Clinical Gastroenterology and Hepatology 2015;13:1670–1675
Psychoactive Medications Increase the Risk of Falls and Fall-related Injuries in Hospitalized Patients With Cirrhosis Elliot B. Tapper, Yesenia Risech-Neyman, and Neil Sengupta Department of Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts BACKGROUND & AIMS:
Reducing or eliminating falls is a focus of patient safety programs as well as health policy. Falls are tied to hospital reimbursement. However, little is known about the risk of falls among hospitalized patients with cirrhosis or factors that affect risk of falling.
METHODS:
We conducted a retrospective cohort study of inpatients with cirrhosis from 2010 to 2013 at a liver transplant center. Our primary aim was to determine the clinical factors associated with falls and fall-related injuries for patients with cirrhosis. Our secondary aim was to describe the rate ratio of falls and fall-related injuries among patients with cirrhosis compared with general medical inpatients.
RESULTS:
During the study period, there were 1749 admissions to the liver service; 55 (3.1%) resulted in falls. Patients who fell were more likely to have received benzodiazepines (50.9% vs 16.7%, P < .0001) and antipsychotic agents (30.9% vs 7.3%, P < .0001). After adjusting for hepatic encephalopathy, the respective odds of a fall after benzodiazepine or antipsychotic exposure were 6.59 (95% confidence interval [CI], 3.76L11.59) and 3.72 (95% CI, 1.90L7.06). The adjusted risk of a fall-related injury was also significantly associated with benzodiazepine and antipsychotic agents, with respective odds ratios of 3.45 (95% CI, 1.39L8.23) and 3.42 (95% CI, 1.09L8.99). Fall-related injuries occurred at a rate of 1.70/1000 patient-days for patients with cirrhosis vs 0.5/1000 patient-days for patients in the general medical service. Accordingly, the rate ratio for a fall-related injury among patients with cirrhosis was 3.37 (95% CI, 1.99L5.72; P < .0001).
CONCLUSIONS:
Psychoactive medications are associated with an increased adjusted risk of falls and fall-related injuries in hospitalized patients with cirrhosis.
Keywords: Model for End-Stage Liver Disease; Hepatic Encephalopathy; Benzodiazepines; Antipsychotics.
Watch this article’s video abstract and others at http://bit.ly/1C2wSLn. Scan the quick response (QR) code to the left with your mobile device to watch this article’s video abstract and others. Don’t have a QR code reader? Get one by searching ‘QR Scanner’ in your mobile device’s app store.
he morbidity and costs associated with hospitalization for patients with decompensated cirrhosis, especially for those with hepatic encephalopathy (HE), are significant.1–5 Hospitalization is associated with many unforeseen risks. Unfortunately, patients with cirrhosis are a particularly vulnerable subset of the hospitalized population and are prone to nosocomial complications that contribute significantly to liverrelated morbidity and mortality.6,7 Falls are an important, preventable cause of morbidity for hospitalized patients. All hospitals strive to avoid
T
falls.8–11 Furthermore, reducing or eliminating falls is a focus of health policy that is tied to hospital reimbursement. The Centers for Medicare and Medicaid Services includes falls in the list of Hospital Acquired Conditions for which steep financial penalties may be assessed. Efforts to curtail falls are important for patient safety above all, but they are also imbued with some urgency owing to the possibility of financial penalty. Attention should therefore be paid toward modifiable risk factors for falls. Some modifiable risk factors include the use of psychoactive medications and physical tethers.11–13 However, there are limited data regarding risk factors for falls among inpatients with cirrhosis. Abbreviations used in this paper: AUROC, area under the receiver operating curve; CI, confidence interval; HE, hepatic encephalopathy; MELD, Model for End-Stage Liver Disease. Most current article © 2015 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2015.03.019
September 2015
Outpatients with cirrhosis and HE have been shown to be at increased risk for falls.14 Roman et al14 and Soriano et al15 previously demonstrated that 1 in 8 outpatients with cirrhosis reported falls, whereas 2 in 5 patients with minimal HE reported falls. Furthermore, this risk was increased by the concurrent use of psychoactive medications and cognitive dysfunction. Patients with cirrhosis who received psychoactive medications and tethers including physical restraints and bladder catheters were more likely to fall. Herein, we describe the risk and clinical factors associated with falls among our inpatients with cirrhosis.
