Comparing Outcomes and Infection Risk in Medical,Surgical,and Trauma Intensive Care Patients with Alcohol Use Disorder

Abstract

Introduction:

Alcohol is the most frequently abused drug in the United States, and alcohol use disorder (AUD) is a common comorbidity in intensive care units (ICUs).

Patients and Methods:

We performed a retrospective chart review of patients admitted to an ICU between January 2017 and March 2019 at a tertiary hospital serving a large rural population. Patients with diagnoses of AUDs were included. Patients were excluded if they did not require ICU care. Patient demographics, hospital course, infection type, culture results, and mortality were evaluated. We compared medical, surgical, and trauma ICU patient outcomes and infections.

Results:

In total, 527 patients met inclusion and exclusion criteria. Trauma ICU patients had the least pre-existing comorbidities, and surgical ICU patients had the longest lengths of stay. There was no difference in in-hospital mortality between ICU groups; however, surgical and medical ICU patients had significantly greater rates of in-hospital mortality compared with trauma ICU patients. Infections were common across all ICU types, occurring in 40% of patients. There was no difference in infection rate between ICU types. In multi-variable analysis controlling for age, gender, liver failure, chronic kidney disease, thrombocytopenia, complications, and blood transfusions, infection remained an independent predictor of in-hospital mortality (adjusted odds ratio 3.3, 95% confidence interval 1.7–6.4). Septic shock occurred in 57% of infections and was associated with an increased risk of mortality (38% vs. 2%, p < 0.001). Pneumonia was the most common infection occurring in 28% of the cohort, followed by bacteremia (7%), skin/soft tissue infections (6%), urinary tract infection (5%), intra-abdominal infections (4%), and C. difficile (2%).

Conclusions:

AUDs in all types of ICU patients are associated with high rates of infections and high morbidity and mortality.

Introduction

Alcohol is the most commonly abused drug in the United States, with more than 15 million people in the United States aged 12 and older meeting criteria for alcohol use disorder (AUD).¹ Alcohol abuse leads to numerous medical complications, including malnutrition, immunosuppression, and liver disease.^2–6 Approximately 95,000 people die from alcohol-related causes annually, making alcohol the third-leading preventable cause of death in the United States.⁷ AUD leads to poorer overall health, and when these patients require hospitalization, whether related to issues directly associated with their alcohol abuse such as liver failure or withdrawal, or indirectly related, such as traumatic injury or infection, these hospital stays are associated with high medical costs and resource utilization, and worse outcomes.^8–11

AUD increases the risk of development of a critical illness, including gastrointestinal (GI) bleeding, sepsis, and severe traumatic injury, and patients with AUD account for an undue proportion of intensive care unit (ICU) admissions.^9,12 The majority of studies evaluating outcomes in critically ill patients with AUD have been performed in large urban areas; however, epidemical data show that AUDs are greater overall in rural areas, but that these rural–urban differences vary by geographic region.¹³ The upper Midwest of the United States has a significantly greater prevalence of binge drinking compared with other regions.¹⁴ We developed this study to better understand the effects of AUD in ICU patients in a large rural population with a high prevalence of AUD. This study aims to describe the population of critically ill patients requiring intensive care at a single, tertiary referral center serving a rural population. The goal of the study was to compare outcomes and infection risk for medical, surgical, and trauma ICU patients and to identify any common risk factors associated with mortality and infectious complications in this high-risk patient population.

Materials and Methods

We performed a retrospective review of all patients admitted to the ICU at Saint Mary’s Medical Center, a Level 1 trauma center and quaternary referral center for northern Minnesota, northern Wisconsin, and the northwestern portion of the upper peninsula of Michigan. The study included all patients admitted to an ICU between January 1, 2017, and March 31, 2019. This analysis is a subgroup analysis of an initial study evaluating chemical dependency screening in ICUs.¹⁵ Patients were included if they were 18 years of age or older and had a diagnosis of alcohol abuse, alcohol dependence, alcoholism, alcoholic hepatitis, or alcoholic cirrhosis. If patients had quit drinking prior to ICU admission but had a history of alcohol abuse and had alcohol-induced liver failure or cirrhosis, they were included in the study. We excluded patients who had only social alcohol use and no diagnosis of alcohol abuse/dependence or its sequela, opted out of research at our institution, were admitted to an ICU for monitoring only and did not require ICU care, or had an ICU stay that was less than 48 h. ICU care was defined as the need for continuous intravenous medications requiring at least hourly titration by an ICU nurse, the need for mechanical ventilation, or evidence of at least one ICU-treated organ dysfunction.

Medical, surgical, and trauma ICU patients were defined initially by their admitting ICU type, and then chart review was performed to ensure that they were correctly grouped (i.e., if a patient with a primary surgical issue was admitted to the medical ICU because of bed availability, they were grouped in the “Surgical ICU” group, not the medical ICU). Manual chart review was performed for all patients, and patient demographics were recorded, including age, gender at birth, self-identified race, body mass index, marital status, and insurance type. Comorbidities were included as individual parameters based on organ system (cardiac disease, chronic respiratory disease, cancer, diabetes mellitus, chronic renal disease, mental health disorders, and thrombocytopenia) and as a comorbidity score of the total number of organ system comorbidities. Rural patients were defined by their home zip code using the Rural-Urban Commuting Area Codes.¹⁶ History of tobacco use, other drug abuse, and if patients were still using alcohol prior to admission were recorded for all patients. Hospital data, including the primary reason for admission, hospital length of stay (LOS), ICU LOS, need for mechanical ventilation, all procedures performed during the hospital stay, and any complications, were collected for all patients. The primary outcomes were in-hospital mortality and mortality within 1 year of ICU discharge. Infection was included as a risk factor for mortality analysis and evaluated as a secondary outcome to identify risk factors associated with the development of infection in this patient population.

Statistical analysis was performed using IBM SPSS version 28 (IBM Corp. Released 2021. IBM SPSS Statistics for Windows, Version 28.0. Armonk, NY: IBM Corp). Continuous variables were compared using Mann Whitney U tests, as all continuous variables were nonparametric. Categorical variables were compared using chi-square tests and odds ratios (ORs). Binary logistic regression was performed to control multi-variable interactions and calculate adjusted ORs and their 95% confidence intervals (CIs). Variables that were significant in uni-variable analysis were included in the multi-variable analysis. p Values <0.05 were considered significant. This study was reviewed, monitored, and approved by the Essentia Health Institutional Review Board.

