Decreased Endotoxin Immunity Is Associated with Greater Mortality and/or Prolonged Hospitalization after Surgery

After obtaining institutional review board approval, patients undergoing major, elective, noncardiac surgery at The Mount Sinai Hospital were enrolled in a prospective blinded observational study. The Mount Sinai Hospital Institutional Review Board (New York, NY) did not require patient consent for this study given its observational nature. Patients undergoing the following elective surgical procedures were enrolled: major orthopedic (e.g. , revision hip arthroplasty, fusion–instrumentation of multiple lumbar or thoracic vertebrae); major general (e.g. , any laparotomy expected to exceed 2-h duration, including partial hepatectomy, pancreatic surgery, colon surgery); major urologic (e.g. , radical cystectomy, radical nephrectomy); major vascular (e.g. , abdominal aortic aneurysm repair); and major gynecologic (e.g. , cancer debulking procedure, abdominal hysterectomy with oophorectomy). These procedures were selected for several reasons: (1) they are routinely performed surgeries; (2) they represent a diverse group of procedure types; and (3) a previous study performed at Duke University Medical Center found these procedures to be associated with prolonged hospitalization and postoperative complications. 4Patients undergoing emergency surgery were excluded from this study. Although every effort was made to enroll the patients in as consecutive a fashion as possible, this was not always possible because of logistical reasons.

Study patients received routine anesthetic care and surgical management, including the routine use of prophylactic antibiotics. Each patient was managed postoperatively according to standard institutional surgical “care maps”; only patients undergoing surgical procedures enabling them to be targeted for discharge from the hospital before the 10th postoperative day were enrolled.

Assays for the determination of immunoglobulin M antiendotoxin core antibody (EndoCAb), immunoglobulin G EndoCAb, total immunoglobulin M, and total immunoglobulin G concentrations were conducted on residual serum that was no longer needed by the Mount Sinai Hospital clinical laboratory. All of these serum samples were obtained preoperatively (within 72 h before surgery) for the determination of routine chemistry tests. Blood samples were collected from patients by venipuncture into nonadditive glass tubes, centrifuged, and serum stored at −70°C until assayed according to the institution’s standard operating procedure. Immunoglobulin M and immunoglobulin G EndoCAb concentrations were measured using an enzyme-linked immunosorbent assay described previously. 19,21In brief, incomplete core, rough, mutant lipopolysaccharide from each of four species of gram-negative bacteria (Escherichia coli , Pseudomonas aeruginosa , Klebsiella aerogenes , and Salmonella typhimurium ) were complexed with polymyxin B and coated on polystyrene microplates. Immunoglobulin G EndoCAb and immunoglobulin M EndoCAb concentrations were determined with alkaline phosphatase–conjugated antibodies specific for human γ-globulin γ or μ heavy chains (Zymed Laboratories, Cambridge BioScience, Cambridge, United Kingdom). Test sera were compared in enzyme-linked immunosorbent assay to a reference serum calibrated in EndoCAb median units, where 100 is the median value for 1,000 healthy adults’ immunoglobulin G or immunoglobulin M, respectively. 19Serum was tested for total immunoglobulin G and immunoglobulin M concentrations by laser nephelometry. 22All laboratory measurements were performed on coded samples so as to blind the investigators to the patients’ identity and outcome.

Demographic information was collected and a preoperative physiologic risk score for mortality and morbidity was assigned according to the established POSSUM criteria, 23which has been recognized as being the most appropriate available score for assessing risk in noncardiac surgical patients. 24 

The POSSUM physiologic score includes the following 12 preoperative factors: age, signs of symptoms of heart failure, abnormal electrocardiogram, respiratory function, blood pressure, heart rate, mental status, hemoglobin concentration, leukocyte count, blood urea concentration, blood sodium concentration, and blood potassium concentration. A point value of 1, 2, 4, or 8 is assigned for 10 of the 12 factors depending on the severity of the abnormality, e.g. , 1 point for no dyspnea and 8 points for dyspnea at rest, 1 point for normal potassium concentration and 8 points for a potassium concentration greater than or equal to 6, 1 point for normal heart rate and 8 points for a heart rate greater than or equal to 121 beats per minute. POSSUM physiologic scores can range between 12 and 88. The preoperative physiologic score predicts the risk of morbidity and mortality for noncardiac surgical patients. 23This score was used in the multivariate logistic regression model to adjust for the effect of patient comorbidities. Use of this risk score obviated the need to adjust for the effect of individual patient comorbidities.

