| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Departments of 1 Geriatrics and Adult Development,
3 Orthopaedics, and
5 Medicine, Mount Sinai School of Medicine, New York, New York.
2 Department of Epidemiology and Preventive Medicine, University of Maryland School of Medicine, Baltimore.
4 Orthopaedic Institute, New York University School of Medicine, New York.
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
Methods. We conducted a prospective cohort study at four New York hospitals that enrolled 541 patients with hip fracture and without delirium. Delirium was identified prospectively by patient interview supplemented by medical record review. Multiple logistic regression was used to identify risk factors.
Results. Eighty-seven of 541 patients (16%) became delirious. Among all subjects, risk factors for delirium were cognitive impairment (relative risk, or RR, 3.6; 95% confidence interval, or CI, 1.8–7.2), abnormal blood pressure (RR 2.3, 95% CI 1.2–4.7), and heart failure (RR 2.9, 95% CI 1.6–5.3). Patients who received less than 10 mg of parenteral morphine sulfate equivalents per day were more likely to develop delirium than patients who received more analgesia (RR 5.4, 95% CI 2.4–12.3). Patients who received meperidine were at increased risk of developing delirium as compared with patients who received other opioid analgesics (RR 2.4, 95% CI 1.3–4.5). In cognitively intact patients, severe pain significantly increased the risk of delirium (RR 9.0, 95% CI 1.8–45.2).
Conclusions. Using admission data, clinicians can identify patients at high risk for delirium following hip fracture. Avoiding opioids or using very low doses of opioids increased the risk of delirium. Cognitively intact patients with undertreated pain were nine times more likely to develop delirium than patients whose pain was adequately treated. Undertreated pain and inadequate analgesia appear to be risk factors for delirium in frail older adults.
Delirium is the most frequent medical complication observed in the 350,000 Americans hospitalized annually with hip fracture (1). The prevalence of delirium following hip fracture ranges from 13% to 61% (2–5), and delirium has been associated with delayed recovery, increased mortality, and poorer physical, cognitive, and affective function 6 months postfracture (2–4,6).
Although risk factors for the development of delirium have been well described in the medically ill and patients undergoing elective noncardiac surgery (7–9), extrapolating the results of these studies and applying previously developed prediction rules to hip fracture patients is problematic. Hip fracture patients are typically older, have poorer baseline functional status, more comorbid medical conditions, and a higher prevalence of dementia than patients previously studied. Indeed, recent data suggest that delirium following hip fracture results from different causes and follows a different clinical course than other delirium syndromes (10). Hip fracture also is associated with considerable pain (11), and undertreated pain has been demonstrated to be an independent risk factor for delirium in healthy older adults undergoing elective surgery (12,13).
This study was performed to examine the relationship among pain, opioid analgesic prescribing, and other risk factors on the development of delirium in frail older adults with hip fracture.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Diagnosis and Detection of Delirium
Within 48 hours of admission, a research nurse interviewed all subjects by using the Confusion Assessment Method (CAM)—a standardized algorithm for the detection of delirium (14). Subjects who were found to be delirious by the CAM at the initial interview were excluded. Subjects were assessed with the CAM on a daily basis, 5 days a week (Monday–Friday) until discharge. Additionally, the research assistant reviewed the medical records daily and spoke to hospital staff to supplement the CAM observations. The medical record was reviewed for the key words "delirious/delirium," "agitated/agitation," "changed/altered mental status," or "new/increased/more confused/confusion" for patients without a diagnosis of dementia and for the key words "delirious/delirium," "changed/altered mental status," or "new confusion" for patients with a history of dementia. These criteria were used because the fluctuating nature of delirium limits the sensitivity of once-daily interviews and because subjects were not interviewed on weekends. Delirium was considered to be present if either the CAM or chart criteria were met. The use of these chart criteria in combination with the CAM to diagnose delirium has been previously validated (7). Patients were not assessed postoperatively on the day of surgery because of the difficulty of distinguishing true delirium from residual effects of anesthesia (7). Although subjects were not interviewed on weekends, patients who developed delirium over the weekend would not have been expected to clear their delirium by Monday, given the reported duration of delirium (3).