Results

During our study period, 649 patients were admitted to an ICU with an associated AUD diagnosis (8% of all ICU admissions). We excluded 122 patients as they did not require ICU care or stayed in the ICU for <48 h, leaving 527 patients in the study cohort. Patient demographics are depicted in Table 1. The age range of patients was broad (18–86 years), and the patient cohort was predominantly male. Although the majority of patients were Caucasian, patients who identified as Native American were overrepresented in our cohort compared with the 2020 census data for our region (1.4% Native American in our region, with 17% of patients in our study identifying as Native American).¹⁷ Almost half of the patients had liver failure or cirrhosis diagnosed prior to admission, and other comorbidities were common. The most common admitting diagnoses were GI bleed (18%), cardiopulmonary failure (18%), trauma (17%), acute withdrawal (13%), sepsis (13%), complications of liver failure (9%), and suicide/overdose (7%). In-hospital mortality was 12%, and mortality within a year of ICU discharge was 27%.

Table 1.

Patient Demographics, Comorbidities, and Hospital Variables

	Entire cohort	Trauma ICU	Surgical ICU	Medical ICU
	n = 527	n = 92	n = 75	n = 360	p Value
Age; years, median (range)	56 (18–86)	55 (18–83)	60 (25–78)	55 (19–86)	0.10
Gender (female), n (%)	160 (30)	21 (23)	21 (28)	118 (33)	0.16
Race n (%)					0.06
White	418 (79)	75 (81)	65 (88)	278 (77)
Black	11 (2)	2 (2)	34 (5)	5 (1)
Native American	88 (17)	13 (14)	6 (8)	69 (19)
Unknown	10 (2)	2 (2)	0 (0)	8 (2)
Body mass index, median (IQR)	27.1 (23.2–32.0)	26.3 (23.0–30.3)	27.7 (22.6–32.0)	27.1 (23.2–32.1)	0.68
Rural, n (%)	189 (36)	37 (40)	31 (42)	121 (34)	0.28
Insurance, n (%)					0.02
Uninsured	115 (22)	17 (19)	11 (15)	87 (24)
Private	71 (14)	18 (120)	12 (16)	41 (11)
Medicare	156 (30)	27 (29)	32 (43)	97 (27)
Medicaid	185 (35)	30 (33)	20 (27)	135 (38)
Comorbidities
Liver failure, n (%)	241 (46)	21 (23)	31 (41)	189 (53)	<0.001
Cardiac disease, n (%)	331 (59)	445 (48)	57 (76)	210 (58)	0.001
Pulmonary disease, n (%)	154 (29)	18 (19)	25 (34)	111 (31)	0.06
Cancer, n (%)	52 (10)	1 (1)	12 (16)	39 (11)	0.003
Diabetes mellitus, n (%)	111 (21)	16 (17)	18 (24)	77 (22)	0.51
Chronic renal disease, n (%)	82 (16)	7 (8)	22 (30)	53 (15)	<0.001
Thrombocytopenia, n (%)	194 (37)	20 (22)	24 (32)	150 (42)	0.001
Mental health disorder, n (%)	307 (58)	46 (50)	41 (55)	220 (61)	0.11
Total number of different organ system comorbidities, median (IQR)	3 (2–6)	2 (1–3)	4 (2–5)	3 (2–4)	<0.001
Meld score, median (IQR)	20 (13–26)	14 (10–24)	20 (14–26)	20 (13–27)	0.29
Two or more comorbidities, n (%)	315 (60)	30 (32)	51 (69)	234 (65)	<0.001
ICU length of stay, median days (IQR)	3 (2–6)	2 (2–6)	6 (3–10)	3 (2–5)	<0.001
Hospital length of stay, median days (IQR)	8 (4–13)	8 (4–14)	13 (9–19)	7 (4–11)	<0.001
Respiratory failure requiring mechanical ventilation, n (%)	263 (50)	56 (60)	55 (74)	152 (42)	<0.001
Still drinking prior to admission, n (%)	463 (88)	88 (95)	62 (84)	313 (87)	0.07
Infection diagnosed any time during admission, n (%)	211 (40)	22 (24)	46 (62)	143 (40)	<0.001
Nosocomial infection, n (%)	163 (31)	22 (24)	31 (41)	11 (31)	0.04
Bacteremia, n (%)	36 (7)	1 (10)	11 (15)	24 (7)	0.002
C. difficile, n (%)	10 (2)	1 (10)	1 (1)	8 (2.2)	0.72
Pneumonia, n (%)	148 (28)	18 (19)	22 (30)	108 (30)	0.12
UTI, n (%)	28 (5)	4 (4)	4 (5)	20 (6)	0.89
Skin and soft tissue infection, n (%)	30 (6)	3 (3)	10 (14)	17 (5)	0.006
Septic shock, n (%)	125 (24)	12 (13)	33 (45)	80 (22)	<0.001
In-hospital complication, n (%)	428 (81)	61 (66)	66 (89)	301 (84)	<0.001
Cardiac complication	86 (16)	5 (5)	23 (31)	58 (16)	<0.001
Acute renal failure	99 (19)	8 (9)	29 (38)	63 (17)	<0.001
ARDS	51 (10)	4 (4)	13 (18)	34 (9)	0.02
Bleeding complication	162 (30)	33 (36)	19 (26)	110 (31)	0.42
Thrombotic complication	36 (7)	6 (7)	5 (7)	25 (7)	0.98
Received blood transfusion	72 (23)	13 (14)	18 (24)	84 (23)	0.13
Alcohol withdrawal during admission	272 (52)	43 (46)	35 (47)	194 (54)	0.31
Readmitted in next 6 months with ETOH/liver complications, n (%)	237 (51)	27 (32)	33 (52)	177 (56)	<0.001
In-hospital mortality, n (%)	63 (12)	8 (9)	10 (14)	45 (13)	0.53
1-year mortality, n (%)	144 (27)	14 (15)	19 (26)	111 (31)	0.009
Discharged home, n (%)	267 (51)	40 (43)	32 (43)	195 (55)	0.06

ARDS = acute respiratory distress syndrome; ICU = intensive care unit; IQR = interquartile range.

