basis; underwent an abdominal, musculoskeletal or urogenital procedure; and had a mean duration of surgery of about 115 minutes. The most common outpatient medications (received by about 15% of patients in each case) were benzodiazepines, angiotensin-converting-enzyme inhibitors, calcium-channel blockers and gastric acid suppressants.

About 7% (*n* = 19 501) of the patients were taking statins before surgery, compared with about 93% (*n* = 264 657) who were not taking statins before surgery. There was about a 2-year difference in mean age between the 2 groups, with more younger patients than older patients using statins. Otherwise, we found no major differences between the 2 groups in demographic characteristics, use of neuropsychiatric or common noncardiovascular medications, or the number of noncardiovascular surgeries performed (Table 1). As expected, use of cardiac medications, cardiac surgeries and peripheral vascular surgeries were more common among patients who used statins than among patients who did not use statins.

View this table:
[Table1](http://www.cmaj.ca/content/179/7/645/T1)

Table 1.

Overall, postoperative delirium was diagnosed in 3195 patients (11 per 1000). The risk of delirium was about 30% higher (95% CI 15%–47%) among patients taking statins before surgery (14 per 1000) than among those not taking statins before surgery (11 per 1000, *p* < 0.001). After adjusting for demographic characteristics, prior admissions and prescriptions and classes of neuropsychiatric medications, we observed a continued increase (OR 1.48, 95% CI 1.38–1.68) in risk of delirium among patients prescribed statins than among those not prescribed statins. We observed a similar increase in risk after adjusting for each cardiovascular and noncardiovascular medication (Figure 1). The final multivariable model, which adjusted for all the preceding predictors along with the type and duration of surgery, showed a 28% increase in the risk of delirium associated with the use of statins (95% CI 12%–46%, *p* < 0.001).

![Figure1](http://www.cmaj.ca/https://www.cmaj.ca/content/cmaj/179/7/645/F1.medium.gif)

[Figure1](http://www.cmaj.ca/content/179/7/645/F1)

**Figure 1:** Risk of postoperative delirium with use of statins among elderly patients. *Derived from the full multivariable model after adjustment for age, sex, social status, prior admissions, prior use of medications, each neuropsychiatric, cardiac, vascular and miscellaneous medication, and duration and type of surgery. Note: CI = confidence interval.

The increased risk of delirium associated with the use of statins was evident across a variety of clinical settings. In each analysis, the results showed a detrimental association, although the confidence intervals were broad in many subgroups (Figure 2). Increases in risk were highly consistent among patients who did not take selected neuropsychiatric medications (that were associated with delirium) or selected common medications (that were not associated with delirium). The relative risk with statins was somewhat higher among patients who had noncardiac surgery than among patients who had cardiac surgery, although the baseline risk among those who had cardiac surgery was twice as high as that among patients undergoing noncardiac surgery. The absolute risk of delirium associated with the use of statins was highest among patients older than 70 years and among patients whose surgeries lasted longer than 3 hours (Appendix 1, available online at [www.cmaj.ca/cgi/content/full/179/7/645/DC2](http://www.cmaj.ca/cgi/content/full/179/7/645/DC2)).

![Figure2](http://www.cmaj.ca/https://www.cmaj.ca/content/cmaj/179/7/645/F2.medium.gif)

[Figure2](http://www.cmaj.ca/content/179/7/645/F2)

**Figure 2:** Relative risk of postoperative delirium associated with the use of statins among elderly patients, by type of medication and surgery. *Adjusted for age, sex, duration of surgery and individual medications. †Excludes patients taking corresponding medication. ‡Further adjusted for type of surgery. §Includes thoracic, neurosurgical, vascular and musculoskeletal procedures. ¶Includes abdominal, retroperitoneal and unclassified procedures. **Includes lower urogenital, breast and skin, external head and neck, and ophthalmological procedures. Note: CI = confidence interval.

The overall resolution of the multivariable model was reasonable (c-index 0.78, *p* < 0.001). The strongest single predictor of postoperative delirium was the duration of surgery, which yielded about a 44% increase in risk for delirium for each hour of surgery (95% CI 41%–46%). The other statistically significant predictors were 2 demographic characteristics (age and sex), use of the 4 neuropsychiatric medications and type of surgery (Table 2). In contrast, use of angiotensin-converting-enzyme inhibitors, calcium-channel blockers or gastric acid suppressors was not associated with delirium. None of the other long-term cardiovascular and noncardiovascular medications were associated with a statistically significant increase in risk of delirium (Table 3).