There is no established definition that encompasses all systemic inflammation–related morbidity in postoperative (non–intensive care unit) patients. End points for single-organ system failure are not applicable to studies such as this one. We prospectively defined adverse postoperative outcome as either in-hospital death or a postoperative length of stay greater than 10 days. This endpoint is identical to that used in a similar study involving more than 300 cardiac surgical patients. 20Postoperative length of stay was defined as the number of days from the day of operation (day 0) to hospital discharge. A 10-day cutoff was selected a priori  for several reasons. First, at our hospital and in many others, care maps target discharge of the patient from the hospital well before the 10th postoperative day, with most patients being discharged within 1 week of surgery. In fact, the 25th percentile for postoperative length of stay was 6 days in this cohort; hence, a 10-day cutoff represented a significant prolongation (> 4 days) in the length of stay, not a trivial prolongation of 1 or 2 days that could be a result of “soft factors” such as physician preference or lack of transportation. Second, a retrospective analysis of patients undergoing noncardiac surgery at our institution demonstrated that the majority of patients with a length of stay of greater than 10 days suffered morbidity that could be consistent with an endotoxin-mediated proinflammatory response. It has been shown that the overwhelming majority of patients with prolonged hospitalization after cardiac and noncardiac surgery have morbidity. 4,20It is important to emphasize that complications potentially related to excessive systemic inflammation are not limited to organ failure observed in patients in the intensive care unit, but include less severe forms of morbidity that nevertheless prolong hospitalization.

To characterize postoperative complications in patients with prolonged hospitalization (for descriptive purposes only), patients were evaluated on postoperative day 11 using a postoperative morbidity survey that has been previously described in a prospective study of 438 patients undergoing routine noncardiac surgery. 4The criteria in this survey identify both severe complications (e.g. , pulmonary failure–acute respiratory distress syndrome) as well as more subtle morbidity that could delay discharge from the hospital (e.g. , moderate pulmonary dysfunction necessitating supplemental oxygen therapy). Existing morbidity scoring systems such as APACHE and SAPS were not used in this study because they are geared to the intensive care unit setting and thus have limited value in most postoperative patients. 25All outcome measurements were performed by investigators blinded to the patients’ laboratory data.

The objective of this study was to test the specific hypothesis that low preoperative serum EndoCAb concentrations are associated with an increased likelihood of postoperative mortality or prolonged hospitalization after controlling for the POSSUM physiologic risk score and total immunoglobulin concentrations. Data from a previous study of 301 patients in a different population showed an incidence of 10% for the primary end point (mortality or prolonged hospitalization). 20We estimated that for this study, a sample size of more than 1,000 would generate at least 100 events and allow for a stable logistic regression model to test the hypothesis of the study. The required sample size for the current study was estimated to be larger than our previous study of cardiac surgical patients because of the greater heterogeneity of procedures in this study.

The relation between each antibody, as well as preoperative POSSUM risk score, and chance of postoperative complication was explored initially with a series of contingency tables. The variables were then entered in a stepwise multiple logistic regression analysis; those with evidence of a linear relation were entered as continuous variables, and the others were dichotomized at the median. The procedure was set to enter and retain variables with a P  value of 0.10 or less.

The associations between pairs of continuous variables (e.g. , immunoglobulin M EndoCAb and POSSUM risk score) were estimated by the Spearman rank correlation coefficient. All statistical calculations and analyses were conducted using the SAS software system (SAS Institute Inc., Cary, NC).

A total of 1,056 patients were enrolled in the study. Perioperative characteristics (table 1) are presented for the study population. Of note, the median operative duration exceeded 3.5 h, which is consistent with our desire to enroll patients undergoing moderate- and high-risk surgical procedures. The mortality rate was 2.1% (n = 22). The causes of death were as follows: multiple system organ failure (n = 15), pulmonary failure (n = 4), bowel necrosis (n = 1), pulmonary embolus (n = 1), and myocardial infarction (n = 1). Two hundred thirty-four of the 1,056 patients (22.1%) fulfilled the prospectively defined composite outcome. For descriptive purposes only, the incidences of complications on postoperative day 11 are shown in table 2.

Immunoglobulin M EndoCAb concentrations (median units) ranged from 9 to 4,946, with a median of 139. Immunoglobulin G EndoCAb concentrations (median units) ranged from 9 to 4,638, with a median of 249. POSSUM preoperative physiologic risk scores ranged from 12 to 42, with a median of 18. Median immunoglobulin M EndoCAb concentrations were 99 and 150 in patients with and without postoperative morbidity, respectively. Median immunoglobulin G EndoCAb concentrations were 232 and 252 in patients with and without postoperative morbidity, respectively.

