Alcohol use disorders affect 21–42% of all hospitalized patients.1–3 Surveys conducted by the Centers for Disease Control and Prevention reveal that 51.5% of nonpregnant and 7.6% of pregnant women reported alcohol use, and 15% of nonpregnant and 1.4% of pregnant women reported binge drinking.4 Other research suggests the rate of binge drinking among pregnant women varies from 2.6% to 4.4%, depending on age.5 Obstetricians typically focus on the well-known fetal risks associated with alcohol exposure, including growth restriction and fetal alcohol syndrome, but not on potential maternal consequences during pregnancy.
Among the general population, patients with alcohol use disorders have an increased risk of sepsis and pneumonia, and this risk is increased further among hospitalized patients.6–8 In addition, both medical and surgical patients with alcohol use disorders are at increased risk for development of hospital-acquired infections, resulting in increased length of stay and mortality.6 Overall, hospital-acquired infections are associated with a two-fold to three-fold increased mortality risk.9
Nearly 1.3 million cesarean deliveries were performed in 2011; cesarean delivery is the most common major surgical procedure performed for women.10 Surgical site infections occur in 7–20% of women undergoing cesarean delivery.11,12 Identification of modifiable risk factors for hospital-acquired infection offers an important opportunity to reduce operative morbidity. The ability to counsel regarding immediate maternal risks may offer another reason for patients to modify alcohol consumption during pregnancy.
Our goal was to investigate whether women with alcohol use disorders have an increased risk of hospital-acquired infection after cesarean delivery.
MATERIALS AND METHODS
We performed a retrospective cohort study using the Nationwide Inpatient Sample database for the years 2002 through 2010.13 Approval by the Virginia Commonwealth University Institutional Review Board was not required because the database is in the public domain and has no personal health information. The total sample included approximately 7 million cesarean deliveries over this 9-year period, from which our cases and controls were drawn. Case group participants were patients who were assigned a diagnosis of alcohol use disorder. Control group participants were patients not assigned a diagnosis of alcohol use disorder. Case and control group participants were matched for age (±2 years), race, severity of illness (within 1 point on the Charlson comorbidity index), elective compared with nonelective admission, smoking, hospital size and teaching status, and patient income.14
The Nationwide Inpatient Sample is the largest all-payer database of inpatient admissions and is designed for research.13 The Nationwide Inpatient Sample details the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) procedures and diagnoses for approximately 8 million patients per year from more than 1,000 hospitals in the United States.13,15
To accommodate the large volume of data, each entry represents patients with the same characteristics. The entry has an associated weight called the discharge weight, which represents the number of patients who match the characteristics of the entry row in the database.16 For example, a 22-year-old woman, gravida 1 para 0, with malposition of the fetus is a common indication for cesarean delivery. Thus, if each separate woman was given her own entry in the database, many rows would include women with these specific characteristics. To limit the size of the data set, one row describing these characteristics is entered in the database as well as an additional column with discharge weight, reflecting the number of patients with these specific characteristics. In the analysis, each row is multiplied by this weight to determine the total number of patients with these characteristics.
Alcohol use disorders were defined using ICD-9-CM codes, as noted in Table 1. A woman was classified as having alcohol use disorder if she had a medical diagnosis of acute alcohol intoxication, unspecified alcohol dependence, alcohol abuse, or alcohol-induced mental disorders during the same admission in which the cesarean delivery was performed.
All patients (case and control group participants) were women aged 18 years or older who underwent cesarean delivery as the primary procedure and had an obstetric indication for cesarean delivery, as reflected by ICD-9-CM codes for cesarean delivery. Adolescents were excluded because their patterns of alcohol use are likely different than those of adult women, and consequences may differ because their duration of drinking is shorter.17 If patients underwent cesarean delivery but it was not the primary procedure, the patient was not included. The primary diagnoses associated with the cesarean delivery were abnormality of forces of labor, other indications for care or intervention related to labor and delivery (not elsewhere classified), malposition and malpresentation of fetus, abnormality of organs and soft tissues of pelvis, and late pregnancy (Table 1).
The hospital-acquired infections evaluated were sepsis, endometritis, pneumonia, urinary tract infection, and surgical site infection. Table 1 details the diagnostic ICD-9-CM codes used to identify patients with hospital-acquired infection. The methodology of Eber et al18 was used to identify patients with health care–associated pneumonia and health care–associated sepsis. Patients with development of surgical site infection and urinary tract infection were identified using the diagnostic criteria of Vogel et al.19 Patients who had more than one hospital-acquired infection were counted only once.
Additional exclusion criteria were the same as those used by Eber et al.18 Patients with preexisting infections, acquired immunodeficiency syndrome, immunocompromised states, and cancer were excluded. Patients who were transferred from another hospital or skilled nursing facility also were excluded.