View this table:
[Table2](http://www.cmaj.ca/content/179/7/645/T2)

Table 2.


View this table:
[Table3](http://www.cmaj.ca/content/179/7/645/T3)

Table 3.

After comparing patients using different statins and receiving different doses, we found no major exceptions. Atorvastatin and simvastatin (2 popular statins with distinct metabolic pathways and half-lives) were each associated with a significantly increased risk of delirium. The odds ratios were 1.68 (95% CI 1.34–2.09) for atorvastatin and 1.46 (95% CI 1.15–1.84) for simvastatin. Pravastatin (the main nonlipophilic statin in our study) was associated with a marginally lower risk of delirium (OR 1.26, 95% CI 0.96–1.64). Patients prescribed the lowest dose of each statin (20 mg) showed a marginally smaller increase in risk of delirium (OR 1.15, 95% CI 0.94–1.40) than patients prescribed higher doses (OR 1.38, 95% CI 1.17–1.63).

We analyzed patients excluded because of pre-existing major vascular disease (*n* = 69 094) and found about the same increase in delirium risk among those taking statins compared with those not taking statins (OR 1.26, 95% CI 1.10–1.45). We analyzed patients admitted for emergency surgery (*n* = 222 615) and found an increase in the risk of delirium among those taking statins compared with those not taking statins (OR 1.24, 95% CI 1.11–1.39). We analyzed patients excluded because the duration of anesthesiology was unavailable (*n* = 100 832), and also found an increase in delirium risk associated with use of statins (OR 1.18, 95% CI 0.92–1.53). In contrast, when we analyzed controls who had received a statin in the past but not in the 90 days before surgery (*n* = 43 152), we found no increase in the risk of delirium associated with use of statins (OR 0.85, 95% CI 0.76–0.96). Finally, in the cohort matched by propensity scores (*n* = 25 244), we found about the same increase in delirium risk associated with the use of statins (OR 1.30, 95% CI 0.96–1.70).

We explored 7 combinations of severity to test whether delirium contributed to other postoperative complications and whether delirium associated with the use of statins was any less severe than delirium that was not associated with the use of statins. For example, patients who experienced delirium were 6 times more likely than those who did not experience delirium to receive a computed tomography scan while in hospital (15.7% v. 2.4%, *p* < 0.001). Patients with delirium were 3 times more likely than patients who did not experience delirium to have a wound infection develop while in hospital (17.0% v. 5.5%, *p* < 0.001) and 4 times more likely to experience a myocardial infarction while in hospital (3.1% v. 0.7%, *p* < 0.001). Finally, patients who experienced delirium were 3 times more likely than those who did not experience delirium to die in hospital (3.5% v. 1.0%, *p* < 0.001). The use of statins was associated with an increased risk of all these severe forms of delirium except the combination of delirium followed by death (Table 4). Overall, we observed about a 10-day absolute increase in the length of stay in hospital for all patients who experienced delirium (16.1 days among patients who were taking statins v. 6.3 days among those who were not taking statins, *p* < 0.001). We also observed about a 10% absolute increase in ongoing use of sedatives after discharge (33.9% among patients taking statins v. 23.3% among patients not taking statins, *p* < 0.001).

View this table:
[Table4](http://www.cmaj.ca/content/179/7/645/T4)

Table 4.

## Interpretation

We found that 1 in 14 elderly patients in our study cohort were taking statins before undergoing elective surgery and about 1 in 90 experienced delirium after elective surgery. Our results suggested that this association was more than a coincidence, particularly among patients who received higher doses of statins and had longer duration noncardiac surgeries. The association between statins and risk of delirium was distinct and was not observed with other lipid-lowering medications, cardiovascular medications or common drugs that reflect underlying chronic diseases but have no major effects on the cardiovascular system. The correlation prevailed among patients in high-risk subgroups and extended to some more complicated combinations of delirium. The magnitude of the association was substantial, but not as substantial as that between risk of delirium and advanced age, baseline neuropsychiatric drug treatment or prolonged surgery.

The most important limitation of our research was that hidden confounding may have biased our results. For example, we were not able to measure intraoperative hypotension or hypoxemia. However, these factors have not been shown to predict postoperative delirium.24 We were able to measure patient age and duration of surgery, the 2 most established predictors of delirium in this setting. Our multivariable models also adjusted for many other features that could remain imbalanced in small trials. Our analysis of selection bias suggested that patients taking statins were healthier than patients who were not taking statins. In addition, prior research indicated that unmeasured comorbidities tended to be biased in a manner favouring patients taking statins.25,26 Hidden confounding would not explain why our results showed the association between postoperative delirium and the use of statins to be distinct, or why we found no association between delirium and the use of other medications.

