Abstract
Background and Aims
Alcohol consumption is a well-documented determinant of adverse perioperative outcome. We sought to determine the effect of active alcohol consumption following elective surgery.
Methods
We queried discharge records from the American College of Surgeons' National Surgical Quality Improvement Program (NSQIP, 2005–2007) for all elective adult admissions. The 7,631 (2.5%) patients with documented alcohol use (active alcohol use of at least two drinks per day within 2 weeks of surgery; ETOH use) underwent elective surgery; 301,994 (97.5%) patients denied ETOH use. Multivariate analysis was performed with adjustments for demographic and comorbid factors. Primary outcome measures included length of stay (LOS), postoperative complications, and death.
Results
ETOH use associated with elective surgery decreased over the course of the study (p < 0.0001). ETOH use was an independent predictor of pneumonia (OR 1.98, 95% CI 1.84–2.13), sepsis (OR 1.19, 95% CI 1.03–1.37), superficial surgical site infection (SSI; OR 1.15, 95% CI 1.02–1.31), wound disruption (OR 1.41, 95% CI 1.11–1.80), and prolonged LOS (OR 1.17, 95% CI 1.08–1.26). Except for SSI, these complications were independent risk factors for postoperative mortality. ETOH use was associated with earlier time to wound disruption (9 vs. 11 days; p = 0.04), longer median hospital stays (5 vs. 3 days; p < 0.0001), and longer LOS after operation (4 vs. 3 days; p < 0.0001).
Conclusions
Active alcohol consumption is a significant determinant of adverse outcomes in elective surgery; patients with ETOH use who are scheduled to undergo elective surgery should be appropriately educated and counseled.
Similar content being viewed by others
References
Alcoholism, N.I.o.A.A.a. [cited 2009 17 September]; Available from: http://www.niaaa.nih.gov/Resources/DatabaseResources/QuickFacts/AlcoholDependence/default.htm.
Organization, W.H., Global Status Report on Alcohol 2004, D.o.M.H.a.S. Abuse, Editor. 2004, World Health Organization: Geneva.
Moore, R.D., et al., Prevalence, detection, and treatment of alcoholism in hospitalized patients. JAMA, 1989. 261(3): p. 403–7.
Umbricht-Schneiter, A., P. Santora, and R.D. Moore, Alcohol abuse: comparison of two methods for assessing its prevalence and associated morbidity in hospitalized patients. Am J Med, 1991. 91(2): p. 110–8.
Szabo, G. and P. Mandrekar, A recent perspective on alcohol, immunity, and host defense. Alcohol Clin Exp Res, 2009. 33(2): p. 220–32.
Nath, B. and G. Szabo, Alcohol-induced modulation of signaling pathways in liver parenchymal and nonparenchymal cells: implications for immunity. Semin Liver Dis, 2009. 29(2): p. 166–77.
Dolganiuc, A., et al., Additive inhibition of dendritic cell allostimulatory capacity by alcohol and hepatitis C is not restored by DC maturation and involves abnormal IL-10 and IL-2 induction. Alcohol Clin Exp Res, 2003. 27(6): p. 1023–31.
Khuri, S.F., et al., The National Veterans Administration Surgical Risk Study: risk adjustment for the comparative assessment of the quality of surgical care. J Am Coll Surg, 1995. 180(5): p. 519–31.
Ghaferi, A.A., J.D. Birkmeyer, and J.B. Dimick, Variation in hospital mortality associated with inpatient surgery. N Engl J Med, 2009. 361(14): p. 1368–75.
Spies, C.D., 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(6): p. 649–56.
Spies C, et al (2001) Perioperative morbidity and mortality in chronic alcoholic patients. Alcohol Clin Exp Res 25(5):164S–170S
Spies, C.D., et al., Altered cell-mediated immunity and increased postoperative infection rate in long-term alcoholic patients. Anesthesiology, 2004. 100(5): p. 1088–100.
Spies, C., et al., Intervention at the level of the neuroendocrine-immune axis and postoperative pneumonia rate in long-term alcoholics. Am J Respir Crit Care Med, 2006. 174(4): p. 408–14.
Schoenfeld, H., et al., The effect of stress-reducing, low-dose ethanol infusion on frequency of bleeding complications in long-term alcoholic patients undergoing major surgery. Am Surg, 2007. 73(2): p. 192–8.
