Loading ...

Patients hospitalized for alcohol-related harms are a diverse population with multiple subgroups, and those with liver disease face the highest risks for in-hospital and post-discharge mortality, new research suggested.

The subgroups followed a severity gradient, and patients in the more severe categories accounted for most of the adverse in-hospital and post-discharge outcomes, according to lead author Erik L. Friesen, PhD, of the Temerty Faculty of Medicine at the University of Toronto, Toronto, Ontario, Canada, and colleagues.

"Previous research on population-level trends in alcohol-related hospitalizations has generally combined all individuals who experience alcohol-related hospitalizations into a single group of people who are assumed to have an alcohol use disorder (AUD)," Friesen told Medscape Medical News. "We felt as though this was an oversimplification of a much more complex clinical reality."

Indeed, his team found that two subgroups — patients with high-frequency alcohol-related health service use and patients with alcohol-related liver disease — had substantially higher rates of short-term readmission and mortality than the rest of the cohort.

In Ontario, the liver disease subgroup made up about 15% of the cohort, but its 1-year mortality rate was about 30%, compared with 12% in the overall cohort. "This was a strikingly high mortality rate, given that the average age of this patient subgroup was 61 years," said Friesen.

The study was published online on January 31 in JAMA Network Open.

Distinct Subgroups

The researchers used latent class analysis (LCA) to identify clinical subgroups of patients with alcohol-related hospitalizations in Manitoba and Ontario. They examined the following four exposures of interest: The alcohol-related diagnostic codes associated with the index hospitalization, the number of alcohol-related outpatient visits, the number of alcohol-related emergency department (ED) visits, and alcohol-related hospitalizations in the 2 years before the admission date of the index hospitalization.

The primary outcomes were in-hospital mortality, time to alcohol-related hospital readmission, and time to mortality in the year following discharge from the index hospitalization.

A total of 34,043 patients were included in the analysis: 4753 from Manitoba (median age, 49 years; 37.6% women) and 29,290 from Ontario (median age, 57 years; 29.1% women).

The researchers identified seven subgroups, following a gradient from low-frequency service use for acute intoxication to high-frequency service use for severe AUD and liver disease.

In Manitoba, the LCA found that a model with five subgroups was the best fit. These subgroups included patients with acute intoxication and a low average frequency of prior alcohol-related health service use (2.7%); those with harmful alcohol use, relatively few alcohol-related comorbidities, and a low frequency of prior alcohol-related health service use (29.2%); those with alcohol dependence, more alcohol-related comorbidities, and an average frequency of prior alcohol-related health service use (31.9%); those presenting for withdrawal with a high average frequency of prior alcohol-related health service use (24.3%); and those with alcohol-related liver disease and the highest frequency of prior alcohol-related health service use (11.8%).

In Ontario, a model with seven subgroups was the best fit. The first five subgroups mirrored those identified in Manitoba, and 3.6% of the cohort was in the acute intoxication subgroup, 19.4% was in the harmful use subgroup, 30.8% was in the alcohol dependence subgroup, 20.4% was in the withdrawal subgroup, and 15.1% was in the liver disease subgroup.

Two additional groups emerged in Ontario. One, representing 5.2% of the cohort, had a high frequency of all types of alcohol-related health service use (ie, outpatient, ED, and inpatient), and the other, representing 5.5% of the cohort, had a high frequency of prior alcohol-related ED visits and hospitalizations but less frequent prior alcohol-related outpatient visits.

In Ontario, 4431 patients in the liver disease subgroup, representing 15.5% of the cohort, were at the highest risk for 1-year mortality (31.2%), relative to the acute intoxication subgroup (4.0%). The adjusted hazard ratio (aHR) was 3.83.

There was also a small subgroup (10.6%) of patients with high-frequency alcohol-related health service use who had a much higher hazard of 1-year readmission following the index hospitalization (46.1% vs 9.8% in the acute intoxication subgroup; aHR, 5.09).

Prognostic Information

Across provinces, 257 patients in Manitoba (5.4%) and 2197 in Ontario (7.5%) died during the index hospitalization. Of those who survived, 965 in Manitoba (20.3%) and 5301 in Ontario (18.1%) were readmitted to the hospital, and 399 in Manitoba (8.4%) and 3544 in Ontario (12.1%) died within 1 year of discharge.

In both cohorts, patients in the liver disease subgroup had the highest incidence of in-hospital and post-discharge mortality.

Compared with these overall trends, patients in the acute intoxication subgroup were proportionately younger (median age, 39 years). More were female (50.7%), and the subgroup had fewer medical comorbidities and more psychiatric comorbidities (78.7% had previous psychiatric care).

"This study identified distinct clinical subgroups of individuals hospitalized for alcohol-related harms," the authors wrote. "Efforts to reduce high rates of readmission and mortality among individuals experiencing alcohol-related hospitalizations may consider prioritizing those at the highest risk of short-term harm, including individuals with alcohol-related liver disease and high-frequency health service use."

The study had limitations. Manitoba and Ontario have universal health insurance, and how people access health services for alcohol-related harms in Canada may differ in places without universal healthcare. There were two subgroups observed in Ontario that were not observed in Manitoba, which could mean that there is regional variability in how people experience alcohol-related hospitalizations. Due to a lack of data availability, the authors did not consider prior use of AUD medications or private addiction services when characterizing prior alcohol-related health service use.

"Our study indicates that individuals presenting with either liver disease or a history of high-frequency health service use are at a disproportionately high risk of readmission and death when they leave the hospital," said Friesen. "This prognostic information could be helpful in conversations between patients, their support networks, and the healthcare team about post-discharge care planning."

Post-Discharge Care

Commenting on the findings for Medscape Medical News, Natalie Klag, MD, assistant professor of psychiatry at the Ohio State University Wexner Medical Center in Columbus, Ohio, said that patients with more severe alcohol use tend to use healthcare resources, particularly hospital-based resources, more frequently. But they are not typically triaged after their hospitalization based on their pattern of alcohol use, she said. Rather, "services are offered to anyone who could benefit from treatment engagement." Klag was not involved in the study.

Missing from this study is a consideration of how engagement in treatment for substance use disorders may modify these outcomes, Klag said. The modification could be global for all study participants or affect each individual group.

"While we know that those with more medical comorbidities as a result of their alcohol use are more likely to engage with the medical system," she said, "there was no delineation of whether [their] subsequent visits were related to alcohol use or a consequence of their medical condition, albeit a condition that was caused by alcohol use."

Clinicians must recognize signs of AUD and problematic alcohol use to refer patients to the resources that could help modify their pattern of use, she said. "Brief medical hospitalization does not change the trajectory of an AUD, so ongoing care after this hospitalization is the best chance there is to modify the disease."

"If someone presents for a medical admission that is a result of alcohol, it is important to refer them for further care or consult an addiction medicine consult service to provide linkage," Klag concluded.

This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health and the Ministry of Long-Term Care. Friesen and Klag disclosed no conflicts.