Immunoglobulin M and immunoglobulin G EndoCAb concentrations were not highly correlated (Spearman correlation coefficient, 0.374), nor were EndoCAb concentrations and the physiologic preoperative POSSUM scores (immunoglobulin M Spearman correlation coefficient, −0.200; immunoglobulin G Spearman correlation coefficient, −0.011). Poor to marginal Spearman rank correlations were found between EndoCAb concentrations and their respective total concentrations (immunoglobulin G, 0.328; immunoglobulin M, 0.653).

Figure 1presents the percentage of patients having a postoperative complication in approximate quartiles of immunoglobulin M EndoCAb concentration. Immunoglobulin G EndoCAb, total immunoglobulin M, and total immunoglobulin G showed no evidence of association with postoperative complication on either the univariate or the multivariate analyses. Immunoglobulin M EndoCAb concentration and POSSUM physiologic risk score were each significantly associated with a poor outcome (table 3) and, in the multiple logistic regression analysis, proved to be independent predictors. Lower immunoglobulin M EndoCAb concentration predicted increased risk of postoperative complication, independent of the effect of POSSUM physiologic risk score (P = 0.048, odds ratio = 0.925, 95% confidence interval, 0.856–0.999 for each 100-point increase in immunoglobulin M EndoCAb concentration). POSSUM physiologic risk score was also significant in this model (P < 0.001, odds ratio = 1.088, 95% confidence interval, 1.062–1.114).

We have completed a large prospective study evaluating the association between preoperative antiendotoxin immune status and death or prolonged hospitalization in 1,056 patients undergoing routine noncardiac surgery. Low preoperative serum immunoglobulin M EndoCAb concentration is associated with adverse outcome after routine noncardiac surgery and may account in part for variability in outcome observed among patients with similar degrees of preoperative risk.

Vital healthcare resources are used to care for patients with postoperative morbidity. Endotoxemia has been reported to occur during major surgery 8–15and may initiate proinflammatory cascades and cause organ dysfunction. 3,5,6The serum from mature humans contains endogenous antibodies to endotoxin, presumably elicited in response to intermittent exposure to intestinal endotoxin. 19Although serotype-specific antibodies are only protective against endotoxin from the homologous serotype, 26antibodies directed against the core structure of endotoxin should theoretically be cross-reactive against most enteric lipopolysaccharides. 27,28Therefore, in this study we used an assay that measures the serum concentration of EndoCAb. 19,21The serum concentration of these antibodies has been previously measured in more than 1,000 human volunteer blood donors and demonstrate marked variability, with some subjects having several hundred-fold higher concentrations than others. 19 

Antiendotoxin core antibody concentrations have been measured previously in patients in the intensive care unit with established sepsis syndrome 29,30and in cardiac surgical patients. 20,31,32Results from studies involving these specialized patient groups cannot be generalized to the large number of patients who undergo noncardiac surgery for several reasons. Patients with sepsis syndrome are often admitted to an intensive care unit and enrolled into clinical studies hours or days after the inciting illness or injury. 29,30Thus, in these studies, it is unclear whether antiendotoxin antibody concentrations are a predictor of disease versus  a marker for patients who have sustained a greater insult. Studies involving cardiac surgery patients enrolled patients undergoing cardiopulmonary bypass, which has been shown to have marked effects on the immunologic and hematologic systems and may result in organ dysfunction in the absence of endotoxemia. Furthermore, in cardiac surgical patients, cardiopulmonary bypass can have deleterious effects on splanchnic perfusion and hence cause leakage of endotoxin. 3This mechanism cannot be relevant to patients undergoing general surgery because they are not subjected to extracorporeal circulation.

In contrast to intensive care unit and cardiac surgical patients, little is known about the role of preoperative antiendotoxin immune status in general surgical patients. In a study of 86 women undergoing major gynecologic surgery, conducted move than a decade ago, the presence of preoperative antiendotoxin antibodies was associated with decreased postoperative infections. 33However, this study did not quantitate antiendotoxin antibodies, measure total immunoglobulins, assign risk scores to patients preoperatively, or comment on postoperative length of stay or other complications.