Spies, C.D., et al., Effects of ethanol on cytokine production after surgery in a murine model of gram-negative pneumonia. Alcohol Clin Exp Res, 2008. 32(2): p. 331–8.
Lau, A., et al., Alcohol use disorder and perioperative immune dysfunction. Anesth Analg, 2009. 108(3): p. 916–20.
Donigan, J.A., et al., Pre-operative and intra-operative factors related to shoulder arthroplasty outcomes. Iowa Orthop J, 2009. 29: p. 60–6.
von Heymann, C., et al., Posttraumatic immune modulation in chronic alcoholics is associated with multiple organ dysfunction syndrome. J Trauma, 2002. 52(1): p. 95–103.
Tonnesen, H., B.T. Schutten, and B.B. Jorgensen, Influence of alcohol on morbidity after colonic surgery. Dis Colon Rectum, 1987. 30(7): p. 549–51.
Felding, C., L.M. Jensen, and H. Tonnesen, Influence of alcohol intake on postoperative morbidity after hysterectomy. Am J Obstet Gynecol, 1992. 166(2): p. 667–70.
Paull, D.E., et al., Complications and long-term survival for alcoholic patients with resectable lung cancer. Am J Surg, 2004. 188(5): p. 553–9.
Lau, A.H., M. Abe, and A.W. Thomson, Ethanol affects the generation, cosignaling molecule expression, and function of plasmacytoid and myeloid dendritic cell subsets in vitro and in vivo. J Leukoc Biol, 2006. 79(5): p. 941–53.
Csikesz, N.G., et al., Nationwide volume and mortality after elective surgery in cirrhotic patients. J Am Coll Surg, 2009. 208(1): p. 96–103.
Eppsteiner, R.W., et al., Surgeon volume impacts hospital mortality for pancreatic resection. Ann Surg, 2009. 249(4): p. 635–40.
Author information
Authors and Affiliations
Corresponding author
Additional information
Discussant
Dr. Craig P. Fischer (Houston, TX): Dr. Nath, and the group at UMass have established a fantastic outcomes research unit which has examined surgical outcomes and have been interested in nomograms that might predict surgical outcomes based upon readily available clinical data.
I have two simple questions.
The first is, your data set tries to look at elective patients, yet I would imagine that most surgeons, if they knew their patient had been drinking alcohol recently, might not operate.
So my question is how good is NSQIP at distinguishing urgent, emergent, and completely elective operations? It's a fine lie between urgent and emergent—and to be useful—this information regarding alcohol consumption must be available prior to a planned operation, with enough time to modify the risk factor.
One of the reasons I liked your abstract was the word in your title, “modifiable.”
So tell me about that. You apparently are interested in finding a way to modify this risk factor. So given this information, this is new, what can you do to now try to find these folks and then limit this risk factor?
Closing Discussant
Dr. Bharath Nath: I'll begin by addressing the first question, which I understood to be how good NSQIP is at distinguishing urgent versus emergent conditions. The code in the database identifies emergent procedures as those that occur within 12 h of admission. So this actually leads to some cases that might be thought of as acute falling into the elective category.
However, our concern in really separating out emergency and elective cases was to discern which patients had been scheduled with enough time for appropriate planning for surgery, and which ones were taken straight to the OR, without time to compensate for other preoperative factors that may predispose to complications.
From that perspective, the difference between emergent and elective codes in NSQIP is sound. Now, that said, it still has the disadvantage of being a binary variable. If time to surgery could be coded as a continuous variable, I think that there might be opportunities to ask some interesting questions for future studies.
Now, to address the second question, namely, if alcohol is a modifiable risk factor, what are the means by which we may modify this factor? Notably, few if any recent discussions on risk factors include alcohol. So I would hope that, on the basis of these data that we've presented today, surgeons would feel comfortable discussing alcohol intake with their patients, and suggest that as long as patients are actively drinking, there's a likelihood that they will have more complications. So from a practice perspective, that's the approach that we would suggest.
Then there's the question, of course, of a hospital process. And I think that as these data become more widely disseminated, it's reasonable to think that a process of identifying these patients on admission and counseling them and making surgeons aware would be useful.
Discussant
Dr. Timothy Pawlik (Baltimore, MD): I want to echo Dr. Fischer's comments about your group at UMass. Really some excellent work.
