Age and Sex Differences in Alcohol Withdrawal Syndrome Authors: Hayrunnisa Unlu, Nan Zhang, Asmaa Yehia, Vugar Isazade, Molly M. Jeffery, Osama A. Abulseoud Published: 2026 ABSTRACT ObjectiveTo investigate alcohol withdrawal syndrome prevalence among women and to study age and sex differences in clinical manifestations and hospital course.Patients and MethodsThis cohort study included all hospitalized patients from June 1, 2019 to June 1, 2022, where the Clinical Institute Withdrawal Assessment for Alcohol Scale, revised, protocol for alcohol withdrawal syndrome was implemented.ResultsA total of 16,190 hospitalizations (10,092 patients aged ≥21 years), with 30.2% women, were included in the study and divided into the following 4 age groups: 21 to 39 years (n=2453 [24.3%]), 40-64 years (n=5128 [50.8%]), 65-74 years (n=1705 [16.9%]), and 75 years or older (n=806 [8%]). We considered the age group (40-64 years) with the highest number of patients as the reference group. Older patients presented with lower blood alcohol concentration and took longer to reach peak withdrawal manifestations than younger patients. Compared with women, men in the youngest age group reached a peak withdrawal earlier (mean, 18.5 hours [95% CI, 17.2-19.8] vs 19.4 hours [95% CI, 17.5-21.3]; P<.001), and there was no sex difference in other age groups. Men required higher benzodiazepine doses during hospitalization in the youngest (mean, 20.1 mg [95% CI, 17.1-23.1] vs 13.6 mg [95% CI, 11.4-15.8]; P=.013) and the reference (mean, 18.0 mg [95% CI, 16.5-19.4] vs 12.7 mg [95% CI, 11.4-14]; P<.001) age groups; there was no sex difference among older adults. The benzodiazepine dose during the first 24 hours of hospitalization dose decreased significantly as age increased (21-39 years: mean, 6.64 mg [95% CI, 6.3-7.0]; 40-64 years: mean, 5.8 mg [95% CI, 5.5-6.0]; 65-74 years: mean, 3.76 mg [95% CI, 3.4-4.1]; ≥75 years: mean, 2.79 mg [95% CI, 2.4-3.2]; P<.001 for all groups). There was no sex difference in the all-cause mortality rate, including posthospitalization mortality, among any age group.ConclusionThe results of this study show a narrowing of the traditional sex gap in alcohol withdrawal. The Royal College of Physicians published the first clinical description of alcohol withdrawal delirium in 1813 by Thomas Sutton.1 Two hundred years later, alcohol withdrawal became a significant public health problem.2 Data from the US Census Bureau documents a steady increase in monthly retail sales of beer, wine, and liquor stores from 2.5 billion dollars in December 1992 to 8.0 billion dollars in December 2023,3 reflecting a marked increase in per capita alcohol consumption over the past 3 decades.4,5 This new heavy drinking pattern is evident among older adults6 and women,7, 8, 9, 10 leading to more emergency room visits and hospitalizations for alcohol-related problems,6,11 including for alcohol withdrawal syndrome (AWS).12 Approximately 5%13 to 7.3%14 of all medical inpatients and 4.4% of patients admitted to the intensive care unit (ICU)15 and 1% of all trauma admissions16 are associated with AWS and subsequent prolonged hospital and ICU length of stay (LOS)15,17 and higher mortality rates.17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27 Complicating matters further is the rapid growth of the older adult population28,29 and the increased prevalence of excessive drinking among older adults9 and women,10,30, 31, 32, 33 combined with a dearth of research on age- and sex-related differences in AWS. The results of our recent sex difference review of 35 studies, including 75,346 patients with AWS, suggest that some AWS manifestations present differently in men and women.34 For example, women were more likely to have comorbid pancreatitis35 and longer hospital LOS,36 whereas men were more likely to have comorbid cardiovascular disorders,37 receive diazepam more frequently at higher doses above 30mg of diazepam,36 and suffer posthospitalization (over 16 years follow-up) mortality approximately 1.5 years sooner than women; P<.05).36 However, most of the included studies were not focused on examining sex differences and had a small number of patients, particularly women.34 Similarly, few studies have examined the impact of age on AWS. Liskow et al38 reported that older patients (≥58 years old, n=26) had a more severe withdrawal and received higher benzodiazepine doses compared to younger patients (21-33 years old, n=24).38 On the other hand, Wojnar et al36 reported that compared to younger patients with AWS (<60 years , n=1163), older patients (≥60 years, n=50) consumed significantly less alcohol during the drinking bout preceding admission for AWS and had more extended hospital LOS, but no differences were found between age groups in the duration of the AWS episode, the severity of AWS symptoms, or benzodiazepine doses.39 One preclinical study compared young (3-4 months) and old (29-30 months) male Wistar rats who received high doses of ethanol through intragastric gavage. The study reported no significant difference in ethanol withdrawal manifestations between the two age groups.40 In this study, we examined in a large sample (n=10,092 patients and 16,190 admissions) age and sex differences in several aspects of AWS, including demographic characteristics, medical and psychiatric comorbidities, hospital course, ICU admission, laboratory findings at the time of admission, and time to reach peak withdrawal severity as assessed by the Clinical Institute Withdrawal Assessment for Alcohol Scale, revised (CIWA-Ar),41 benzodiazepine treatment, and mortality. Patients and Methods Study Cohort The study was approved by the institutional review board of Mayo Clinic (ID#22-008591). We collected the electronic medical records of all admissions at Mayo Clinic Hospitals (Rochester, Florida, and Arizona) and Mayo Clinic Health System from June 1, 2019 to June 1, 2022 for all patients who received an order to implement the CIWA-Ar protocol. Patients were placed on the CIWA-Ar protocol when they report a recent cessation of heavy drinking, present with a high blood alcohol concentration (BAC), or exhibit AWS manifestations. The CIWA-Ar scale is a 10-item survey used to quantify the severity of withdrawal manifestations and guide lorazepam dose administration : (1) nausea and vomiting; (2) tremors; (3) sweating; (4) anxiety; (5) agitation; (6) tactile disturbances; (7) auditory disturbances; (8) visual disturbances; (9) headache; and (10) disorientation or clouding of the sensorium. Each item is scored 0 to 7 based on severity, except item 10, scored 0 to 4. Total scores on the CIWA-Ar range from 0 to 67 points.41 Patients who score 0 to 9 do not receive lorazepam treatment; 10 to 12 receives 1 mg of lorazepam orally or intravenously (PO/IV); 13-14 receives 2 mg of lorazepam PO/IV; 15-17 receives 3 mg of lorazepam PO/IV; and 18 or greater receives 4 mg of lorazepam PO/IV. Alcohol withdrawal syndrome symptoms are reassessed 30 minutes after lorazepam administration for further dosing. Compared with fixed benzodiazepine dose, this symptom-triggered lorazepam approach has been associated with lowering benzodiazepine doses, which yields less sedation, shorter hospitalizations, fewer adverse reactions, and a decreased risk of respiratory depression.42 Adolescents and young adults (≤20 years, n=130 patients, and 148 admissions) were reported in a separate article.43 This study reports on adult patients (≥21 years old). In cases where patients had more than 1 admission during the study period, all admissions were retained in the analysis. Demographic characteristics is reported as each patient’s first identified encounter. The active problem list at the time of each admission was used to determine medical and psychiatric comorbidities and substance use profiles. Hospital and ICU LOS were calculated for each admission episode. The timestamps of individual CIWA-Ar assessments and benzodiazepine treatment were collected for each admission to identify the time to peak withdrawal severity and total benzodiazepine dose throughout the hospital LOS. All administered benzodiazepine doses were converted into lorazepam dose equivalents.44 Mortality data were ascertained through June 2023. For demographic characteristics and mortality analyses, unique patient-level data based on each individual’s first admission were used. For analyses of comorbid medical and psychiatric conditions and hospital course, each admission episode was used as a separate data point. Statistical Analyses The normality of continuous variables was assessed using the Shapiro-Wilk test. Data were presented as mean and SD if normally distributed or as median, quartiles, and interquartile range (IQR) if not normally distributed, whereas categorical data were expressed as percentages with 95% CIs. We grouped the subjects into age groups: 21-39, 40-64, 65-74, and 75 years or older and considered the age group (40-64 years) with the highest number of patients as the reference group for age comparisons. The 95% CIs and P values for admission-level summary statistics were