Another important limitation of our analysis was outcome definition. That is, the database analyses provided a specific, but insensitive, indicator for postoperative delirium. As such, our study detected extreme cases and underestimated the overall incidence of postoperative delirium, particularly in our analyses that focused on severe combinations of delirium. Consequently, all estimates were biased conservatively, and the absolute increase in risk associated with statin use was understated (assuming no systematic coding bias). If the true incidence of delirium were 10%,27 for example, an observed odds ratio of 1.28 might amount to a number needed to harm of about 35. The true incidence may be impossible to determine with certainty given the vagaries of distinguishing delirium from other forms of postoperative cognitive dysfunction.28

Shifts in the distribution of capillary blood flow leading to neuronal ischemia scattered throughout the brain could explain postoperative delirium. However, results of our goodness-of-fit test suggested that other mechanisms also contributed. These mechanisms merit research because delirium could be associated with many of them, including altered isoprenoid metabolism in neurons, inhibition of membrane receptors, toxicity at synaptic clefts, blood–brain barrier dysfunction or drug interactions if such statin pleiotropic effects are tolerated under normal circumstances but not during anesthesia.29,30 Our findings are not easily explained as drug withdrawal, because statin cessation generally led to a decreased effect (not rebound harm) and because the observed risk was no lower for the statin with the longest half-life (atorvastatin).31 The endothelial nitric oxide synthase mechanism emphasized the role of microvascular blood flow autoregulation for preserving neuronal function.32

Our results are not readily attributable to unmeasured underlying severity of vascular disease for several reasons. First, we observed no adverse association with 20 other medications, each a marker of vascular disease. Second, the analyses adjusted for 30 measured factors and yielded stable results. Hence, a postulated underlying disease would need to be unrelated to all these factors to extinguish the finding. Third, the adverse association was no more or less substantial in our post-hoc analysis of patients who had major vascular disease. Fourth, no risk was observed in earlier studies correlating statin use with postoperative myocardial infarction. Hence, the adverse association between the use of statins and postoperative delirium is distinct. Conducting a randomized trial would be a different way to check for bias due to severity of disease. However, randomized trials are generally an unreliable method for studying adverse drug effects, whereas nonrandomized studies are a powerful method for identifying potential toxicity.33

We caution that our findings do not imply that statins are harmful under normal circumstances. On the contrary, trials involving outpatients show many benefits from long-term statin therapy in cardiovascular care. Furthermore, our primary analysis was restricted to patients who did not have major vascular disease; therefore, we did not assess their potential benefits in preventing myocardial infarction.34

The implication of our study is that statins, unlike other cardiac medications, might contribute to delirium after elective surgery and can be discontinued temporarily before surgery. Studies suggest that expression of endothelial nitric oxide synthase returns to normal within 2–4 days after interrupting statin therapy.35 If needed, statin therapy can be reinstated on the first postoperative day, which would re-establish activity of endothelial nitric oxide synthase within 1–2 days (the interval most common for a postoperative myocardial infarction).36 Such a strategy would have the additional advantage of reducing the risk of inadvertent drug interactions during the hospital stay and reducing the risk of postoperative hepatitis.

Further research could be designed as either a discontinuation trial or a prophylaxis trial, depending on the feasibility of recruitment. We applaud recent calls for more clinical trials in perioperative medicine to directly determine whether medications that are beneficial in the outpatient setting are equally safe in the perioperative setting.37 The results of our study should inform the design of future studies, bolster recruitment in trials that are currently underway and underscore the need to carefully assess both cardiac and adverse neurologic outcomes after surgery. Until more data are available, a recommendation to temporarily interrupt the use of statins before surgery may be a reasonable compromise.

@@ See related commentary by Marcantonio, page [627](http://www.cmaj.ca/lookup/volpage/179/627?iss=7)

## Footnotes

*   Une version française de ce résumé est disponible à l'adresse [www.cmaj.ca/cgi/content/full/179/7/645/DC1](http://www.cmaj.ca/cgi/content/full/179/7/645/DC1)
    
    This article has been peer reviewed.
    