Comparing medical, surgical, and trauma ICU patients

We compared patients with AUD by their ICU admission type (medical, surgical, and trauma ICU admissions). The different ICU groups were similar in age, rurality, gender, and race (Table 1). Twenty-two percent of patients were uninsured, and medical ICU patients were the most likely to be uninsured. There were high rates of chronic comorbidities in all ICU groups, but liver failure was highest in the medical ICU group, and cardiac disease, cancer, chronic renal disease, and thrombocytopenia were highest in the surgical ICU group. The surgical ICU group had a significantly greater number of total pre-existing comorbidities than the medical or trauma ICU groups. More than half of the cohort had a diagnosed mental health disorder, in addition to AUD. Hospital LOS was long, with a median of 8 days overall, and again surgical ICU patients had significantly longer hospital and ICU stays than the other ICU groups. Medical ICU patients had a significantly lower need for mechanical ventilation. Although the vast majority of patients had complications in the hospital (81%), trauma ICU patients had the lowest complication rates. Bleeding complications were the most common complication and similar between ICU groups. Acute renal failure, cardiac complications, and acute respiratory distress syndrome (ARDS) occurred most frequently in surgical ICU patients. Approximately half of all patients had alcohol withdrawal during their hospital stay, and this was similar between ICU groups. Although in-hospital mortality rates did not differ significantly between ICU groups, patients admitted to surgical and medical ICUs had significantly greater 1-year post-discharge mortality rates compared with trauma ICU patients. Medical and surgical ICU patients were also more likely to be readmitted within 6 months of their initial ICU admission for alcohol-related problems.

Operative intervention

Forty-three percent of patients in the cohort underwent an invasive procedure. The most common procedure type was endoscopic (19%), followed by operative procedure (15%) and interventional radiology (IR) procedure (14%). No patients in the medical ICU group underwent an operative procedure, but 65% of surgical ICU and 35% of trauma ICU patients had an operative procedure. Only 7% of the surgical ICU patients were admitted to the ICU after an elective procedure. Trauma ICU patients had a significantly increased risk of morbidity and mortality after operative procedures (Table 2), but there was no difference in the surgical ICU group. IR procedures were associated with a significantly increased OR for in-hospital mortality in our cohort. This was true for all ICU types. The most common IR procedures were paracentesis (51%) and embolization of bleeding (24%). Transjugular intrahepatic portosystemic shunt was performed in 15% of IR procedures.

Table 2.

Procedures Performed by ICU Admission Group and Their Associated Odds Ratio (OR) for In-Hospital Mortality

	All patientsn = 527	Trauma ICUn = 92	Surgical ICUn = 75	Medical ICUn = 360	p Value
Any procedure, n (%)	224 (43)	36 (39)	55 (74)	113 (37)	<0.001
OR for in-hospital mortality with operative procedure (95% CI)	1.0 (0.6–1.7)	1.7 (0.8–7.1)	1.4 (0.3–7.5)	0.8 (0.4–1.6)
OR for complication with operative procedure (95% CI)	1.3 (0.8–2.2)	4.1 (1.5–11.6)	0.3 (0.1–3.3)	0.9 (0.5–1.6)
Operative procedure, n (%)	81 (15)	32 (35)	49 (65)	0 (0)	<0.001
OR for in-hospital mortality with operative procedure (95% CI)	0.9 (0.4–1.9)	1.7 (1.01–4.1)	0.7 (0.2–2.9)	NA (none underwent)
OR for complication with operative procedure (95% CI)	1.1 (0.6–2.1)	3.2 (1.2–8.9)	1.6 (1.3–1.9)	NA (none underwent)
Endoscopic procedure, n (%)	100 (19)	1 (1)	11 (15)	88 (24)	<0.001
OR for in-hospital mortality with endoscopic procedure(95% CI)	1.0 (5.1–2.0)	^a	4.5 (0.9–23)	0.7 (0.3–1.3)
OR for complication with endoscopic procedure (95% CI)	1.0 (0.6–1.7)	^a	1.2 (0.9–1.4)	0.7 (0.3–1.3)
Interventional radiology procedure, n (%)	71 (14)	7 (8)	11 (15)	53 (15)	0.19
OR for in-hospital mortality with interventional radiology (95% CI)	3.1 (1.7–5.8)	5.3 (1.1–33.5)	5.4 (1.2–24)	2.4 (1.2–5.1)
OR for complication with interventional radiology (95% CI)	2.3 (0.9–5.2)	1.1 (0.9–1.2)	1.2 (0.9–1.4)	1.3 (0.6–3.1)

only one patient in group, odds ratio not calculated. Forty-three percent of patients underwent one or more invasive procedures. Surgical ICU patients were the most likely to undergo any invasive procedure and the most likely to undergo operative intervention. Trauma ICU patients who underwent an operative procedure had a significantly increased OR for in-hospital mortality and morbidities compared with those who did not undergo an operative procedure. There was no significant increased risk of mortality or morbidity in surgical ICU patients. Interventional radiology procedures were associated with an increased odds of mortality and morbidity in all groups. Endoscopic procedures were not associated with increased mortality or morbidity.

95% CI = 95% confidence interval; ICU = intensive care unit; OR = odds ratio.

When we only looked at patients with severe liver disease (Child’s-Pugh Score B or C, n = 212 patients in the cohort), there was still no significant difference in in-hospital mortality rate associated with invasive procedures. Overall mortality rate was 25% for all patients with Child B and C cirrhosis, and there was no difference in mortality if they underwent procedures (20% mortality for those who underwent any procedure [p = 0.10], 27% for operative procedures [p = 0.75], and 31% for IR procedures [p = 0.18]). Mortality rate was lower for patients who underwent GI procedures (16%, p = 0.03).

Infections

Infections were present during the hospital stay in 40% of patients overall. The admitting diagnosis requiring ICU care was sepsis in 13%, 25%, and 1% of medical, surgical, and trauma ICU patients, respectively (p < 0.001). Thirty-three percent of patients developed a nosocomial infection (an infection diagnosed during the hospital stay that was not present on admission) during their hospital stay. Nosocomial infections occurred in 32% of patients not admitted with an infection and 45% of those admitted with an infection (p = 0.03). Surgical ICU patients had the highest rate of nosocomial infections (46%) compared with medical ICU (34%) and trauma ICU patients (23%), p = 0.008. Infection overall in this cohort was associated with a significantly increased mortality rate compared with patients who did not have an infection while in the hospital (23% vs. 5%, p > 0.001). In-hospital mortality rates for those with an infection were not significantly different between ICU types (25% for medical ICU, 17% for surgical ICU, and 14% for trauma ICU, p = 0.27).

Patient and hospital factors associated with infections in our cohort of ICU patients with AUD are described in Table 3. Patients with and without infections were similar in age, gender, race, and primary insurance type. However, infection rates were greater in rural patients. Surgical ICU patients had a significantly greater rate of infection than medical or trauma ICU patients. Comorbid conditions present at admission, including liver failure, chronic renal disease, and thrombocytopenia, were associated with an increased risk of infection. Interestingly, we did not find increased rates of infection in patients with diabetes mellitus. Patients with infections had significantly longer ICU and hospital stays, and infections were associated with a significantly greater risk of death during the hospital stay and in the year after ICU admission.