We prospectively studied the association of preoperative antiendotoxin immune status and death or prolonged hospitalization in 1,056 patients undergoing routine noncardiac surgery. Several conclusions can be drawn from our study. First, low preoperative immunoglobulin M EndoCAb concentration predicted morbidity after noncardiac surgery independent of several known preoperative risk factors and overall humoral immunity. Our findings, in conjunction with similar results described in previous work involving cardiac surgical patients, suggest that endotoxemia may be an important cause of postoperative morbidity in a broad range of surgical procedures and does not appear to be limited to patients undergoing cardiopulmonary bypass.

Second, preoperative immunoglobulin M and immunoglobulin G EndoCAb concentrations were higher in our study involving patients in the United States compared with a series of 1,000 healthy Scottish volunteers. 32This variation in endogenous immunity may reflect genetic differences, environmental factors (e.g. , diet, which can alter intestinal flora), and/or a greater degree of exposure to endotoxin in patients with medical problems requiring surgery. This issue does not impact on our results because we did not analyze the data with regard to “normal” values. Differences of these concentrations in varying populations warrant further study.

A potential limitation of our study relates to the selection of a primary end point, a challenge to any investigator interested in studying postoperative morbidity attributable to excessive systemic inflammation. Mortality may be an appropriate end point in sepsis studies involving critically ill patients given the high incidence of death and the overwhelming magnitude of systemic inflammation in these patients. Surgical patients, however, appear to develop a less severe form of systemic inflammation that rarely results in death but nevertheless causes morbidity and prolonged hospitalization. There are no validated postoperative morbidity scores that can be easily used in large prospective studies such as ours. Existing morbidity scoring systems such as APACHE were not used in this study because they are geared to the intensive care unit setting, generally rely on laboratory tests that are not routinely available in postoperative patients, and thus have limited value in patients after routine elective surgery. 25The morbidity observed in patients with prolonged hospitalization in our study involved organ systems unrelated to the type or site of surgery. Infectious, pulmonary, and renal complications were observed frequently. Although our study was not designed to prove a causal relation, the pattern of complications observed is consistent with an exaggerated systemic inflammatory response, perhaps partly as a result of exposure to endotoxin.

A major limitation of any observational study such as ours is that low antiendotoxin immunity may not be a cause of poor outcome but merely a marker for sicker patients with higher operative risk and poorer general immunity. Our study minimized this possibility by using an established preoperative risk scoring system to quantify degree of risk, and we also measured and accounted for total immunoglobulin concentrations. Preoperative antiendotoxin antibody concentrations and preoperative risk score and total immunoglobulin concentrations were not well correlated, which is consistent with the results of the multivariate analysis in which immunoglobulin M EndoCAb concentration was an independent predictor of postoperative morbidity. Results from controlled studies in which endotoxin is selectively neutralized will be necessary to confirm the clinical relevance of endotoxemia. Several antiendotoxin-related agents are currently under development that may potentially benefit surgical patients if given prophylactically. 34–39 

There are two commonly stated criticisms of the theory that endotoxemia is an important cause of postoperative morbidity. One criticism relates to the low incidence of culture-proven bacteremia in surgical patients. However, endotoxemia is clearly prevalent in these patients and usually exists in the setting of negative blood cultures. 12In fact, the routine administration of prophylactic antibiotics to surgical patients would be expected to kill or prevent the growth of susceptible gram-negative bacteria and might actually elevate endotoxin blood concentrations through increased shedding of endotoxin. 40Studies attempting to detect endotoxemia probably underestimate its incidence given the intermittent nature of endotoxemia. Another criticism stems from the failure of two anti–lipid A monoclonal antibodies (HA-1A, Centocor, Malvern, PA; and E5, Xoma, Berkeley, CA) to improve outcome on an intention-to-treat basis in intensive care unit patients with established sepsis. 41,42These monoclonal antibodies, however, were tested in patients with established sepsis and organ failure, which is an entirely different setting than elective surgical patients who are more likely to benefit from prophylaxis with other endotoxin-related strategies. Moreover, these anti–lipid A monoclonal antibodies do not bind to endotoxin with high affinity, 43which may also explain their lack of demonstrable efficacy.

In summary, low preoperative endogenous antiendotoxin antibody concentrations were associated with death and/or prolonged hospitalization after routine noncardiac surgery independent of the POSSUM physiologic risk score and overall humoral immunity. Our findings suggest that endotoxemia may be a cause of postoperative morbidity after routine noncardiac surgery.

The authors thank Sabera Hossain, M.Sc., and Seth Strope, B.S. (both Research Assistants, The Mount Sinai School of Medicine, New York, NY), for assistance in conducting this study.