    **Contributors:** All of the authors participated in the study design, interpretation of results, the drafting of the manuscript and approval of the final version for publication. Donald Redelmeier had full access to all the data, final authority for the decision to submit for publication and ongoing responsibility for accuracy.
    
    **Acknowledgements:** We thank Marianne Alexiadis, Edward Etchells, David Juurlink, Daniel Hackam, Alexander Kopp, Beverley Orser, Damon Scales, Michael Schull, Steven Shumak, Homer Tien, Joel Ray, Sunu Thomas, Charlie Wang and Brian Wong for their helpful comments on earlier drafts of the manuscript.
    
    The views expressed are not necessarily those of the Ontario Ministry of Health and Long-Term Care.
    
    **Funding:** This work was supported by the Canada Research Chair in Medical Decision Sciences, the Canadian Institutes of Health Research, the National Institutes of Health Resuscitation Outcomes Consortium, the Clinician Scientist Training Program of the University of Toronto and the Physician's Service Incorporated Foundation of Ontario.
    
    **Competing interests:** None declared.

## REFERENCES

1.  1. Ely EW, Shintani A, Truman B, et al. Delirium as a predictor of mortality in mechanically ventilated patients in the intensive care unit. JAMA 2004;291:1753-62.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1001/jama.291.14.1753&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=15082703&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000220788700033&link_type=ISI) 

2.  2. Thomason JW, Shintani A, Peterson JF, et al. Intensive care unit delirium is an independent predictor of longer hospital stay: a prospective analysis of 261 nonventilated patients. Crit Care 2005;9:R375-81.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1186/cc3729&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=16137350&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000232151500013&link_type=ISI) 

3.  3. Cole MG. Delirium in elderly patients. Am J Geriatr Psychiatry 2004;12:7-21.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1176/appi.ajgp.12.1.7&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=14729554&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000188001000002&link_type=ISI) 

4.  4. Roche V. Etiology and management of delirium. Am J Med Sci 2003;325:20-30.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1097/00000441-200301000-00005&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=12544081&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000180459300005&link_type=ISI) 

5.  5. Moller JT, Cluitmans P, Rasmussen LS, et al. Long-term postoperative cognitive dysfunction in the elderly ISPOCD1 study. ISPOCD investigators. International Study of Post-Operative Cognitive Dysfunction. Lancet 1998;351:857-61.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1016/S0140-6736(97)07382-0&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=9525362&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000072645500011&link_type=ISI) 

6.  6. Inouye SK, Charpentier PA. Precipitating factors for delirium in hospitalized elderly persons: predictive model and interrelationship with baseline vulnerability. JAMA 1996;275:852-7.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1001/jama.1996.03530350034031&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=8596223&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=A1996TZ97100028&link_type=ISI) 

7.  7. Yokota H, Ogawa S, Kurokawa A, et al. Regional cerebral blood flow in delirium patients. Psychiatry Clin Neurosci 2003;57:337-9.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1046/j.1440-1819.2003.01126.x&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=12753576&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 

8.  8. Kalisvaart KJ, de Jonghe JF, Bogaards MJ. Haloperidol prophylaxis for elderly hip-surgery patients at risk for delirium: a randomized placebo controlled study. J Am Geriatr Soc 2005;53:1658-66.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1111/j.1532-5415.2005.53503.x&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=16181163&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000232006000002&link_type=ISI) 

9.  9. Liptzin B, Laki A, Garb JL, et al. Donepezil in the prevention and treatment of post-surgical delirium. Am J Geriatr Psychiatry 2005;13:1100-6.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1176/appi.ajgp.13.12.1100&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=16319303&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000233614500010&link_type=ISI) 

10. 10. Leung JM, Sands LP, Vaurio LE, et al. Nitrous oxide does not change the incidence of postoperative delirium or cognitive decline in elderly surgical patients. Br J Anaesth 2006;96:754-60.
    
    [Abstract/FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NjoiYnJqYW5hIjtzOjU6InJlc2lkIjtzOjg6Ijk2LzYvNzU0IjtzOjQ6ImF0b20iO3M6MjA6Ii9jbWFqLzE3OS83LzY0NS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

11. 11. McGown CC, Brookes ZL. Beneficial effects of statins on the microcirculation during sepsis: the role of nitric oxide. Br J Anaesth 2007;98:163-75.
    