Methods We conducted a retrospective cohort study including all admissions to the liver unit at the Beth Israel Deaconess Medical Center (Boston, MA) between March 1, 2010 and September 15, 2013. Criteria for admission to this service entail an established diagnosis of cirrhosis with decompensation, acute alcoholic hepatitis, or a medically complicated liver transplant. We reviewed all patient charts and excluded the admissions involving patients with liver transplants that have not been complicated by cirrhosis (n ¼ 314). All clinical care was provided on the dedicated inpatient hepatology unit on a hospital floor reserved for patients with decompensated cirrhosis. It is a teaching service staffed by resident physicians and a transplant hepatologist. No changes in the number of staff, nursing, or house staff occurred during the time under study. This study was conducted in accordance with the Declaration of Helsinki and approved by our Institutional Review Board as exempt from review. All data were collected from clinical activities maintained in a Health Insurance Portability and Accountability Act compliant clinical information database. The cohort design and analysis of this study were performed consistent with STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines.16 Clinical data and demographics were abstracted from the clinical information repository. The Model for EndStage Liver Disease (MELD) score was calculated by using the United Network for Organ Sharing modification according to previously described algorithms.17 The Charlson comorbidity index was calculated by using International Classification of Diseases, 9th Revision codes according to previously described processes.18 The 90-day mortality rate for all patients was determined by using a validated search of the Social Security Death Index.19 Length of stay, medication administration, and procedures were recorded for all patients and censored at the time of a fall (ie, clinical data were only tracked until the time of a fall). The study outcome was an inpatient fall. All falls occurring at Beth Israel Deaconess Medical Center are reported through a dedicated fall monitoring system and are graded in a standard fashion by using the rubric described by the American Nurses Association–National
Psychoactive Medications, Cirrhosis, and Falls 1671
Database of Nursing Quality Indicators as 0 (no intervention), 1 (requires monitoring), 2 (minor harm), 3 (moderate harm), and 4 (major harm).20 Fall-related injuries were defined as any fall graded as a 2, 3, or 4. These grades are assigned by an independent reviewer in the Department of Patient Safety and Quality. Two lines of analysis were developed to assess the risk of falls for inpatients with cirrhosis. First, we assessed the risk factors associated with falls within the population presenting to the liver unit. Exposure variables included age, MELD score, sodium level, Charlson comorbidity index, alcoholic liver disease, acute alcoholic hepatitis, as well as the use of restraints, bladder catheters, opiates, benzodiazepines, or antipsychotics. The opiates assessed included morphine, oxycodone, hydromorphone, and hydrocodone. The benzodiazepines included diazepam, lorazepam, clonazepam, alprazolam, zolpidem, and chlordiazepoxide. Zolpidem, a non-benzodiazepine sedative hypnotic, was included in the benzodiazepine category because its sedative effects are based on the fundamental mechanism of neurotransmission.21 It was analyzed separately to demonstrate a consistent trend. The antipsychotics assessed included olanzapine, risperidone, quetiapine, and haloperidol. The restraints assessed included mitts and soft and leather tethers. The bladder catheters assessed included internal catheters and condom catheters. For patients who experienced a fall, medications, restraints, and catheters were only included if they were present before or on the day of the fall. Second, we determined the rate ratio for a fall in the liver unit’s population compared with patients admitted to a general medical service. Two general medicine units were included in the analysis. The rates of any fall as well as fall-related injuries per patient-day were calculated.
Analysis Data were summarized as mean þ standard deviation for normally distributed, median (25th and 75th percentiles) for non-normally distributed continuous outcomes, or counts and percentages for categorical outcomes. Normal distributions were determined by visual inspection (for obvious skew or kurtosis) and otherwise confirmed by a Shapiro-Wilk test. For two-group comparisons, we compared means of normally distributed continuous outcomes, medians of non-normally distributed continuous outcomes, and frequencies of categorical outcomes. Because of the rarity of falls, we used propensity scores to increase the efficiency of logistic regression in the setting of rare outcomes. The propensity score approach seeks to reduce biases by constructing the probability of an outcome given numerous covariates. The propensity score for a given patient yields the probability (on a scale of 0 to 1) of being exposed to our chosen factors (psychotropic medications) conditional on that patient’s pre-exposure demographics and clinical characteristics by using multivariable logistic
1672 Tapper et al
Clinical Gastroenterology and Hepatology Vol. 13, No. 9
Table 1. Demographics and Clinical Characteristics of Admissions to the Liver Unit Admissions with a fall
Admissions without a fall
Table 2. Description of Circumstances Surrounding Falls Among Cirrhotic Patients P value
Number (N) 55 1694 Average 55.0 (9.7) 56.8 (11.7) .36 age, y (SD) Male, n (%) 32 (58.2) 1052 (62) .48 Admission MELD 18.3 (14.4–28.3) 17.5 (12.6–23.4) .01 score, median (IQR) Admission median 135 (129–138) 135 (131–138) .40 sodium level, meq/L (IQR) Active HE, 38 (69) 898 (53) .002 n (%) Alcoholic liver 24 (43.6) 528 (31) .05 disease, n (%) Alcoholic hepatitis, 8 (14.5) 148 (8.7) .16 n (%) Median length of 6 (1.9–17.3) 4 (2–8)
Psychoactive Medications Increase the Risk of Falls and Fall-related Injuries in Hospitalized Patients With Cirrhosis Elliot B. Tapper, Yesenia Risech-Neyman, and Neil Sengupta Department of Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts BACKGROUND & AIMS:
Reducing or eliminating falls is a focus of patient safety programs as well as health policy. Falls are tied to hospital reimbursement. However, little is known about the risk of falls among hospitalized patients with cirrhosis or factors that affect risk of falling.