Table 3.

Patient Demographics, Comorbidities, and Hospital Variables Associated with Infections in Critically Ill Patients with Alcohol Use Disorder

	Entire cohort	Patients with infection			Patients with pneumonia
	n = 527	n = 211	Odds ratio(95% CI)	p Value	n = 148	Odds ratio(95% CI)	p Value
Age; years, median (range)	56 (18–86)	57 (20–86)	—	0.26^a	57 (20–78)	—	0.04
Gender (female), n (%)	160 (30)	68 (32)	1.2 (0.8–1.7)	0.45	45 (30)	1.0 (0.6–1.5)	0.99
Race, n (%)			N	0.3		—	0.66
White	418 (79)	160 (76)			115 (78)
Black	11 (2)	4 (2)			2 (1.4)
Native American	88 (17)	43 (20)			27 (18)
Asian	6 (1)	2 (1)			2 (1)
Unknown	4 (1)	2 (1)			2 (1)
Rural, n (%)	189 (36)	88 (42)	1.5 (1.1–2.1)	0.02	65 (44)	1.6 (1.1–2.4)	0.02
Insurance, n (%)			—	0.22		—	0.35
Uninsured	115 (22)	41 (19)			35 (24)
Private	71 (14)	27 (13)			16 (11)
Medicare	156 (30)	73 (35)			50 (34)
Medicaid	185 (35)	70 (33)			47 (32)
BMI >30	94 (30)	76 (36)	1.4 (0.9–2.0)	0.11	52 (31)	1.22 (0.8–1.8)	0.33
Comorbidities
Liver failure/cirrhosis, n (%)	241 (46)	116 (55)	1.9 (1.3–2.7)	<0.001	76 (51)	1.4 (0.9–2.0)	0.11
Cardiac disease, n (%)	331 (59)	134 (64)	1.4 (0.9–2.0)	0.09	90 (61)	1.1 (0.8–1.6)	0.60
Pulmonary disease, n (%)	154 (29)	78 (37)	1.9 (1.3–2.7)	0.001	61 (41)	2.2 (1.4–3.2)	<0.001
Cancer, n (%)	52 (10)	30 (14)	2.2 (1.2–4.0)	0.006	20 (14)	1.7 (0.9–3.1)	0.08
Diabetes mellitus, n (%)	111 (21)	48 (27)	1.1 (0.7–1.8)	0.44	30 (20)	0.9 (0.6–1.5)	0.78
Chronic renal disease, n (%)	85 (15)	48 (23)	2.4 (1.5–3.9)	<0.001	27 (18)	1.3 (0.8–2.2)	0.39
Thrombocytopenia, n (%)	194 (37)	94 (45)	1.7 (1.2–2.5)	0.003	65 (44)	1.5 (1.03–2.2)	0.04
Mental health disorder, n (%)	307 (58)	126 (60)	1.1 (0.8–1.6)	0.58	86 (58)	1.0 (0.7–1.5)	0.97
Two or more comorbidities, n (%)	244 (77)	180 (86)	1.8 (1.1–2.9)	0.01	124 (29)	1.5 (0.9–2.5)	0.13
ICU type, n (%			—	<0.001			0.12
Medical ICU	360 (68)	143 (40)			108 (30)
Surgical ICU	74 (14)	46 (62)			22 (29)
Trauma ICU	93 (18)	22 (24)			18 (20)
ICU length of stay, median days (IQR)	3 (2–6)	5 (3–10)	—	<0.001	6 (3–10)	—	<0.001
Hospital length of stay; median days (IQR)	8 (4–13)	10 (6–18)	—	<0.001	10 (6–18)	—	<0.001
Respiratory failure requiring mechanical ventilation, n (%)	263 (50)	154 (73)	5.1 (3.5–7.5)	<0.001	123 (83)	8.4 (5.2–13.6)	<0.001
Alcohol withdrawal during hospital stay, n (%)	272 (52)	116 (55)	1.3 (0.9–1.8)	0.21	96 (65)	2.1 (1.4–3.2)	<0.001
Received blood transfusion, n (%)	72 (23)	43 (20)	0.9 (0.6–1.3)	0.51	30 (20)	0.9 (0.6–1.4)	0.51
In-hospital mortality, n (%)	63 (12)	48 (23)	5.9 (3.1–10.8)	<0.001	33 (22)	3.3 (2.0–5.7)	<0.001
1-year mortality, n (%)	144 (27)	89 (40)	2.6 (1.8–3.9)	<0.001	57 (39)	2.1 (1.4–3.2)	<0.001

ICU = intensive care unit; IQR = interquartile range.

Pneumonia was the most common infection occurring in 28% of the cohort, followed by bacteremia (7%), skin/soft tissue infections (6%), urinary tract infections (5%), intra-abdominal infections (4%), and C. difficile (2%). Not surprisingly, surgical ICU patients were most likely to have intra-abdominal infections and were also the most likely to have bacteremia and skin and soft tissue infections (Table 1). Infection was associated with a significantly increased in-hospital and 1-year post-discharge OR for mortality: OR of 5.9 for in-hospital mortality and 2.6 for 1-year mortality (Table 3). As pneumonia was by far the most common infection identified, further analysis of this subgroup was performed. Pneumonia was associated with older age, longer lengths of stay in the hospital and ICU, and a significantly increased mortality rate (Table 2). There was no difference in risk of pneumonia associated with gender, insurance status, race, or ICU type. Rural patients were significantly more likely than urban patients to develop pneumonia. We found that patients with pre-existing chronic respiratory disease had greater rates of pneumonia (OR 2.2; 95% CI 1.4–3.2), as did patients with thrombocytopenia (OR 1.5; 95% CI 1.03–2.2). Other comorbidities were not associated with an increased likelihood of pneumonia diagnosis. Alcohol withdrawal during their ICU admission was associated with an increased risk of pneumonia (OR 2.1; 95% CI 1.4–3.2). Patients requiring mechanical ventilation were at highest risk for development of pneumonia (OR 8.4; 95% CI 5.2–13.5). In multi-variable analysis using binary logistic regression, we found that the need for mechanical ventilation in the ICU, alcohol withdrawal in the ICU, increasing age, and thrombocytopenia were independently associated with pneumonia diagnosed during their hospital stay. Cultures were obtained in 74% of pneumonias, where the most common isolated organisms were S. aureus (17%; 11% Methicillin sensitive S. aureus and 6% Methicillin resistant S. aureus), H. influenzae (13%), and S. pneumoniae (12%). Enteric pathogens were isolated in 28% of cultures.