calculated using generalized estimation models to account for correlation from multiple observations for some patients. Both the mean or median values and P values for in Supplementary Tables 1 to 3 were adjusted using generalized estimating equations to account for repeated measurements from multiple hospital visits during the clinical course. We used linear regression to examine the relationship between the number of comorbid medical conditions and ICU LOS. Kaplan-Meier survival analysis was used to estimate differences in time from hospital discharge to death between age and sex groups. P values of ≤.05 were considered statistically significant for the purposes of hypothesis testing. All analyses were performed using R software (version 4.2.1; R Foundation for Statistical Computing). Results Demographic Characteristics Between June 2019 and June 2022, we identified 10,092 patients aged 21 years and older, with 16,190 hospital admissions implementing the CIWA-Ar protocol. Within this cohort, 30.2% (n=3046) were women. Most patients were White (90%; 95% CI, 89%-90%) and non-Hispanic (94%; 95% CI, 93%-94%). More than half of the patients were unemployed or retired (53%; 95% CI, 52%-54%), and 40% (95% CI, 39%-41%) were single. Most of the patients were in the 40-64-year-old age group (50.8% [n=5128]), followed by the 21-39-year-old (24.3% [n=2453]), 65-74-year-old (16.9% [n=1705]), and 75-year-old or older (8% [n=806] groups. Moreover, 24% (95% CI, 23%-25%) of the patients had more than 1 hospital admission with CIWA-Ar implementation for AWS during the study interval. The reference age group (40-64 years) had the highest rate of multiple hospital admissions where the CIWA-Ar protocol was initiated for AWS during the study period (27%; 95% CI, 25-28), with no significant difference than the youngest group (26%; 95% CI, 25%-28%). In contrast, older patients had significantly lower rates of multiple admissions (65-74 years: 19% [95% CI, 18%-21%]; P<.001; ≥75 years: 14% [95% CI, 12%-17%]; P<.001) (Supplementary Table 1). Please also see Supplementary Table 2 for commonly used substances and Supplementary Table 3 for Demographics of patients with multiple admissions. Medical Comorbidities Among all patients, the most common medical comorbidities were cardiovascular (47.4%; 95% CI, 46.3%-48.5%), followed by gastrointestinal (40.4%; 95% CI, 39.3%-41.4%), endocrinological (35%; 95% CI, 34%-36.1%), and neurological disorders (35%; 95% CI, 34%-36.1%). As expected, we observed an increased prevalence of medical comorbidities with increasing age. Among all patients, women had a higher prevalence of comorbid gastrointestinal disorders (women vs men: 42.3% [95% CI, 40.5%-44%] vs 38.5% [95% CI, 37.4%-39.6%]; P=.004), anemia (20.9% [95% CI, 19.5%-22.4%] vs 16.1% [95% CI, 15.3%-17%]; P<.001), nutritional disorders (12.8% [95% CI, 11.7%-14%] vs 9.6% [95% CI, 8.9%-10.3%]; P<.001), orthopedic problems (18.2% [95% CI, 16.9%-19.6%] vs 14.9% [95% CI, 14.1-15.7]; P<.001), and infectious diseases (16% [95% CI, 14.7%-17.3%] vs 13.1% [95% CI, 12.4%-13.9%]; P=.002). The higher prevalence of comorbid gastrointestinal disorders among women was driven by the 40- to 64-year-old age group and was not observed in other age groups. Conversely, men had a higher prevalence of cardiovascular (men vs women: 53.3% [95% CI, 52.2%-54.5%] vs 41.5% [95% CI, 39.8%-43.3%]; P<.001), neurological (37.1% [95% CI, 35.9%-38.2%] vs 33.1% [95% CI, 31.4%-34.7%]; P=.005), and endocrinological disorders (37.1% [95% CI, 35.9%-38.2%) vs 33.1% [95% CI, 31.4%-34.7%]; P=.002). The male predominance for cardiovascular disorders was consistent across age groups, except in patients aged 75 years or older, where no significant sex difference was observed (Supplementary Table 4). Psychiatric Comorbidities We observed a decreased prevalence of psychiatric comorbidities with increasing age. There was no significant sex difference in the prevalence of previously diagnosed alcohol use disorder (AUD) (men vs women: 64.1% [95% CI, 63-65.2] vs 62.8% [95% CI, 61.1-64.4]; P=.836) or substance use disorder (SUD) other than alcohol (8.7% [95% CI, 7.7-9.7] vs 8.8% [95% CI, 8.2-9.4]; P>.99]. However, a notable age difference was observed in both disorders. Alcohol use disorder was most prevalent among individuals aged 40 to 64 years, and SUD was most prevalent among individuals aged 21 to 39 years, with a decline in prevalence as age increased. Compared with men, women had a high prevalence of bipolar disorder (women vs men: 7.3% [95% CI, 6.4%-8.2%), vs 3.7% [95% CI, 