Other variables collected included the hospital day when cesarean delivery was performed, age, race, tobacco dependence (defined as ICD-9-CM 305.1), and severity of illness. Severity of illness was defined by using the Charlson comorbidity index, which is a standardized method of comparing mortality risk among inpatients and has been used previously in studies examining the effect of alcohol use disorders among hospitalized patients.14,20 Admissions were classified as elective or nonelective. Patient income, hospital size, and hospital teaching status also were collected. The Nationwide Inpatient Survey divides patient income into quartiles based on the U.S. Postal Service ZIP code of the patient's residence. Hospital size was divided into small or medium hospitals and large hospitals.
The primary outcome of the study was to compare the risk of development of any hospital-acquired infection in patients with and without alcohol use disorders. A secondary outcome was to compare hospital length of stay.
Univariate analyses examining the relationship between alcohol use disorders and hospital-acquired infection, hospital day when cesarean delivery was performed, age, race, alcohol use disorders, tobacco dependence, severity of illness, hospital size, hospital teaching status, and income were performed.
The decision to match within a single point on Charlson comorbidity illness was made consistent with previously published similar analyses of alcohol use disorders and surgical morbidity.21 Patients undergoing cesarean delivery are, in general, much healthier than those in other studies. Among healthier patients, a single point difference, the smallest difference possible on the scale, is likely clinically significant. To adjust for this potential difference, we adjusted for Charlson comorbidity index in the multivariable analyses.
Multivariable models using logistic regression were computed to determine the effects of alcohol use disorders on development of hospital-acquired infection. Multivariable analyses were performed to determine the effect of alcohol use disorders on hospital length of stay in patients with hospital-acquired infection. The data were adjusted for age, race, tobacco dependence, severity of illness, elective compared with nonelective admission, hospital size, hospital teaching status, income, and all possible two-way interactions. Using a backward elimination strategy, terms remained in the model if the P value was less than the alpha value. Because of the large sample size, we set the alpha at 10−6 and considered significance only when P value was less than the alpha value. Analysis was performed with SAS 9.3 and JMP Pro 10.
A total of 2,484 database entries including a diagnosis of alcohol use disorders were identified. Each of these 2,484 entries was matched with a control entry. After multiplying each entry by its corresponding discharge weight, a total of 12,081 women with alcohol use disorders were identified and matched with 11,960 women without alcohol use disorders who underwent cesarean delivery. Because of the nature of the database, the total number of women with alcohol use disorders is slightly different than the number of matched controls without alcohol use disorders. A total of 24,041 women were evaluated, with approximately half having alcohol use disorders and half not having alcohol use disorders.
Table 2 details the characteristics of the case and control group participants. Patients with and without alcohol use disorders were otherwise similar except for Charlson comorbidity index. Univariate analysis (Table 3) suggests that patients with alcohol use disorders were more likely to have development of any type of hospital-acquired infection, particularly urinary tract infection and sepsis. Because of our stringent alpha, the rate of endometritis was not statistically different between patients with and without alcohol use disorders, although it did occur approximately twice as frequently. The two groups did not differ in rate of development of surgical site infection, with similar point estimates. No patient had development of pneumonia. Patients with alcohol use disorders had a longer length of stay of 3.3 days (95% confidence interval [CI] 3.2–3.3) compared with 3.1 days (95% CI 3.0–3.1; P=8×10−10). Alcohol use disorders were not associated with a higher probability of death (P=1.9×10−4).
Multivariable models were fit for each primary and secondary outcome, and they are presented by outcome. No interaction term was significant in any of the models. Multivariable analysis revealed that Charlson comorbidity index did not affect the rate of hospital-acquired infection (P=.41). The odds of development of any type of hospital-acquired infection were higher in the group of patients with alcohol use disorders (odds ratio [OR] 2.2, 95% CI 1.9–2.7; P=2.6×10−19; number needed to harm 55). Among patients with alcohol use disorders, 397 had development of hospital-acquired infections, whereas only 179 patients without alcohol use disorders had development of these infections.
Neither adjusting for severity of illness by Charlson comorbidity index (P=.07) nor hospital-acquired infection (P=.99) altered the difference in length of stay between groups. Alcohol use disorders were an independent predictor of increased length of stay: 3.3 days (95% CI 3.2–3.3) in patients with alcohol use disorders compared with 3.1 days (95% CI 3.0–3.1) in patients without alcohol use disorders (P=4×10−7).
Multivariable analysis adjusting for Charlson comorbidity index found that this index had no effect on mortality (P=.99). Similarly, multivariable analysis adjusting for hospital-acquired infection found that it had no effect on mortality (P=.99). Thus, severity of illness and hospital-acquired infection did not affect mortality.