    [Abstract/FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NjoiYnJqYW5hIjtzOjU6InJlc2lkIjtzOjg6Ijk4LzIvMTYzIjtzOjQ6ImF0b20iO3M6MjA6Ii9jbWFqLzE3OS83LzY0NS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

12. 12. Buerk DG, Atochin DN, Riva CE. Investigating the role of nitric oxide in regulating blood flow and oxygen delivery from in vivo electrochemical measurements in eye and brain. Adv Exp Med Biol 2003;530:359-70.
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=14562731&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000186182800033&link_type=ISI) 

13. 13. Chen J, Zacharek A, Zhang C, et al. Endothelial nitric oxide synthase regulates brain-derived neurotrophic factor expression and neurogenesis after stroke in mice. J Neurosci 2005;25:2366-75.
    
    [Abstract/FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Njoiam5ldXJvIjtzOjU6InJlc2lkIjtzOjk6IjI1LzkvMjM2NiI7czo0OiJhdG9tIjtzOjIwOiIvY21hai8xNzkvNy82NDUuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 

14. 14. Fong TG, Bogardus ST Jr, Daftary A. Cerebral perfusion changes in older delirious patients using 99mTc HMPAO SPECT. J Gerontol A Biol Sci Med Sci 2006;61:1294-9.
    
    [Abstract/FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NjoiZ2Vyb25hIjtzOjU6InJlc2lkIjtzOjEwOiI2MS8xMi8xMjk0IjtzOjQ6ImF0b20iO3M6MjA6Ii9jbWFqLzE3OS83LzY0NS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

15. 15. Redelmeier D. New thinking on postoperative delirium. CMAJ 2007;177:424.
    
    [FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiRlVMTCI7czoxMToiam91cm5hbENvZGUiO3M6NDoiY21haiI7czo1OiJyZXNpZCI7czo5OiIxNzcvNC80MjQiO3M6NDoiYXRvbSI7czoyMDoiL2NtYWovMTc5LzcvNjQ1LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

16. 16. Kapoor AS, Kanji H, Buckingham J, et al. Strength of evidence for perioperative use of statins to reduce cardiovascular risk: systematic review of controlled studies. BMJ 2006;333:1149.
    
    [Abstract/FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiYm1qIjtzOjU6InJlc2lkIjtzOjEzOiIzMzMvNzU3OS8xMTQ5IjtzOjQ6ImF0b20iO3M6MjA6Ii9jbWFqLzE3OS83LzY0NS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

17. 17. Juurlink DN, Mamdani MM, Lee DS, et al. Rates of hyperkalemia after publication of the Randomized Aldactone Evaluation Study. N Engl J Med 2004;351:543-51.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1056/NEJMoa040135&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=15295047&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000223069900007&link_type=ISI) 

18. 18. Tu JV, Penfold SP, McColgan et al, eds. *Access to health services in Ontario: ICES Atlas*. Toronto: Institute for Clinical Evaluative Sciences; 2005.
    
    

19. 19. Horwitz RI, McFarlane MJ, Brennan TA, et al. The role of susceptibility bias in epidemiologic research. Arch Intern Med 1985;145:909-12.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1001/archinte.1985.00360050177030&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=3994467&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=A1985AGN4100025&link_type=ISI) 

20. 20. Levy AR, O'Brien BJ, Sellors C, et al. Coding accuracy of administrative drug claims in the Ontario Drug Benefit database. Can J Clin Pharmacol 2003;10:67-71.
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=12879144&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 

21. 21. Redelmeier DA, Thiruchelvam D, Daneman N. Introducing a methodology for estimating duration of surgery in health services research. J Clin Epidemiol 2008;61:882-9.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1016/j.jclinepi.2007.10.015&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=18468859&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000258557100005&link_type=ISI) 

22. 22. Romano PS, Chan BK, Schembri ME, et al. Can administrative data be used to compare postoperative complication rates across hospitals? Med Care 2002;40:856-67.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1097/00005650-200210000-00004&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=12395020&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000178435400004&link_type=ISI) 

23. 23. Daneman N. *The epidemiology of surgical site infections* [master's thesis]. Toronto (ON): University of Toronto; 2008.
    
    

24. 24. Marcantonio ER, Goldman L, Orav EJ, et al. The association of intraoperative factors with the development of postoperative delirium. Am J Med 1998;105:380-4.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1016/S0002-9343(98)00292-7&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=9831421&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000077055200004&link_type=ISI) 

25. 25. Redelmeier DA, Tan SH, Booth GL. The treatment of unrelated disorders in patients with chronic medical diseases. N Engl J Med 1998;338:1516-20.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1056/NEJM199805213382106&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=9593791&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000073699100006&link_type=ISI) 

26. 26. Brookhart MA, Patrick AR, Dormuth C, et al. Adherence to lipid-lowering therapy and the use of preventive health services: an investigation of the healthy user effect. Am J Epidemiol 2007;166:348-54.
    
