This prospective, observational study performed on a large population of critically ill patients assessed upon admission to the ICU, and distinguished according to the presence or absence of infection, shows that previous alcohol consumption has an impact on counts of circulating WBC involved in both innate and adaptive immunity. The only difference in circulating WBC counts between at-risk and not-at-risk drinkers was in B lymphocytes, which were significantly lower in at-risk drinkers. We found that B lymphocytes, cytotoxic T lymphocytes, noncytotoxic T lymphocytes, and CD16– monocytes in at-risk drinkers, with or without infection, were similar to those in not-at-risk drinkers with infection and significantly lower than those in not-at-risk drinkers without infection. When severity of infection was considered, none of the subsets of circulating WBC studied varied significantly in at-risk drinkers. Neutrophil CD64 index varied significantly by severity of infection both in at-risk and not-at-risk drinkers.
Because individuals with a history of alcohol abuse are more likely to develop severe pneumonia leading to ICU admission, the effect of chronic alcohol exposure on immunity of the lung has been assessed in many in vitro and in vivo studies [10, 12]. In particular, it has been shown that ethanol consumption reduces neutrophil recruitment and neutrophil superoxide production during pulmonary bacterial infection and diminishes phagocytic activity, as well as cytokine and chemokine production by alveolar macrophages after lipopolysaccharide treatment [33, 34]. Paradoxically, alterations in count and function of WBC associated with excessive chronic alcohol consumption have been less studied in the circulation than in the lung. To our knowledge, this is the first study performed in a large population of critically ill patients to assess the behavior of various subsets of circulating WBC in response to infection depending on whether or not patients were at-risk drinkers.
The low WBC counts in at-risk drinkers observed in our study are in accordance with published data showing that chronic alcohol exposure results in hyporesponsiveness of neutrophils to chemotactic signals, reduces delayed hypersensitivity response of isolated lymphocytes after stimulation in vitro by mitogens, and blunts CD4+ and CD8+ lymphocytes [2, 10, 12, 14, 35].
Various abnormalities in circulating neutrophils also have been described with chronic alcohol consumption, ranging from an increase in the number of these cells in the peripheral blood to neutropenia in those with the most severe form of infection or severe underlying hepatic disease . For our part, we did not find that neutrophil counts differed significantly when compared between at-risk and not-at-risk drinkers at ICU admission. Interestingly, when functional activity of neutrophils was assessed by the expression of Fcγ receptor I, which is a marker of neutrophil activation recognized by the monoclonal antibodies CD64, we found that neutrophil CD64 expression varied significantly with infection and severity of infection, but not with alcohol consumption.
In the present study, most of the at-risk drinkers had been exposed to alcohol for many years. The low circulating B-lymphocyte count found in at-risk drinkers is in agreement with results from previous studies showing that the number of peripheral blood B cells is diminished after long-term alcohol consumption. Laslo et al. [36, 37] previously showed that there is a decrease in the number of total B cells and the CD5+/CD19+ subset of B cells in patients with alcoholic liver cirrhosis and a decrease in number of CD5+ B cells in patients with active alcoholism that do not have liver disease. The number of circulating B cells also is reduced in mice undergoing chronic ethanol consumption . Our results could be explained by a lower production of B lymphocytes in critically ill patients that were chronically exposed to alcohol and by alterations in the interactions between T and B lymphocytes. Previous experimental or clinical studies have shown a reduced number of cells in the thymus of patients chronically exposed to alcohol, a decreased activation of lymphocytes after antigen stimulation, reduced cytokine production by macrophages and T lymphocytes, and inhibited monocyte-derived myeloid cell capacity to induced T-cell activation [10, 11].
Results listed in Table 2 are suggestive of an important impact of infection and chronic alcohol consumption on counts of circulating B lymphocytes, cytotoxic T lymphocytes, and CD16– monocytes. However, for these subsets of WBC, multivariate analysis failed to demonstrate that at-risk drinking was an independent predictor when infection was included in the model. Because our findings suggest that at-risk drinkers admitted to the ICU with infection are less prompt to develop intense immune response than not-at-risk drinkers, we believe that systematic and accurate identification of patients with prior alcohol misuse will lead to improved care for these patients. In addition, our results suggest that neutrophil CD64 index may help physicians to diagnose infection in at-risk drinkers.
Whilst our study has some strengths, including a large number of patients and a long history of alcohol consumption in most of the at-risk drinkers, our study also has some limitations. The impact of alcohol exposure on the functionality of WBC was not assessed, except for neutrophils by looking at CD64 expression. Also, serum levels of immunoglobulins and cytokines were not analyzed in this study, whereas previous authors have shown that, at the onset of infection and during early septic shock, chronic alcoholic patients had lower plasma levels of proinflammatory interleukins than nonalcoholic patients . Patients were assessed at different times during the course of infection, but it must be noted that at-risk and not-at risk drinkers did not differ in the duration of symptoms or in antibiotic therapy before admission.
Patients were not screened for illicit drugs and assessment of blood alcohol was not systematically performed; therefore, we cannot exclude that acute alcohol consumption was unrecognized in some patients. In human studies focusing on defects in the immune system associated with alcohol abuse, it is important to differentiate between acute and chronic alcohol exposure, and the presence or absence of acute hepatitis or liver cirrhosis. Acute alcohol consumption has effects on inflammatory cell activation opposite to those seen with chronic alcohol consumption [9, 10]. In the present study, patients admitted with acute alcohol intoxication were excluded and very few patients had liver cirrhosis. Determination of at-risk drinking was based on results of interviews with patients and relatives regarding preadmission alcohol drinking habits. Therefore, some patients may have been misclassified, particularly due to underestimation of daily alcohol intake. However, the general characteristics of at-risk drinkers were similar to those previously reported by us  and others [38–40]. Even if biological tests have poor performance for screening at-risk drinking in critically ill patients, not-at-risk and at-risk drinkers differed significantly for liver enzymes, MCV, and GGT levels; thus, we believe that, in general, our patient groups were correctly classified. The proportion of patients with liver cirrhosis may have been underestimated. A liver biopsy should have been performed to eliminate or confirm with certainty the diagnosis of cirrhosis. Although not the main focus on the study, it is notable that as reported in a previous study  the proportion of patients with ICU-acquired infection was significantly higher in at-risk drinkers than in not-at-risk drinkers (data not shown).