3.3%-4.2%]; P<.001), major depression disorder (41.8% [95% CI, 40.1%-43.6%] vs 25.4% [95% CI, 24.4%-26.4%]; P<.001), generalized anxiety disorder (36% [95% CI, 34.3%-37.7%] vs 20.1% [95% CI, 19.2%-21.1%]; P<.001), posttraumatic stress disorder (7.9% [95% CI, 7%-8.9%] vs 3.2% [95% CI, 2.8%-3.7%]; P<.001), sleep disorders (9% [95% CI, 8%-10%] vs 6.4% [95% CI, 5.9%-7%]; P<.001), and previous suicidal ideation (6.6% [95% CI, 5.8%-7.5%] vs 4.2% [95% CI, 3.8%-4.7%]; P<.001) or attempts (2.8% [95% CI, 2.3%-3.5%] vs 1.7% [95% CI, 1.4%-2%]; P=.002) (Supplementary Table 5 ). Hospital LOS As detailed in Supplementary Table 6, the mean hospital LOS was 121.1 hours (95% CI, 117.4-124.8); 80% of patients were discharged within 7 days. Compared with the reference age group (40-64 years: mean LOS, 119.6 hours; 95% CI, 114.5-124.7), older patients had significantly longer LOS (65-74 years: mean, 149.9 hours [95% CI, 137.2-162.5]; ≥75 years: mean, 133.7 hours [95% CI, 123.5-143.8]; P=.01 for both), whereas younger patients had significantly shorter LOS (21-39 years: mean, 101.8 hours; 95% CI, 95.8-107.8; P=.01). Only the oldest age group (≥75 years) showed sex difference in-hospital LOS: men had a longer stay than women (mean, 152 hours [95% CI, 137-167] vs 116 hours [95% CI, 102-129]; P=.011]. ICU Admission Rates and ICU LOS Among all patients, the percentage of patients who were admitted to the ICU was 24.2% (95% CI, 23.4%-25%), with higher rates in men (26.1% [95% CI, 25.2%-27.1%] vs 22.4% [95% CI, 21.1%-23.7%]; P<.001). However, when divided into age groups, no significant sex difference was observed in any age group. No significant age difference was observed between age groups for ICU admission rates. The mean ICU LOS was 65.1 hours (95% CI, 62.4-67.7), with no sex or age differences (Supplementary Table 6). BAC and Liver Function Test Results on Admission The mean BAC at the time of hospital admission among all patients who had BAC checked (16.8%; n=3551 admissions) was 219 mg/dL (95% CI, 241-225). There was no sex difference in the mean BAC level in any age group. The median BAC was highest in the youngest age group (21-39 years: 240 mg/dL; IQR, 230-250 mg/dL) and significantly decreased by increasing age (40-64 years: 217 mg/dL [IQR, 210-224]; 65-74 years: 192 mg/dL [IQR, 177-207]; ≥75 years: 160 mg/dL [IQR, 140-180]; P<.001 in all age groups]. Among all patients, alkaline phosphatase was significantly higher in women than that in men (women vs men: 125 IU/L [IQR, 118-132] vs 112 IU/L [IQR, 109-115]; P=.039]. However, when divided into age subgroups, this significant difference persisted only in the reference age group (women vs men: 133 IU/L [IQR, 126-140] vs 118 IU/L [IQR, 113-123]; P=.03). The highest alkaline phosphatase concentration was detected in the reference age group (40-64 years) compared with other age groups. There was no age difference in the mean serum alanine aminotransferase (ALT) levels. There was no sex difference in the serum ALT levels, except in the reference age group, where men had significantly higher ALT levels than women (men vs women: 72.5 IU/L [IQR, 65.6-79.5] vs 57.2 IU/L [IQR, 51.5-62.9]; P=.015). There was no sex difference in the mean serum aspartate aminotransferase levels in any age group. The mean serum aspartate aminotransferase level was highest in the youngest age group (21-39 years: 117 IU/L; IQR, 106-128 IU/L) and significantly decreased by increasing age (40-64 years: 109 IU/L [IQR, 96.4-121]; 65-74 years: 68.7 IU/L [IQR, 62.1-75.2]; ≥75 years: 56.6 IU/L [IQR, 46.3-66.9]; P<.001 in all age groups] (Supplementary Table 6). Peak CIWA-Ar Scores Regarding CIWA-Ar scores, 79% of patients scored 4 or more. Among all patients, the mean peak CIWA-Ar score was 12.7 (95% CI, 12.5-12.8), with a statistically significant sex difference (men vs women: 13.0 [95% CI, 12.8-13.2] vs 12.3 [95% CI, 12.1-12.6]; P<.001). However, when divided into age groups, this significant sex difference only persisted in the youngest age group (men vs women: 13.9 [95% CI, 13.5-14.3] vs 12.8 [95% CI, 12.3-13.3]; P=.012). The mean peak CIWA-Ar score was highest in the youngest age group (21-39 years: 13.4; 95% CI, 13.0-13.7) and significantly decreased by increasing age (40-64 years: 13.1 [95% CI, 12.9-13.3]; 65-74 years: 11.15 [95% CI, 10.8-11.5]; ≥75 years: 10.6 [95% CI, 10.0-11.1]; P<.001 in all age groups] (Supplementary Table 6). Time to Reach Peak CIWA-Ar Score In all patients, the mean time from admission to reach the peak total CIWA-Ar score was 23.5 hours (95% CI, 22.8-24.2 hours, and