My comment specifically regards a statistical issue. You showed that the alcohol consumption group and the group that didn't consume alcohol are very disparate. Although you use multivariable logistic regression to control for some of that, I'm sure your group is well aware that this does not suffice when the groups are so different. In fact, causal inferences from that type of modeling, when the groups so different, can be misleading.
Did you use other statistical modeling, like a propensity index, which I think may have worked nicely with this data set? And did you find similar results, if indeed you did use that other modeling?
Closing Discussant
Dr. Bharath Nath: That's a great point. In terms of the work that we presented here today, we did not. However, that is exactly the direction that we are working on right now. So we hope to present those data in the near future.
Discussant
Dr. Steven Demeester (Los Angeles, CA): You know, I'll just follow up on that and say that in the types of complications you presented, I wouldn't have anticipated from an alcohol type thing; they struck me more as smoking things, pneumonia, sepsis, respiratory, well-known problems with smoking.
A simple thing to do is take your people that are drinking alcohol, divide them by those that are active or former smokers versus those that are nonsmokers, and see if your differences really hold up in that alcohol-only group. That would be a quick test to see if you are really onto something or whether you are being confounded by the smoking issue.
Closing Discussant
Dr. Bharath Nath: Absolutely. I think that's also a great point. A synergistic effect between smoking and alcohol would be a very intriguing one to uncover. Now that said, I came to the study after completing my Ph.D. work in the area of alcohol and its effects on pathogenesis of liver disease.
One of the things that becomes clear is that chronic alcohol use tends to predispose to hyperresponsive macrophage activation, particularly when macrophages are challenged with immunogenic particles such as the gram-negative cell wall product lipopolysaccharide. So from that perspective, there is a rationale to think that pneumonia and sepsis could indeed be worsened by chronic alcohol exposure quite independently of smoking.
But I absolutely agree that it would be intriguing to validate that observation, made in the laboratory, with a study that stratified the effect of smoking and ethanol on surgical outcomes.
Discussant
Dr. Gerard V. Aranha (Maywood, IL): In your alcohol group, what percentage of patients had ascites or varices? If you are operating upon Child C group patients, wouldn't you have a higher mortality and morbidity?
Closing Discussant
Dr. Bharath Nath: That's a good point. The prevalence of ascites in the alcohol-exposed gruop was 1.6%, whereas in the non-alcohol-exposed group it was 0.8%. These differences were indeed statistically significant between the two groups. However, ascites was one of the preoperative patient characteristics that we considered when performing our adjusted analysis. Within the limits of performing an adjusted analysis, differences in the prevalence of ascites did not affect our findings.
Discussant
Dr. John Bowen (New Orleans, LA): I don't doubt your correlations; I do question one word you use, which is “determination.” In other words, you correlate these factors with alcoholism, but is it not possible that alcoholism is simply an indicator of a type of personality or person that's coming in and continuing to drink prior to surgery, rather than a specific kind of physiological defect because of the alcohol?
The only way I could see you could unravel this is to identify these people before you operate, not operate on them, have them stay off the alcohol for a couple of weeks, and then operate on them and see if you had a difference, if you really did have all these complications.
It's my feeling that this is more of a proxy for a group of patients that are acting badly, for whatever reason, and it's having a bad effect on them physiologically, emotionally, or whatever. And I would be very interested to see that part of it unraveled.
Closing Discussant
Dr. Bharath Nath: I think that's a great comment as well. I would be very interested to see some sort of clinical trial that compared patients in those two arms. That said, whether it's a physiological process or whether it's an indicator of bad habits, minimizing those may still have the same effect. And so, knowing this data as we do, I would feel that counseling a patient to avoid alcohol might still be in the best interests of the patient.
Supported by American Society of Transplant Surgeons Faculty Development Award and Worcester Foundation for Biomedical Research (SAS). BN is supported by a F30 grant from the National Institute of Alcohol Abuse and Alcoholism of the National Institutes of Health.
Rights and permissions
About this article
Cite this article
Nath, B., Li, Y., Carroll, J.E. et al. Alcohol Exposure as a Risk Factor for Adverse Outcomes in Elective Surgery. J Gastrointest Surg 14, 1732–1741 (2010). https://doi.org/10.1007/s11605-010-1350-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11605-010-1350-4