Using a national cohort of patients, this study demonstrates an increased risk of hospital-acquired infection, including urinary tract infection and sepsis, in women with alcohol use disorders undergoing cesarean delivery. Pregnant women with alcohol use disorders were more than twice as likely to experience a hospital-acquired infection (OR 2.2) than women without alcohol use disorders. The most common hospital-acquired infection in this group was urinary tract infection (OR 2.5), which was found to be significant. Endometritis (OR 2.1) and surgical site infection (OR 1.2) were increased, but these were not significant. Overall, the number needed to harm was 55, indicating that for 55 women with alcohol use disorders undergoing cesarean delivery, 1 additional patient has development of hospital-acquired infection. These data do not reflect readmissions or outpatient visits, possibly leading to an underestimation of the risk of surgical site infections, which are often diagnosed and managed subsequent from the initial admission.
These findings are consistent with research evaluating the risk associated with alcohol use disorders in other patient populations. A recent review of the effect of alcohol use disorders on a variety of patient groups showed that general surgical patients with long-term alcohol abuse are at an increased risk for major postoperative infectious complications. Recent research by de Wit21 has linked alcohol use disorders to an increased risk of hospital-acquired infection in a variety of surgical patients.
Clinicians rightly focus on the fetal risks of alcohol use in pregnancy, and the role of alcohol as a teratogen and its adverse effects on fetal development are well-known.22 To date, however, there is a paucity of research evaluating the effect of alcohol use disorders on maternal outcomes, specifically as they relate to infectious morbidity after cesarean delivery. Using the Nationwide Inpatient Sample from 2002 to 2010, including more than 7 million cesarean deliveries, we evaluated the effect of alcohol use disorders on hospital-acquired infection in patients undergoing cesarean delivery. As noted, the overall risk of hospital-acquired infection and specific risks of urinary tract infection and sepsis significantly increased in women with alcohol use disorders. As noted, we set a very stringent P value (P<10−6) to minimize the risk of type II error given such a large sample size and corresponding high power.
Use of administrative databases for research has significant limitations and has been justifiably criticized.23 The Nationwide Inpatient Sample is designed for research and, hence, is derived from administrative databases that rely on the accuracy with which ICD-9-CM codes are abstracted from patient records. Clinical diagnoses were based on definitions of alcohol use disorders and hospital-acquired infection as assigned by the providers and may not necessarily reflect the Centers for Disease Control and Prevention definitions. Some data items, like mortality and cesarean delivery, are likely to be coded accurately. Some, like alcohol use disorders, are clearly undercoded. In fact, only approximately 12,000 women were coded as having alcohol use disorders, approximately 0.9% of the total. According to the 2011 National Household Survey on Drug Abuse, 9.4% of pregnant women reported current alcohol use and 2.6% reported current binge drinking, and similar research places this rate of binge drinking at up to 4.4%.5 Our data clearly show that screening for alcohol use disorders is inadequate as recorded in the medical records of pregnant patients. It is possible that such undercoding of alcohol use disorders would lead to only the most severely affected cases being reflected in our data. However, we anticipate that such undercoding of alcohol use disorders is more likely to result in underestimation of the effect of alcohol use disorders. We had no ability to determine the volume of alcohol consumption and, hence, are unable to address whether there is a threshold of consumption reflecting a significantly increased risk of complications. The database does not include data on outpatient care and, because some delayed complications may be largely managed in an ambulatory environment, the effect of their omission is difficult to predict. However, the database should encompass significant complications occurring during the index hospitalization, a timeframe clinically relevant in itself. Despite the limitations of the data set, the effects seen here are consistent with known effects of alcohol and similar risks noted in other surgical populations, lending further credence to the observed relationships.
Our study demonstrates an increased risk for hospital-acquired infection in patients with alcohol use disorders who undergo cesarean delivery. These results highlight the importance of screening pregnant patients for alcohol use disorders and engaging these patients in appropriate counseling and treatment programs as needed using one of a variety of validated simple screening tools.24
Reducing alcohol use during pregnancy would serve to reduce the well-known fetal risks of alcohol use during pregnancy and may reduce the likelihood of adverse surgical infectious complications. All women who either are pregnant or are trying to become pregnant are advised to avoid any alcohol consumption.22,24 Obstetrician–gynecologists in particular, as well as other providers of primary care services to women, have a unique opportunity to help identify women with at-risk alcohol use, to intervene to promote healthy behaviors, and to refer patients to substance abuse specialists as appropriate.
1. Gerke P, Hapke U, Rumpf HJ, John U. Alcohol-related diseases in general hospital patients. Alcohol Alcohol 1997;32:179–84.
2. Smothers BA, Yahr HT, Ruhl CE. Detection of alcohol use disorders in general hospital admissions in the United States. Arch Intern Med 2004;164:749–56.