there was no sex difference between men (24.3 hours; 95% CI, 23.5-25.1) and women (22.7 hours; 95% CI, 21.6-23.9; P=.3]. We observed an increased time to reach the peak CIWA-Ar with increasing age. Compared with the reference age group (40-64 years: 22.8 hours; 95% CI, 21.9-23.7), the 21-39-year-old group reached the peak earlier (18.9 hours; 95% CI, 17.8-20.1; P<.001], whereas older patients reached the peak later (65-74 years: 30.6 hours [95% CI, 28.2-33.1]; P<.001; ≥75 years: 32.9 hours [95% CI, 29.9-35.7]; P<.001]. There was no sex difference in time to reach the peak CIWA-Ar score in any age group (Supplementary Table 6). Benzodiazepine Treatment (Lorazepam Dose Equivalents) The mean benzodiazepine dose during the first 24 hours of hospitalization was significantly decreased by increasing age. The youngest age group (21-39 years: 6.64 mg; 95% CI, 6.26-7.02) received the highest benzodiazepine dose during the first 24 hours, followed by the 40- to 64-year-old (5.8 mg; 95% CI, 5.5-6.0), 65- to 74-year-old (3.8 mg; 95% CI, 3.4-4.1), and 75-year or older (2.8 mg; 95% CI, 2.4-3.2) age groups (P<.001 for all groups). A similar age pattern of benzodiazepine treatment was observed over the whole hospital LOS. Compared with women, men received higher doses of benzodiazepine in the first 24 hours of hospitalization in the youngest and the reference age groups (men vs women—21-39 years: 7.5 mg [95% CI, 7-8] vs 5.8 mg [95% CI, 5.2-6.3]; 40-64 years: 6.5 mg [95% CI, 6.2-6.8] vs 5.1 mg [95% CI, 4.7-5.4]; P<.001 for both). Meanwhile, in older age groups, men and women received similar doses of benzodiazepine (men vs women—65-74 years: 4.2 mg [95% CI, 3.8-4.7] vs 3.3 mg [95% CI, 2.7-3.9]; P=.2; ≥75 years: 3.1 mg [95% CI, 2.5-3.6] vs 2.5 mg [95% CI, 1.9-3.2]; P>.99). A similar age and sex patterns of benzodiazepine treatment were observed over the whole hospital LOS (Supplementary Table 6). High-dose thiamine (≥500 mg) was administered in only 2% of patients. All-Cause Mortality The all-cause mortality rate, including posthospitalization mortality, was 11% (95% CI, 10%-12%) among all patients with no sex difference. Among all patients, the all-cause in-hospital mortality rate was 1.6% (95% CI, 1.3%-1.8%), whereas the posthospitalization mortality was 9.5% (95% CI, 8.9%-10%), with no sex difference. As expected, the mortality rate significantly increased by increasing age (≥75 years: 26% [95% CI, 23%-29%]; 65-74 years: 17% [95% CI, 15%-18%]; 40-64 years: 10% [95% CI, 9.2%-11%]; 21-39 years: 4.3% [95% CI, 3.6%-5.2%). There was no sex difference in any age group, including both in-hospital and posthospitalization mortality rates. Among all patients, the median duration between the first hospitalization and mortality was 0.43 years (IQR, 0.06-1.12 years). Using Kaplan-Meier survival analysis, we did not observe sex or age differences in the median (IQR) duration between hospitalization and death (Supplementary Table 6 and Figure A, B).FigureMortality duration between first hospitalization and death. (A) The survival curve comparison between the men (blue) and women (red) shows there is no significant difference in survival between men and womens. (B) The survival curves for different age groups illustrate that there is no significant difference in survival probability among the various age groups. Mortality duration between first hospitalization and death. (A) The survival curve comparison between the men (blue) and women (red) shows there is no significant difference in survival between men and womens. (B) The survival curves for different age groups illustrate that there is no significant difference in survival probability among the various age groups. Discussion To the best of our knowledge, this is the first study to provide detailed phenotyping of AWS course and treatment outcomes in a large cohort of patients with a wide age distribution and a large group of women. Our results show that the prevalence of medical comorbidities in hospitalized patients with alcohol withdrawal is cumulative. As patients age, all medical conditions increase steadily, whereas certain psychiatric conditions peak between 21-39 and 40-64 years of age. The highest prevalence of prior AUD diagnosis (65%) is documented in patients between 40 and 64 years, whereas 13% of patients between 21 and 39 showed substance use, 7.1% had a bipolar disorder diagnosis, and 3.2% recorded a previous suicide attempt. These very high rates of medical and psychiatric comorbidities are several-fold higher than the general public45,46 and