Collection of Delirium Risk Factors
Risk factors were collected from medical records, patient or proxy interviews, nursing staff interviews, and patient observation. Selection of risk factors was based on previous studies of delirium (2,7–9,13) and on clinical relevance and experience. Risk factors were grouped into nine variable sets: patient demographic or background characteristics, cognitive status, functional status, biomedical factors, abnormal laboratory or clinical findings on admission, medical complications, nonoperative process variables, operative process variables, and pain-related variables (Table 1). For the cognitive status variable, patients were classified as cognitively intact if they had no diagnosis of memory impairment or a dementing illness and were able to give correct answers to a four-item screen that contained questions about their orientation (place and time), circumstances of their fracture (place, time, and circumstances of their fracture), immediate recall of the nature and purpose of the research study, and recall of the name or position of the person administering the informed consent. Patients were classified as cognitively impaired if they had a diagnosis or history of memory impairment or a dementing illness or if they made one or more errors in answering the four-item screen.
|
Nonoperative process variables included time spent in the emergency department, time spent without oral intake (NPO), and sedative or hypnotic medication use. Days spent NPO were counted as a risk factor if they occurred before the episode of delirium or through postoperative Day 3 for nondelirious subjects or hospital Day 5 for nondelirious subjects who received nonoperative management. The operative process variables included surgery more than 24 hours after admission, surgical repair technique, type of anesthesia, and duration of anesthesia.
We collected data on pain and opioid analgesic intake in morphine sulfate equivalents through postoperative Day 3.
Patients were interviewed on a daily basis and asked to report the average amount of pain that they experienced at rest over the previous 24-hour period on a 1 (no pain) to 5 (very severe) scale. As patients were not interviewed over the weekend, patients interviewed on Monday were asked to rate their average pain at rest over the preceding 48 hours on the same 1–5 scale. Pain was included as a risk factor if the subject had a pain at rest score of 4 or 5 (severe to very severe pain) within 48 hours before the episode of delirium or through postoperative Day 3 for nondelirious subjects.
We calculated the total daily opioid dose in parenteral morphine sulfate equivalents for the 24 hours preceding the delirious episode and the highest 24-hour cumulative opioid dose for the first 3 postoperative days for nondelirious subjects. Doses of all opioids (including continuous infusions and patient-controlled analgesia) were converted into parenteral morphine sulfate equivalents by using the AHCPR equianalgesic dosing guidelines (18). As meperidine has been associated with an increased risk of delirium in previous studies, we also examined whether the use of meperidine increased the risk of delirium (19). Meperidine was considered a risk factor if it was administered within 24 hours of a delirious episode or at any time during postoperative Days 1–3 for nondelirious subjects. Finally, we created a dichotomous variable to indicate whether opioid administration increased immediately following an episode of severe pain prior to a delirious episode or at any time during the first 3 postoperative days for nondelirious subjects. This variable was included in an effort to determine whether it was pain or an increase in the analgesic dose in response to pain that was associated with delirium.
Statistical Analysis
Risk factors known prior to surgery.--
Logistic regression was used to identify risk factors for delirium. The first model examined the effects of risk factors that are known to the medical team at hospital admission. All variables with p <.15 in univariate comparisons were entered into the model. However, if a variable set from Table 1 (e.g., biomedical factors) did not contain a variable that was significantly associated with delirium, one or more representative variables from the missing set(s) were entered on the basis of clinical relevance. The amount of missing data from patients with a diagnosis of dementia who could not self-report their pain (over 30% of subjects) precluded the pain variable being entered into the overall model.
Risk factors known at the time of surgery.-- A second regression model examined the association of risk factors on postoperative delirium. The postoperative model was created by entering the operative and postoperative process variables that were associated with the development of delirium following surgery at a significance level of p <.15 into the regression equation, along with the variables previously included in the postadmission delirium model just described.
Delirium in cognitively intact patients and cognitively impaired and dementia patients.-- As patients with cognitive impairment were at increased risk of developing delirium as compared with cognitively intact patients, and because pain was not reliably reported in cognitively impaired patients, we performed separate analyses examining risk factors for the development of delirium in (a) cognitively intact and (b) cognitively impaired patients. Model development proceeded as just described. We included the pain self-report variables in models for the cognitively intact patients.
Goodness of fit was assessed for all models by using the area under the receiver operating characteristic (ROC) curve and the Hosmer–Lemeshow statistic (20). Analyses were performed by using STATA (College Station, TX).