Mortality risk factors for all ICU patients

As ICU type was not associated with differences in in-hospital mortality, we included all ICU types in an analysis of factors that were predictors of mortality in our cohort of critically ill patients with AUD. ICU type was associated with differences in 1-year mortality, with surgical ICU and medical ICU patients having significantly greater post-discharge mortality rates than trauma ICU patients. Parameters that were associated with increased in-hospital and 1-year mortality were varied, including a significantly increased risk of mortality with older age, rural patients, pre-existing liver failure, and pre-existing comorbidities (Table 4). In addition, the diagnosis of an infection while in the hospital was associated with a significant increase in in-hospital and 1-year mortality. Septic shock was diagnosed in 57% of patients with infections and was associated with an increased risk of in-hospital mortality (38% vs. 2% p < 0.001). ARDS was diagnosed in 51 patients (10% of the cohort) and was associated with a significantly increased risk of death in the hospital (OR 11.1, 95% CI 5.8–21.0), with 49% of patients with ARDS dying in the hospital, compared with 8% without ARDS. Of the 63 patients who died in the hospital, just over a third (37%) died after transitioning to comfort-focused care. In multi-variable analysis using binary logistic regression, controlling for age, gender, number of comorbidities, need for mechanical ventilation, and in-hospital complications, infection remained an independent predictor of in-hospital mortality in this critically ill patient population with AUD (Fig. 1A depicts all variables included in the model). In multi-variable analysis controlling for the same variables, 1-year mortality was significantly associated with increasing age, rurality, in-hospital complications, and increased number of chronic comorbidities (Fig. 1B depicts all variables included in the model). To assess for collinearity in the variables included in our regression model, a Pearson’s correlation coefficient matrix was created, and none of the values had a correlation coefficient of (absolute value) 0.6 or greater.

FIG. 1.

Forest plots depicting adjusted ORs for mortality in critically ill patients with alcohol used disorder. (A) Adjusted ORs for n-hospital mortality. (B) Adjusted ORs for 1-year mortality. OR = odds ratio; 95% CI = 95% confidence interval.

Table 4.

Patient Demographics, Comorbidities, and Hospital Variables Associated with Mortality in Critically Ill Patients with Alcohol Use Disorder

	Entire cohort	In-hospital mortality			1-year mortality
	n = 527	n = 63	Odds ratio(95% CI)	p Value	n = 144		Odds ratio(95% CI)	p Value
Age; years median (range)	56 (18–86)	59 (50–67)	—	0.008	59 (52–67)	—		<0.001
Gender (female), n (%)	160 (30)	23 (37)	1.4 (0.8–2.4)	0.26	53 (37)	1.5 (1.0–2.3)		0.05
Race, n (%)			—	0.65	—	—		0.81
White	418 (79)	52 (83)			114 (79)
Black	11 (2)	0 (0)			2 (1)
Native American	88 (17)	10 (16)			26 (18)
Asian	6 (1)	1 (1.6)			2 (1.4)
Unknown	4 (1)	0 (0)			0 (0)
Rural, n (%)	189 (36)	30 (48)	1.7 (1.1–3.0)	0.04	65 (47)	1.9 (1.3–2.8)		0.002
Insurance, n (%)			—	0.02		—		<0.001
Uninsured	115 (22)	10 (16)			35 (24)
Private	71 (14)	5 (8)			11 (8)
Medicare	156 (30)	29 (46)			59 (41)
Medicaid	185 (35)	19 (30)			39 (27)
ICU type			—	0.53		—		0.009
Medical ICU	360 (68)	45 (13)			111 (31)
Surgical ICU	74 (14)	10 (14)			19 (26)
Trauma ICU	93 (18)	8 (9)			14 (15)
Comorbidities
Liver failure, n (%)	241 (46)	55 (87)	10.3 (4.8–22.1)	<0.001	114 (80)	7.7 (4.9–12.1)		<0.001
Cardiac disease, n (%)	331 (59)	48 (76)	2.4 (1.3–4.5)	0.003	100 (69)	1.9 (1.2–2.8)		0.003
Pulmonary disease, n (%)	154 (29)	29 (46)	2.3 (1.4–4.0)	0.002	60 (42)	2.2 (1.5–3.3)		<0.001
Cancer, n (%)	52 (10)	12 (19)	2.5 (1.2–5.1)	0.009	27 (19)	3.3 (1.9–5.9)		<0.001
Diabetes mellitus, n (%)	111 (21)	17 (27)	1.5 (0.8–2.7)	0.22	40 (28)	1.7 (1.1–2.6)		0.02
Chronic renal disease, n (%)	85 (15)	20 (32)	3.0 (1.7–5.5)	<0.001	37 (26)	2.6 (1.6–4.2)		<0.001
Thrombocytopenia, n (%)	194 (37)	42 (67)	4.1 (2.3–7.2)	<0.001	77 (54)	2.6 (1.8–3.9)		<0.001
Mental health disorder, n (%)	307 (58)	42 (67)	1.5 (0.9–2.6)	0.15	83 (57)	0.9 (0.7–1.4)		0.86
Two or more comorbidities, n (%)	244 (77)	63 (100)	1.2 (1.1–1.2)	<0.001	136 (94)	5.5 (2.6–11.6)		<0.001
ICU length of stay, median days (IQR)	3 (2–6)	6 (3–10)	—	<0.001	3 (2–5)	—		0.02
Hospital length of stay, Median days (IQR)	8 (4–13)	8 (3–13)	—	0.43	8 (4–12)	—		0.30
Respiratory failure requiring mechanical ventilation, n (%)	263 (50)	56 (89)	9.9 (4.4–22.3)	<0.001	84 (58)	0.6 (1.1–2.4)		0.02
Alcohol withdrawal during admission, n (%)	272 (52)	23 (37)	0.5 (0.3–0.9)	0.01	56 (39)	0.5 (0.3–0.7)		<0.001
Infection, n (%)	211 (40)	48 (76)	2.6 (1.8–3.9)	<0.001	82 (57)	2.6 (1.8–3.9)		<0.001
Septic shock, n (%)	125 (24)	47 (75)	14.5 (7.8–26.9)	<0.001	63 (48)	5.4 (3.5–8.3)		<0.001
In-hospital complication, n (%)	428 (81)	62 (98)	16.6 (2.3–121)	<0.001	134 (93)	4.1 (2.1–8.1)		<0.001
Cardiac complication	86 (16)	31 (49)	7.2 (4.1–12.7)	<0.001	45 (31)	3.8 (2.3–6.1)		<0.001
Acute renal failure	99 (19)	34 (54)	7.2 (4.1–12.6)	<0.001	54 (38)	4.5 (2.9–7.1)		<0.001
ARDS	51 (10)	25 (40)	11.1 (5.8–21.1)	<0.001	35 (24)	7.4 (3.9–13.8)		<0.001
Bleeding complication	162 (30)	23 (37)	1.3 (0.8–2.3)	0.31	53 (37)	1.5 (0.9–2.2)		0.6
Thrombotic complication	36 (7)	7 (11)	1.9 (0.8–4.5)	0.15	15 (11)	2.0 (1.0–4.0)		0.04
Received blood transfusion	72 (23)	20 ()	1.8 (1.01–3.2)	0.04	43 (30)	1.8 (1.2–2.9)		0.006