even higher than patients with AUD.47 This may reflect the concept that alcohol withdrawal presents as an advanced stage of severe alcohol use.48 In our cohort, older patients presented with lower BAC and took longer to reach peak withdrawal manifestations than younger patients. They also received lower benzodiazepine doses but had longer hospital stays. One study reported that older adults (≥58 years, n=26) had more severe withdrawal and received higher benzodiazepine doses than younger patients (21-33 years, n=24).38 In contrast, another study reported no difference in AWS severity or benzodiazepine dose between older (≥60 years, n=50) and younger (<60 years, n=1163) patients.39 These different results could be attributed, at least in part, to the small number of older patients in these studies. Among all admissions with AWS in our cohort, only 16.8% (95% CI, 16.1%-17.5%) had a documented BAC level. In literature, few clinical studies have reported BAC levels in patients with AWS.34 Given the potential value of BAC in confirming recent alcohol use and contextualizing withdrawal severity, we encourage clinicians to consider obtaining BAC, particularly in patients with suspected AWS or a history of chronic or heavy alcohol use. In addition to BAC, objective alcohol biomarkers may assist clinicians in evaluating recent alcohol exposure and estimating the risk of AWS. Two increasingly used biomarkers are serum phosphatidylethanol (PEth) and urinary ethyl glucuronide (EtG). These biomarkers are direct metabolites of ethanol and provide additional information about alcohol consumption patterns that may not be captured through patient self-report or BAC alone. Phosphatidylethanol is a phospholipid formed in cell membranes only in the presence of ethanol and can detect alcohol intake up to 4 weeks after last use, depending on the amount and frequency of alcohol consumption. Owing to its longer detection window, it is particularly useful for identifying chronic or heavy alcohol use.49 In contrast, EtG is a minor metabolite of ethanol that can be detected in various body fluids, tissues, and hair; however, urine testing is most commonly used in clinical practice due to its availability. Depending on the assay cutoff and amount consumed, EtG can be detected in the blood for up to 36 hours and in the urine for up to 5 days after heavy alcohol use.50,51 In high-risk settings such as liver transplantation, these biomarkers may complement clinical evaluation by providing objective evidence of recent and chronic alcohol exposure, which is critical for both perioperative risk assessment and posttransplant outcomes.52 The use of these biomarkers may enhance clinical risk stratification for AWS, particularly when the drinking history is uncertain or unavailable. Elevated PEth levels suggest sustained or heavy alcohol use, which is associated with a higher risk of alcohol withdrawal. A positive urine EtG result indicates recent alcohol intake and may support the diagnosis of early withdrawal in patients presenting with nonspecific symptoms. However, each marker has limitations that should be considered in clinical practice. These tests are specimen-dependent and may increase health care costs. Phosphatidylethanol, although highly specific and strongly correlated with cumulative alcohol exposure,53, 54, 55 may yield false-positive results in patients who have received a blood transfusion recently56 and may not detect very recent drinking immediately before admission. Urine EtG, on the contrary, may have better accuracy for alcohol use occurring proximal to specimen collection; the combination of uEtG and PEth together may bring the greatest accuracy for uniform testing protocols in pre–liver transplant evaluation.57 However, uEtG is not feasible in anuric liver transplant patients with hepatorenal syndrome, and it has a shorter detection window and may be susceptible to false-positive (eg, incidental exposure to alcohol-containing products) or false-negative (eg, bacterial degradation in urine samples) results.58 Additionally, uEtG results may be affected by urine infections and reduced kidney function.59 Neither biomarker directly measures physiologic dependence or predicts withdrawal severity; rather, they provide supportive information that should be interpreted along with a history, laboratory findings, and validated withdrawal scales such as the CIWA-Ar. Furthermore, the results of this study show a narrowing of the traditional sex gap in alcohol withdrawal6,7,9,34,60 and how heavy drinking men and