METHODS:
We conducted a retrospective cohort study of inpatients with cirrhosis from 2010 to 2013 at a liver transplant center. Our primary aim was to determine the clinical factors associated with falls and fall-related injuries for patients with cirrhosis. Our secondary aim was to describe the rate ratio of falls and fall-related injuries among patients with cirrhosis compared with general medical inpatients.
RESULTS:
During the study period, there were 1749 admissions to the liver service; 55 (3.1%) resulted in falls. Patients who fell were more likely to have received benzodiazepines (50.9% vs 16.7%, P < .0001) and antipsychotic agents (30.9% vs 7.3%, P < .0001). After adjusting for hepatic encephalopathy, the respective odds of a fall after benzodiazepine or antipsychotic exposure were 6.59 (95% confidence interval [CI], 3.76L11.59) and 3.72 (95% CI, 1.90L7.06). The adjusted risk of a fall-related injury was also significantly associated with benzodiazepine and antipsychotic agents, with respective odds ratios of 3.45 (95% CI, 1.39L8.23) and 3.42 (95% CI, 1.09L8.99). Fall-related injuries occurred at a rate of 1.70/1000 patient-days for patients with cirrhosis vs 0.5/1000 patient-days for patients in the general medical service. Accordingly, the rate ratio for a fall-related injury among patients with cirrhosis was 3.37 (95% CI, 1.99L5.72; P < .0001).
CONCLUSIONS:
Psychoactive medications are associated with an increased adjusted risk of falls and fall-related injuries in hospitalized patients with cirrhosis.
Keywords: Model for End-Stage Liver Disease; Hepatic Encephalopathy; Benzodiazepines; Antipsychotics.
Watch this article’s video abstract and others at http://bit.ly/1C2wSLn. Scan the quick response (QR) code to the left with your mobile device to watch this article’s video abstract and others. Don’t have a QR code reader? Get one by searching ‘QR Scanner’ in your mobile device’s app store.
he morbidity and costs associated with hospitalization for patients with decompensated cirrhosis, especially for those with hepatic encephalopathy (HE), are significant.1–5 Hospitalization is associated with many unforeseen risks. Unfortunately, patients with cirrhosis are a particularly vulnerable subset of the hospitalized population and are prone to nosocomial complications that contribute significantly to liverrelated morbidity and mortality.6,7 Falls are an important, preventable cause of morbidity for hospitalized patients. All hospitals strive to avoid
T
falls.8–11 Furthermore, reducing or eliminating falls is a focus of health policy that is tied to hospital reimbursement. The Centers for Medicare and Medicaid Services includes falls in the list of Hospital Acquired Conditions for which steep financial penalties may be assessed. Efforts to curtail falls are important for patient safety above all, but they are also imbued with some urgency owing to the possibility of financial penalty. Attention should therefore be paid toward modifiable risk factors for falls. Some modifiable risk factors include the use of psychoactive medications and physical tethers.11–13 However, there are limited data regarding risk factors for falls among inpatients with cirrhosis. Abbreviations used in this paper: AUROC, area under the receiver operating curve; CI, confidence interval; HE, hepatic encephalopathy; MELD, Model for End-Stage Liver Disease. Most current article © 2015 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2015.03.019
September 2015
Outpatients with cirrhosis and HE have been shown to be at increased risk for falls.14 Roman et al14 and Soriano et al15 previously demonstrated that 1 in 8 outpatients with cirrhosis reported falls, whereas 2 in 5 patients with minimal HE reported falls. Furthermore, this risk was increased by the concurrent use of psychoactive medications and cognitive dysfunction. Patients with cirrhosis who received psychoactive medications and tethers including physical restraints and bladder catheters were more likely to fall. Herein, we describe the risk and clinical factors associated with falls among our inpatients with cirrhosis.