ARDS = acute respiratory distress syndrome; ICU = intensive care unit; IQR = interquartile range.

Discussion

We present findings from a large cohort of patients with AUD requiring ICU care at a tertiary hospital serving a broad rural population. ICU admissions associated with alcohol abuse accounted for 8% of our ICU admissions. This is similar to studies performed in the United States, United Kingdom, Finland, Denmark, and Australia, which have reported rates of ICU admissions associated with alcohol abuse between 7% and 25%.^12,18–20 However, because our study excluded patients admitted to an ICU who did not require intensive care interventions, the prevalence of alcohol abuse as related to overall ICU utilization may be greater than presented in this article. In our cohort, we had long ICU and hospital lengths of stay and high resource utilization, including high rates of mechanical ventilation. Many of these patients did not have health insurance (22%) or were on government-sponsored insurance (Medicare or Medicaid; 65%). This points to a significant cost burden to hospital systems and highlights the importance of appropriate AUD treatment, patient education, and prevention.

We found that patients with AUD admitted to medical, surgical, and trauma ICUs were overall similar. Surgical ICU patients had the longest lengths of stay and the highest rates of respiratory failure requiring mechanical ventilation. Although there were some differences in patient demographics and hospital factors between ICU types, the groups were quite similar, and in-hospital mortality rates were not statistically different. However, complication rates and 1-year mortality were significantly different between ICU groups, with trauma ICU patients having the lowest complication rates and lowest 1-year mortality rate. Although the cause for this is not known, it could be because of the lower rates of comorbidities and liver failure in the trauma group. It is possible that traumatic ICU admissions associated with AUD occur in the earlier stages of AUD and may be an opportunity for early intervention. In a previous analysis of this cohort, we found that provider-directed chemical dependency screening was associated with significantly decreased 1-year mortality and readmissions.¹⁵ Although our hospital has nursing protocols that are the same across all ICUs (for ventilator weaning, central line care and removal, antibiotic agent stewardship, etc.), we cannot be sure that there are not differences in care based on intensivist training (medical vs. surgical) or individual providers. Trauma and surgical ICU patients are cared for by eight double-boarded trauma critical care/general surgical physicians, and the medical ICU is staffed by pulmonary critical care and anesthesia critical care providers. However, that analysis was beyond the scope of this study. Additionally, it should be noted that this study may be underpowered to identify a difference, as in-hospital mortality rates ranged from 14% to 25%; larger studies may identify a more significant relationship.

In this study of critically ill patients, we found that trauma ICU patients had a significantly increased risk of morbidity and mortality; when undergoing operative intervention, this increased risk was not identified in our other ICU groups. Although 65% of the surgical ICU patients underwent an operative procedure, we did not find an increased risk of death or complication related to operative interventions in this group. Previous studies have reported an increased risk of morbidity in patients with AUD and have hypothesized that this increased risk is directly because of the alcohol use and differences in host response to stress.^21,22 It is possible that our study design biased our results; by including only critically ill patients with AUD, the critical illness and AUD may have been the primary drivers of mortality in this study, not the operative procedure performed. Additionally, as only 7% of the surgical ICU patients underwent an elective procedure, the remaining procedures were all emergent and most were done to perform source control for an infection or bleeding as the treatment intervention for their critical illness. In contrast, in the trauma ICU group, most patients underwent orthopedic procedures, and the operative stress may have added to their critical illness. It is also possible that too few of the patients underwent a surgical procedure to produce an effect, or that our providers are not offering operative procedures to these patients because of their high risk. Interestingly, IR procedures were found to have a significant effect on mortality. This is likely because of a combination of factors. Paracentesis was the most common IR procedure performed and correlates with worsening liver disease. Ascites has been reported to be a predictor of mortality even in patients with lower MELD scores,²³ and we hypothesize that those patients with enough ascites to require paracentesis while critically ill may be in that greater risk group.

The patients in our study had a high risk of in-hospital death (12%), with that risk of mortality extending a year after ICU discharge. Not surprisingly, we found that liver failure, respiratory failure requiring mechanical ventilation, complications, and septic shock were associated with the highest risk of in-hospital death. Many other studies in patients with and without AUD have reported similar risk factors for death.^11,24–26 In our study, the most common complications besides infection, which will be discussed subsequently, were bleeding, thrombosis, ARDS, renal failure, and cardiac complications. The usual ICU quality measures of appropriate ventilator bundles, venous thromboembolism prevention, catheter removal, goal-directed sepsis treatment, and early mobility to prevent these complications are imperative in all critically ill patients, but our findings highlight the increased risk to patients with AUD. These factors remained associated with an increased risk of death in the year following discharge, likely pointing to their association with worse overall health of those patients. Patients diagnosed with two or more chronic comorbidities had an increased risk of death in the hospital and 1-year mortality, and in multi-variable analysis increasing number of pre-existing chronic comorbidities was significantly associated with an increased rate of 1-year mortality. This finding of increased risk of death with increasing chronic comorbidities is echoed in numerous other studies, both in patients with alcohol abuse^27,28 and without.^29,30 Appropriate recognition and management of pre-existing comorbidities in patients with AUD are imperative to reduce mortality.