women have different vulnerabilities to alcohol.61, 62, 63, 64 Women represented 30% of all admissions in this cohort, and they had a similar prevalence to men in prior AUD and SUD diagnoses, rates of multiple admissions, BAC levels, ICU LOS, and time to peak CIWA-Ar scores, mortality rates, and similar duration between hospitalization and posthospital mortality. In contrast, women had a higher prevalence of comorbid gastrointestinal and neurological disorders and more comorbid psychiatric disorders, specifically bipolar disorder, than men. Nevertheless, men had a higher prevalence of cardiovascular and endocrinological disorders, higher rates of ICU admission, and higher peak CIWA-Ar scores except in the older adult groups and received higher doses of benzodiazepine. Compared with men, heavy drinking among women places them at higher risk for certain medical conditions, such as alcoholic hepatitis and certain cancers.65, 66, 67, 68 The potential mechanisms for this sex difference in medical and psychiatric comorbidities remain under investigation. The neurobiological underpinnings for the differential effect of alcohol on women could be attributed to several factors, such as their ability to achieve a higher BAC than men after ingesting the same dose of ethanol per kilogram of body weight69; the faster hepatic oxidation of ethanol, which accentuates alcohol toxicity by increasing the generation of acetaldehyde70; or sex differences in glutamatergic signaling.71 Alcohol is metabolized first to acetaldehyde, mainly by hepatic72 and gastric73 alcohol dehydrogenase (ADH) enzymes and, to a lesser extent, by cytochrome P450 isoenzyme CYP2E1 and catalase.74 Alcohol dehydrogenase metabolizes alcohol in the presence of Nicotineamide Adenine Dinucleotide NAD into acetaldehyde and NADH. CYP2E1 also oxidizes alcohol into acetaldehyde through the reduction of oxygen molecules to water and NADPH into NADP, generating several reactive oxygen species.74 Both acetaldehyde and reactive oxygen species are toxic to various organs.75,76 Acetaldehyde is metabolized by aldehyde dehydrogenase to acetate.77 Astrocytic aldehyde dehydrogenase mediates the production of acetate in the cerebellum, which contributes to ethanol-induced γ-amino butyric acid synthesis and motor impairment after low-dose ethanol administration.78 The activity of gastric ADH is higher in men and younger adults than that in women and older adults, as reported in some70,79, 80, 81, 82 but not all83, 84, 85, 86, 87 studies. In comparison, the activity of hepatic ADH is higher in women,88,89 with inconsistent results.90 The reduction in gastric ADH in women and older adults is associated with higher blood alcohol concentrations, which may contribute to the enhanced acute effects of alcohol in women and older adults. On the contrary, the increased activity of hepatic ADH in women is associated with enhanced hepatic metabolism of alcohol, increased production of toxic acetaldehyde, and accentuated vulnerability of certain organs to chronic complications. Drinking patterns also contribute to the observed age and sex differences in withdrawal manifestations. In general, the frequency of drinking and the number of heavy drinking episodes91 are higher among men, who tend to drink more liquor92 than those among women, who tend to drink more wine.92 Additionally, there is a steady increase in drinking frequency with age93 compared with binge drinking, which tends to be the prevailing pattern among young, college-age individuals.94 We have recently shown that alcohol withdrawal in young individuals is associated with high suicidal rates.43 In this study, we observed high rates of posthospital mortality. Alcohol is a leading cause of premature mortality worldwide.95 At an average consumption of 4 standard drinks per day, the risk of dying is 4-fold compared with current nondrinkers.96 The number of death certificates indicating alcohol involvement in mortality in the United States doubled between 1999 and 2017,97 and the average annual percentage changes in alcohol-induced death rates increased substantially between 2000 and 2016 in men (1.4%; 95% CI, 1.0%-1.8%) and women (3.1%; 95% CI, 2.6%-3.6%). Furthermore, the reported all-cause mortality rates in patients with AWS vary widely from approximately 1% of trauma patients17,21, 22, 23, 24 to 4.2% in hospitalized patients with AWS and concurrent arrhythmias25 to 8.5% (46/539) of patients admitted for AWS26 to 10% (8/80) over a 10-year follow-up in patients diagnosed with delirium tremens.27 Prior