Methods We conducted a retrospective cohort study including all admissions to the liver unit at the Beth Israel Deaconess Medical Center (Boston, MA) between March 1, 2010 and September 15, 2013. Criteria for admission to this service entail an established diagnosis of cirrhosis with decompensation, acute alcoholic hepatitis, or a medically complicated liver transplant. We reviewed all patient charts and excluded the admissions involving patients with liver transplants that have not been complicated by cirrhosis (n ¼ 314). All clinical care was provided on the dedicated inpatient hepatology unit on a hospital floor reserved for patients with decompensated cirrhosis. It is a teaching service staffed by resident physicians and a transplant hepatologist. No changes in the number of staff, nursing, or house staff occurred during the time under study. This study was conducted in accordance with the Declaration of Helsinki and approved by our Institutional Review Board as exempt from review. All data were collected from clinical activities maintained in a Health Insurance Portability and Accountability Act compliant clinical information database. The cohort design and analysis of this study were performed consistent with STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines.16 Clinical data and demographics were abstracted from the clinical information repository. The Model for EndStage Liver Disease (MELD) score was calculated by using the United Network for Organ Sharing modification according to previously described algorithms.17 The Charlson comorbidity index was calculated by using International Classification of Diseases, 9th Revision codes according to previously described processes.18 The 90-day mortality rate for all patients was determined by using a validated search of the Social Security Death Index.19 Length of stay, medication administration, and procedures were recorded for all patients and censored at the time of a fall (ie, clinical data were only tracked until the time of a fall). The study outcome was an inpatient fall. All falls occurring at Beth Israel Deaconess Medical Center are reported through a dedicated fall monitoring system and are graded in a standard fashion by using the rubric described by the American Nurses Association–National
Psychoactive Medications, Cirrhosis, and Falls 1671
Database of Nursing Quality Indicators as 0 (no intervention), 1 (requires monitoring), 2 (minor harm), 3 (moderate harm), and 4 (major harm).20 Fall-related injuries were defined as any fall graded as a 2, 3, or 4. These grades are assigned by an independent reviewer in the Department of Patient Safety and Quality. Two lines of analysis were developed to assess the risk of falls for inpatients with cirrhosis. First, we assessed the risk factors associated with falls within the population presenting to the liver unit. Exposure variables included age, MELD score, sodium level, Charlson comorbidity index, alcoholic liver disease, acute alcoholic hepatitis, as well as the use of restraints, bladder catheters, opiates, benzodiazepines, or antipsychotics. The opiates assessed included morphine, oxycodone, hydromorphone, and hydrocodone. The benzodiazepines included diazepam, lorazepam, clonazepam, alprazolam, zolpidem, and chlordiazepoxide. Zolpidem, a non-benzodiazepine sedative hypnotic, was included in the benzodiazepine category because its sedative effects are based on the fundamental mechanism of neurotransmission.21 It was analyzed separately to demonstrate a consistent trend. The antipsychotics assessed included olanzapine, risperidone, quetiapine, and haloperidol. The restraints assessed included mitts and soft and leather tethers. The bladder catheters assessed included internal catheters and condom catheters. For patients who experienced a fall, medications, restraints, and catheters were only included if they were present before or on the day of the fall. Second, we determined the rate ratio for a fall in the liver unit’s population compared with patients admitted to a general medical service. Two general medicine units were included in the analysis. The rates of any fall as well as fall-related injuries per patient-day were calculated.
Analysis Data were summarized as mean þ standard deviation for normally distributed, median (25th and 75th percentiles) for non-normally distributed continuous outcomes, or counts and percentages for categorical outcomes. Normal distributions were determined by visual inspection (for obvious skew or kurtosis) and otherwise confirmed by a Shapiro-Wilk test. For two-group comparisons, we compared means of normally distributed continuous outcomes, medians of non-normally distributed continuous outcomes, and frequencies of categorical outcomes. Because of the rarity of falls, we used propensity scores to increase the efficiency of logistic regression in the setting of rare outcomes. The propensity score approach seeks to reduce biases by constructing the probability of an outcome given numerous covariates. The propensity score for a given patient yields the probability (on a scale of 0 to 1) of being exposed to our chosen factors (psychotropic medications) conditional on that patient’s pre-exposure demographics and clinical characteristics by using multivariable logistic
1672 Tapper et al
Clinical Gastroenterology and Hepatology Vol. 13, No. 9
Table 1. Demographics and Clinical Characteristics of Admissions to the Liver Unit Admissions with a fall
Admissions without a fall
Table 2. Description of Circumstances Surrounding Falls Among Cirrhotic Patients P value
Number (N) 55 1694 Average 55.0 (9.7) 56.8 (11.7) .36 age, y (SD) Male, n (%) 32 (58.2) 1052 (62) .48 Admission MELD 18.3 (14.4–28.3) 17.5 (12.6–23.4) .01 score, median (IQR) Admission median 135 (129–138) 135 (131–138) .40 sodium level, meq/L (IQR) Active HE, 38 (69) 898 (53) .002 n (%) Alcoholic liver 24 (43.6) 528 (31) .05 disease, n (%) Alcoholic hepatitis, 8 (14.5) 148 (8.7) .16 n (%) Median length of 6 (1.9–17.3) 4 (2–8)