Patients with AUD and critical illness had high rates of infection during their hospital stay. Having an infection in the hospital was associated with six-fold increased odds of death. This risk of death remained high in the year following admission. Our infection rate of 40% is similar to other reports of ICU infection rates.^25,31,32 We found that surgical ICU patients had the highest infection rate. In our cohort, patients with septic shock had a 38% in-hospital mortality rate. This rate is similar to a recent meta-analysis evaluating current mortality rates from septic shock in ICU patients.³² This meta-analysis included 71 publications from Europe and North America and found that the mean in-hospital mortality rate was 37.3%.³⁰ We would have initially expected the rate of infection and mortality from sepsis to be greater in our cohort than the general population given their risk factor of AUD, but it is possible that AUD conveys similar risk to other critical illness. This would be something that should be investigated further.

Pneumonia was by far the most common infection diagnosed in our cohort. We hypothesize this may be because of the increased aspiration risk associated with alcohol abuse, as well as the high rate of mechanical ventilation in our patient population. The organisms isolated in our respiratory cultures are similar to other hospital-associated pneumonias, with the most common isolates being S. pneumonia, S. aureus, and H. influenza; however, we did identify many patients that had enteric bacteria as the isolated organism in their respiratory cultures. De Roux et al. also found an increased prevalence of S. pneumoniae isolated in patients with alcohol misuse with community-acquired pneumonia than those without alcohol misuse, and that these infections were more severe, although mortality rates did not differ.³³ Alcohol abuse has been associated with an increased risk of pneumonia and acute lung injury. Two separate meta-analyses have found an increased risk of community-acquired pneumonia in patients with alcohol consumption, and both also identified a dose-related response, with increasing risk with increasing consumption.^34,35 Numerous clinical and experimental evidences have pointed to several pathways in which alcohol may affect the respiratory and immune systems that may account for the increased risk in patients with alcohol abuse. These include decreased innate immunity because of decreased production of bactericidal substances such as complement or lysozyme,³⁶ decreased alveolar macrophage activity,³⁷ and decreased ability to recruit polymorphonuclear leukocytes because of decreased production and response to chemotactic signals.^36,38 These studies and our results highlight the need to be vigilant in patients with AUD who are at greater risk for aspiration, may present with altered mental status and increased sedation, and with less immune function because of alcohol abuse to fight off infection.

This study is limited by its retrospective design, carried out at a single institution, which may lead to bias and cannot define causality. In addition, this study was not planned as a comparative study, and without a control group of patients without alcohol abuse, we were unable to evaluate if alcohol abuse or the critical illness was the driving force for the high risks of mortality and morbidity in this cohort. We hope to further study this in the future.

Conclusion

In conclusion, AUD in a cohort of critically ill patients was associated with high rates of infections and a high risk of death. Surgical ICU patients were at the greatest risk of infection, and although there was no significant difference in in-hospital mortality, medical and surgical ICU patients had significantly greater rates of 1-year mortality. Although we were not able to find a definite reason for this difference in 1-year mortality, it does appear that our trauma ICU patients were less likely to have pre-existing liver failure and had fewer comorbidities, leading us to hypothesize that trauma ICU admissions may be one of the earlier ICU presentations for this high-risk population and interventions to decrease alcohol use and address comorbidities may be able to improve outcomes. We found that infections in this critically ill cohort with AUD lead to significantly longer lengths of stay, high ICU resource utilization, and significant morbidity and mortality. We believe that this study highlights the importance of recognizing AUD in all patients and the need for ongoing patient counseling and treatment of AUD. AUD is a modifiable risk factor, and clinicians should be aware of the high risks associated with it. Each interaction with medical providers should be used to support patients with AUD as an opportunity for support and intervention. With the high 1-year mortality, this risk is ongoing after the critical illness resolves; these patients should have close follow-up and support after discharge.

Footnotes

Authors’ Contributions

Colling was involved in all parts of this work, from the conceptualization, methodology, formal analysis, curating data, writing the article, and preparing the visualization results. Kraft was involved in the data curation, writing the abstract, presented the abstract at the Surgical Infection Society, helped to prepare visual result, and participated in article revision. Harry was involved in the conceptualization of the study, methodology, and article revision.

Author Disclosure Statement

The authors have no conflicts of interest to disclose.

Funding Information

No funding was received for this article.

References

Richesson

, Hoenig

. Key Substance Use and Mental Health Indicators in the United States: Results from the 2020 National Survey on Drug Use and Health. (HHS Publication No. PEP21-07-01-003, NSDUH Series H-56). Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration: Rockville, MD. Available from: https://www.samhsa.gov/data/ [Last accessed: February 15, 2022].

Kimball

, Lang

. Mechanisms underlying muscle protein imbalance induced by alcohol. Annu Rev Nutr, 2018; 38:197–217; doi: 10.1146/annurev-nutr-071816-064642

Lieber

. ALCOHOL: Its metabolism and interaction with nutrients. Annu Rev Nutr, 2000; 20:395–430; doi: 10.1146/annurev.nutr.20.1.395

Osna

, Donohue

, Kharbanda

. Alcoholic liver disease: Pathogenesis and current management. Alcohol Res, 2017; 38(2):147–161.

Bishehsari

, Magno

, Swanson

, et al. Alcohol and gut-derived inflammation. Alcohol Res, 2017; 38(2):163–171.

Chiappelli

, Kung

, Stefanini

, et al. Alcohol and immune function. In: Nutrition and Immunology.( Gershwin

M.E

, German

J.B

, Keen

C.L

eds) Humana Press: Totowa, NJ; 2000; doi: 10.1007/978-1-59259-709-3_21

National Institute on Alcohol Abuse and Alcoholism. (NIAAA). Alcohol Facts and Statistics. National Institute on Alcohol Abuse and Alcoholism (NIAAA); 2019 [cited 2023 Feb 18]. Available from: https://www.niaaa.nih.gov/publications/brochures-and-fact-sheets/alcohol-facts-and-statistics

Peterson

, Li

, Xu

, et al. Assessment of annual cost of substance use disorder in US hospitals. JAMA Netw Open, 2021; 4(3):e210242; doi: 10.1001/jamanetworkopen.2021.0242

Cervellione

, Shah

, Patel

, et al. Alcohol and drug abuse resource utilization in the ICU. Subst Abuse, 2019; 13:1178221819869327; doi: 10.1177/1178221819869327

10.

Secombe

, Stewart

. The impact of alcohol-related admissions on resource use in critically Ill patients from 2009 to 2015: An observational study. Anaesth Intensive Care, 2018; 46(1):58–66; doi: 10.1177/0310057X1804600109

11.

Lone

, Lee

, Walsh

. Long term mortality and hospital resource use in ICU patients with alcohol-related liver disease. Crit Care Med, 2019; 47(1):23–32; doi: 10.1097/CCM.0000000000003421

12.