studies reported higher mortality and shorter intervals from hospitalization to death in men with AUD and AWS than those in women.34 In contrast, our study reported a significant age difference but no sex difference in all-cause mortality rates. Comparable mortality risk observed in women in our study may be attributable to the greater representation of women in our cohort and to emerging evidence of a narrowing of the historical sex gap in alcohol use and related outcomes. The results of this study should be viewed in the context of its strengths and limitations. Our large sample size, including a large number of women, detailed phenotyping for comorbid conditions, and hospital course, all add more validity to the results. On the contrary, the retrospective nature of the study—with no access to data on drinking patterns, the severity of AUD, psychiatric consultation during hospitalization, referral for addiction treatment, or initiating pharmacotherapy at discharge—and the fact that most patients were Whites and non-Hispanics could limit the generalizability of the results. The mortality data were ascertained through June 2023, whereas AWS data were collected from June 2019 to June 2022. As such, patients who were admitted earlier had longer observation time than those who were admitted later during the study interval. This discrepancy could have an effect on the results. Similarly, we collected mortality data recorded only in our electronic medical recotds. It is possible that our results underestimated the true posthospital mortality rate. Conclusion Despite these limitations, our study provides deep phenotyping for AWS, including among women and older adults, and suggests clinical pearls for busy clinicians. For all new admissions, males, and females, include in your routine medical history a few screening questions about typical alcohol drinking patterns, last drink, and any history of alcohol withdrawal. If you suspect recent heavy drinking or observe withdrawal manifestations, check BAC, place the patient on CIWA-Ar protocol and start IV thiamine 500 mg thrice daily for 3 to 5 days. Approximately a quarter of patients placed on the CIWA-Ar protocol get admitted to the ICU for acute decompensation in comorbid medical conditions or for severe withdrawal. Alcohol withdrawal syndrome in the ICU setting could be difficult to distinguish from hemodynamic instability, medication-associated nausea and vomiting, and ICU delirium. Differentiating ICU delirium from alcohol withdrawal delirium tremens requires a detailed alcohol history from the patient or family members on admission if possible. When drinking history is uncertain or unavailable, objective alcohol biomarkers can be considered to enhance risk stratification. Urine EtG is a sensitive marker of recent alcohol consumption and can be detected in the blood for up to 36 hours and in the urine for up to 5 days after heavy alcohol use. In contrast, PEth has a longer detection window (2-4 weeks after last use) and is most useful for identifying chronic or heavy alcohol use, which can help recognize patients at higher baseline risk for AWS. Neither biomarker directly measures physiologic dependence or predicts withdrawal severity; results should be integrated along with a history, laboratory findings, and validated withdrawal scales such as the CIWA-Ar. Keep in mind that withdrawal manifestations reach peak severity later in older adults, so they may require a longer hospital LOS despite scoring moderately high on CIWA-Ar and despite receiving lower benzodiazepine doses than younger adults. In contrast, younger patients are more likely to develop earlier and more severe withdrawal manifestations and may require higher benzodiazepine doses, particularly during the first 24 hours of hospitalization. Patients treated for AWS often experience multiple comorbid medical and psychiatric conditions, and they have high rates of posthospital mortality. Screen for comorbidities and provide outpatient follow-up appointments within a reasonable time frame and provide community resources to help them achieve and maintain sobriety. Potential Competing Interests The authors report no competing interests. Ethics Statement The study was approved by the institutional review board of Mayo Clinic (ID#22-008591). We collected the electronic medical records of all admissions at Mayo Clinic Hospitals (Rochester, Florida, and Arizona) and Mayo Clinic Health System from June 2019 to June 2022 for all patients who received an order to implement the CIWA-Ar protocol.