3. Lau K, Freyer-Adam J, Coder B, Riedel J, Rumpf HJ, John U, et al.. Dose-response relation between volume of drinking and alcohol-related disease in male general hospital inpatients. Alcohol Alcohol 2008;43:34–8.
4. Centers for Disease Control and Prevention. Alcohol use and binge drinking among women of childbearing age—United States, 2006-2010. MMWR Morb Mortal Wkly Rep 2012;61:534–8.
5. Substance Abuse and Mental Health Services Administration. Results from the 2011 national survey on drug use and health: summary of national findings, NSDUH series H-41, HHS publication no. (SMA). 11–4658. Rockville (MD): Substance Abuse and Mental Health Services Administration; 2012.
6. Spies CD, Nordmann A, Brummer G, Marks C, Conrad C, Berger G, et al.. Intensive care unit stay is prolonged in chronic alcoholic men following tumor resection of the upper digestive tract. Acta Anaesthesiol Scand 1996;40:649–56.
7. Gacouin A, Legay F, Camus C, Volatron AC, Barbarot N, Donnio PY, et al.. At-risk drinkers are at higher risk to acquire a bacterial infection during an intensive care unit stay than abstinent or moderate drinkers. Crit Care Med 2008;36:1735–41.
8. Moss M, Parsons PE, Steinberg KP, Hudson LD, Guidot DM, Burnham EL, et al.. Chronic alcohol abuse is associated with an increased incidence of acute respiratory distress syndrome and severity of multiple organ dysfunction in patients with septic shock. Crit Care Med 2003;31:869–77.
9. Klevens RM, Edwards JR, Richards CL Jr, Horan TC, Gaynes RP, Pollock DA, et al.. Estimating health care-associated infections and deaths in U.S. hospitals, 2002. Public Health Rep 2007;122:160–6.
10. Hamilton BE, Martin JA, Ventura SJ. Births: preliminary data for 2011. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics National Vital Statistics Report. 2012;61:5. Available at: http://www.cdc.gov/nchs/data/nvsr/nvsr61/nvsr61_05.pdf
. Retrieved January 10, 2013.
11. Yokoe DS, Christiansen CL, Johnson R, Sands KE, Livingston J, Shtatland ES, et al.. Epidemiology of and surveillance for postpartum infections. Emerg Infect Dis 2001;7:837–41.
12. Ramsey PS, White AM, Guinn DA, Lu GC, Ramin SM, Davies JK, et al.. Subcutaneous tissue reapproximation, alone or in combination with drain, in obese women undergoing cesarean delivery. Obstet Gynecol 2005;105:967–73.
13. HCUP databases. Healthcare cost and utilization project (HCUP). 2012. Agency for Healthcare Research and Quality, Rockville (MD). Available at: www.hcup-us.ahrq.gov/nisoverview.jsp
. Retrieved January 10, 2013.
14. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83.
15. Centers for Disease Control and Prevention. Classification of diseases, functioning, and disability. Available at: http://www.cdc.gov/nchs/icd.htm
. Retrieved January10, 2013.
16. HCUP NIS. Description of data elements. Healthcare cost and utilization project (HCUP). 2008. Agency for Healthcare Research and Quality; Rockville (MD). Available at: www.hcup-us.ahrq.gov/db/vars/discwt/nisnote.jsp
. Retrieved January 10, 2013.
17. Substance Abuse and Mental Health Services Administration, Results from the 2011 national survey on drug use and health: summary of national findings, NSDUH series H-44, HHS publication No. (SMA) 12–4713. Rockville (MD): Substance Abuse and Mental Health Services Administration; 2012. Available at: http://www.samhsa.gov/data/NSDUH/2k11Results/NSDUHresults2011.pdf
. Retrieved December 20, 2012.
18. Eber MR, Laxminarayan R, Perencevich EN, Malani A. Clinical and economic outcomes attributable to health care-associated sepsis and pneumonia. Arch Intern Med 2010;170:347–53.
19. Vogel TR, Dombrovskiy VY, Lowry SF. In-hospital delay of elective surgery for high volume procedures: the impact on infectious complications. J Am Coll Surg 2010;211:784–90.
20. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, et al.. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005;43:1130–9.
21. de Wit M, Goldberg S, Hussein E, Neifeld JP. Health care-associated infections in surgical patients undergoing elective surgery: are alcohol use disorders a risk factor? J Am Coll Surg 2012;215:229–36.
23. Grimes DA. Epidemiologic research using administrative databases: garbage in, garbage out. Obstet Gynecol 2010;116:1018–9.
© 2013 by The American College of Obstetricians and Gynecologists.
24. At-risk drinking and alcohol dependence: obstetric and gynecologic implications. Committee Opinion No. 496. American College of Obstetricians and Gynecologists. Obstet Gynecol 2011;118:383–8.