Uusaro

, Parviainen

, Tenhunen

, et al. The proportion of intensive care unit admissions related to alcohol use: A prospective cohort study. Acta Anaesthesiol Scand, 2005; 49(9):1236–1240; doi: 10.1111/j.1399-6576.2005.00839.x

13.

Borders

, Booth

. Rural, suburban, and urban variations in alcohol consumption in the united states: Findings from the national epidemiologic survey on alcohol and related conditions. J Rural Health, 2007; 23(4):314–321; doi: 10.1111/j.1748-0361.2007.00109.x

14.

CDC. Data on Excessive Drinking. CDC. 2022 [cited 2022 Nov 18]. Available from: https://www.cdc.gov/alcohol/data-stats.htm

15.

Colling

, Kraft

, Harry

. Alcohol use disorder in the intensive care unit a highly morbid condition, but chemical dependency discussion improves outcomes. Acute Crit Care, 2023; 38(1):122–133; doi: 10.4266/acc.2022.00584

16.

United States Department of Agriculture Economic Research Service. Documentation - 2010 Rural-Urban Commuting Area (RUCA) Codes. 2019. Available from: https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes/documentation/

17.

U.S. Census Bureau QuickFacts: Minnesota. [cited 2023 Apr 26]. Available from: https://www.census.gov/quickfacts/MN

18.

Lynch

, Pugh

, Battle

, Welsh Intensive Care Society Audit and Research Group (WICSARG). A multicentre prospective evaluation of alcohol-related admissions to intensive care units in Wales. J Intensive Care Soc, 2017; 18(3):193–197; doi: 10.1177/1751143717698977

19.

Geary

, O’Brien

, Ramsay

, et al.; Scottish Intensive Care Trainees’ Audit Share Group. A national service evaluation of the impact of alcohol on admissions to Scottish intensive care units. Anaesthesia, 2012; 67(10):1132–1137; doi: 10.1111/j.1365-2044.2012.07233.x

20.

McKenny

, O’Beirne

, Fagan

, et al. Alcohol-related admissions to an intensive care unit in Dublin. Ir J Med Sci, 2010; 179(3):405–408; doi: 10.1007/s11845-010-0498-y

21.

Rolfzen

, Mikulich-Gilbertson

, Natvig

, et al. Association between alcohol use disorder and hospital outcomes in colectomy patients - A retrospective cohort study. J Clin Anesth, 2022; 78:110674; doi: 10.1016/j.jclinane.2022.110674

22.

Shabanzadeh

, Sørensen

. Alcohol consumption increases post-operative infection but not mortality: A systematic review and meta-analysis. Surg Infect (Larchmt), 2015; 16(6):657–668; doi: 10.1089/sur.2015.009

23.

Somsouk

, Kornfield

, Vittinghoff

, et al. Moderate ascites identifies patients with low model for end-stage liver disease scores awaiting liver transplantation who have a high mortality risk. Liver Transpl, 2011; 17(2):129–136; doi: 10.1002/lt.22218

24.

Huang

, Lin

, Cherng

, et al. A population-based cohort study of mortality of intensive care unit patients with liver cirrhosis. BMC Gastroenterol, 2020; 20(1):15; doi: 10.1186/s12876-020-1163-1

25.

Toufen

, Franca

, Okamoto

, et al. Infection as an independent risk factor for mortality in the surgical intensive care unit. Clinics (Sao Paulo), 2013; 68(8):1103–1108; doi: 10.6061/clinics/2013(08)07

26.

Tran

, Groeneveld

, van der Meulen

, et al. Age, chronic disease, sepsis, organ system failure, and mortality in a medical intensive care unit. Crit Care Med, 1990; 18(5):474–479; doi: 10.1097/00003246-199005000-00002

27.

Rivas

, Sanvisens

, Bolao

, et al. Impact of medical comorbidity and risk of death in 680 patients with alcohol use disorders. Alcohol Clin Exp Res, 2013; 37(Suppl 1):E221–E227; doi: 10.1111/j.1530-0277.2012.01861.x

28.

Moore

, Giuli

, Gould

, et al. Alcohol use, comorbidity, and mortality. J Am Geriatr Soc, 2006; 54(5):757–762; doi: 10.1111/j.1532-5415.2006.00728.x

29.

Simpson

, Puxty

, McLoone

, et al. Comorbidity and survival after admission to the intensive care unit: A population-based study of 41,230 patients. J Intensive Care Soc, 2021; 22(2):143–151; doi: 10.1177/1751143720914229

30.

Esper

, Martin

. The impact of comorbid conditions on critical illness. Crit Care Med, 2011; 39(12):2728–2735; doi: 10.1097/CCM.0b013e318236f27e

31.

Barrasa-Villar

, Aibar-Remón

, Prieto-Andrés

, et al. Impact on morbidity, mortality, and length of stay of hospital-acquired infections by resistant microorganisms. Clin Infect Dis, 2017; 65(4):644–652; doi: 10.1093/cid/cix411

32.

Vincent

, Jones

, David

, et al. Frequency and mortality of septic shock in Europe and North America: A systematic review and meta-analysis. Crit Care, 2019; 23(1):196–111; doi: 10.1186/s13054-019-2478-6

33.

de Roux

, Cavalcanti

, Marcos

, et al. Impact of alcohol abuse in the etiology and severity of community-acquired Pneumonia. Chest, 2006; 129(5):1219–1225; doi: 10.1378/chest.129.5.1219

34.

Samokhvalov

, Irving

, Rehm

. Alcohol consumption as a risk factor for pneumonia: A systematic review and meta-analysis. Epidemiol Infect, 2010; 138(12):1789–1795; doi: 10.1017/S0950268810000774

35.

Simou

, Britton

, Leonardi-Bee

. Alcohol and the risk of pneumonia: A systematic review and meta-analysis. BMJ Open, 2018; 8(8):e022344; doi: 10.1136/bmjopen-2018-022344

36.

Zhang

, Bagby

, Happel

, et al. Alcohol abuse, immunosuppression, and pulmonary infection. Curr Drug Abuse Rev, 2008; 1(1):56–67; doi: 10.2174/1874473710801010056

37.

Szabo

, Mandrekar

, Catalano

. Inhibition of superantigen-induced T cell proliferation and monocyte IL-1 beta, TNF-alpha, and IL-6 production by acute ethanol treatment. J Leukoc Biol, 1995; 58(3):342–350; doi: 10.1002/jlb.58.3.342

38.

Neuman

. Cytokines—Central Factors in Alcoholic Liver Disease. Alcohol Res Health, 2003; 27(4):307–316.