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Risk Factors for Delirium for All Patients
Table 1 lists univariate associations for the possible risk factors for delirium following hospital admission. Table 2 lists the multiple logistic regression analysis results. Independent predictors of new delirium following hospital admission were as follows: evidence of cognitive impairment (relative risk, or RR, 3.6; 95% confidence interval, or CI, 1.8–7.2), abnormal blood pressure on admission (RR 2.3, 95% CI 1.2–4.7), heart failure on admission (RR 2.9, 95% CI 1.6–5.3), meperidine use (RR 2.4, 95% CI 1.3–4.5), and receiving very low doses of opioid analgesia (RR 5.4, 95% CI 2.4–12.3) for patients receiving <10 mg of parenteral morphine sulfate equivalents per day and RR 1.4, 95% CI 0.6–3.3 for patients receiving 10–30 mg of parenteral morphine sulfate equivalents per day compared with patients receiving <30 mg of morphine sulfate equivalents). Among all subjects, independent predictors of postoperative delirium were identical to those for postadmission delirium, and no operative process variables were independently associated with delirium. Too few patients received benzodiazepines or other sedatives or hypnotics for meaningful analyses to be performed. The area under the ROC curve for the post-admission delirium model was 0.80.
|
|
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Data from this study suggest that undertreated pain is a significant contributor to the development of delirium. Furthermore, our data demonstrate that opioids, with the exception of meperidine, do not precipitate delirium in patients with acute pain and that avoiding opioids or administering very low doses of opioids is associated with an increased risk of delirium. A major barrier to the treatment of pain in older adults has been the fear that opioids cause delirium. In the subgroup of cognitively intact adults in this study, severe pain was associated with a ninefold risk of developing delirium. Receiving no opioid analgesia or a very low dose of an opioid increased the risk of developing delirium for both cognitively intact and cognitively impaired patients. Our data also confirm previous findings that meperidine is associated with an increased risk of delirium (19)—perhaps because of the action of its neuroexcitatory metabolite normeperidine—and thus should not be administered to geriatric patients. Doses of meperidine were significantly higher than those of other opioids, providing evidence that the association of delirium with meperidine was not the result of meperidine being administered at lower equivalent doses as compared with other analgesics.
Although it is possible that clinicians may have identified patients at increased risk of delirium and thus altered their analgesic prescribing practices (i.e, administering lower doses of analgesia to patients at high risk of delirium), we believe that this is unlikely. We found no evidence that doses of opioids were reduced in the immediate 48 hours preceding an episode of delirium, and we did not observe significant differences in opioid prescribing between patients with and without other delirium risk factors. Studies in relatively healthy older adults undergoing elective surgery have revealed similar findings with respect to the relationship between pain and delirium, further supporting our data (12,13).
Although this study cannot determine whether improved pain management will reduce the incidence and duration of delirium, it provides evidence to support future studies of intensive pain management in older adults. Two studies employing geriatric consultation, targeting known delirium risk factors, and providing recommendations for pain management have recently demonstrated reductions in the incidence (22) and duration of delirium (23) following hip fracture. However, the studies could not distinguish the independent effect of enhanced analgesia on the prevention of delirium, and compliance with the analgesic recommendations was poor.
This study has limitations. We observed patients for 5 days of every week the patient was hospitalized, and subjects were observed once per day; as a result, some episodes of delirium might have been missed. Nevertheless, we supplemented our observations with chart notations that were previously shown to be relatively sensitive and with specific markers for delirium (7), and, given the duration of delirium observed in this cohort and other studies (10,24), it is unlikely that we missed a substantial number of cases. We were unable to include pain as a variable in our overall or subanalysis of cognitively impaired patients because of missing data in dementia patients unable to self-report their pain. Thus, we do not know whether untreated pain is a risk factor for delirium in the cognitively impaired.
The findings from this study both extend previous research on delirium and provide new insights into this complicated medical syndrome in hip fracture patients. This study suggests that hip fracture patients at increased risk for delirium can be identified at admission on the basis of a limited number of variables and that untreated pain and underuse of opioid analgesics are important contributors to its development. Our data will help guide studies of interventions aimed at improving pain management and preventing the development of delirium in high-risk patients hospitalized with hip fracture.
|
|
Acknowledgments |
|---|
We gratefully acknowledge the invaluable assistance of our project manager, Stacey Silberzweig, and of our research assistants, Camille Cohen, Linda Dutcher, and Sameh Samy.
Address correspondence to R. Sean Morrison, MD, Department of Geriatrics, Box 1070, The Mount Sinai School of Medicine, One Gustave L. Levy Place, New York, NY 10029. E-mail: sean.morrison{at}mssm.edu
Received July 23, 2002
Accepted August 8, 2002
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. S. Kornitzer, L. C. Manace, D. J. Fischberg, and R. M. Leipzig Prevalence of Meperidine Use in Older Surgical Patients Arch Surg, January 1, 2006; 141(1): 76 - 81. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