    [Abstract/FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NzoiYW1qZXBpZCI7czo1OiJyZXNpZCI7czo5OiIxNjYvMy8zNDgiO3M6NDoiYXRvbSI7czoyMDoiL2NtYWovMTc5LzcvNjQ1LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

27. 27. Siddiqi N, House AO, Holmes JD. Occurrence and outcome of delirium in medical in-patients: a systematic literature review. Age Ageing 2006;35:350-64.
    
    [Abstract/FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NjoiYWdlaW5nIjtzOjU6InJlc2lkIjtzOjg6IjM1LzQvMzUwIjtzOjQ6ImF0b20iO3M6MjA6Ii9jbWFqLzE3OS83LzY0NS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

28. 28. Newman S, Stygall J, Hirani S, et al. Postoperative cognitive dysfunction after noncardiac surgery: a systematic review. Anesthesiology 2007;106:572-90.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1097/00000542-200703000-00023&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=17325517&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000244617400022&link_type=ISI) 

29. 29. Hering H, Lin CC, Sheng M. Lipid rafts in the maintenance of synapses, dendritic spines, and surface AMPA receptor stability. J Neurosci 2003;23:3262-71.
    
    [Abstract/FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Njoiam5ldXJvIjtzOjU6InJlc2lkIjtzOjk6IjIzLzgvMzI2MiI7czo0OiJhdG9tIjtzOjIwOiIvY21hai8xNzkvNy82NDUuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 

30. 30. Stang P, Morris L, Kempf J, et al. The coprescription of contraindicated drugs with statins: continuing potential for increased risk of adverse events. Am J Ther 2007;14:30-40.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1097/01.mjt.0000208875.61480.c8&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=17303973&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 

31. 31. Cubeddu LX, Seamon MJ. Statin withdrawal: clinical implications and molecular mechanisms. Pharmacotherapy 2006;26:1288-96.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1592/phco.26.9.1288&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=16945051&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000240281100010&link_type=ISI) 

32. 32. Han HS, Suk K. The function and integrity of the neurovascular unit rests upon the integration of the vascular and inflammatory cell systems. Curr Neurovasc Res 2005;2:409-23.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.2174/156720205774962647&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=16375722&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000234340000006&link_type=ISI) 

33. 33. Papanikolaou PN, Christidi GD, Ioannidis JP. Comparison of evidence on harms of medical interventions in randomized and nonrandomized studies. CMAJ 2006;174:635-41.
    
    [Abstract/FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoiY21haiI7czo1OiJyZXNpZCI7czo5OiIxNzQvNS82MzUiO3M6NDoiYXRvbSI7czoyMDoiL2NtYWovMTc5LzcvNjQ1LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

34. 34. Lindenauer PK, Pekow P, Wang K, et al. Lipid-lowering therapy and in-hospital mortality following major noncardiac surgery. JAMA 2004;291:2092-9.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1001/jama.291.17.2092&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=15126437&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000221174300026&link_type=ISI) 

35. 35. Gertz K, Laufs U, Lindauer U, et al. Withdrawal of statin treatment abrogates stroke protection in mice. Stroke 2003;34:551-7.
    
    [Abstract/FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6OToic3Ryb2tlYWhhIjtzOjU6InJlc2lkIjtzOjg6IjM0LzIvNTUxIjtzOjQ6ImF0b20iO3M6MjA6Ii9jbWFqLzE3OS83LzY0NS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

36. 36. Kapoor AS, Kanji H, Buckingham J, et al. Strength of evidence for perioperative use of statins to reduce cardiovascular risk: systematic review of controlled studies. BMJ 2006;333:1149.
    
    [Abstract/FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiYm1qIjtzOjU6InJlc2lkIjtzOjEzOiIzMzMvNzU3OS8xMTQ5IjtzOjQ6ImF0b20iO3M6MjA6Ii9jbWFqLzE3OS83LzY0NS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

37. 37. Fleisher LA, Poldermans D. Perioperative beta blockade: Where do we go from here? Lancet 2008;371:1813-4.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1016/S0140-6736(08)60662-5&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=18479743&link_type=MED&atom=%2Fcmaj%2F179%2F7%2F645.atom)