Excessive Bleeding and Transfusion in a Prior Cardiac Surgery is Associated with Excessive Bleeding and Transfusion in the Next Surgery

Abstract

If there is a genetic predisposition to excessive bleeding, there should be an association in excessive blood loss between multiple cardiac surgeries. We retrospectively determined in 174 patients the association of excessive bleeding between 2 cardiac surgeries with cardiopulmonary bypass between January 19, 1990 and June 25, 2002. Excessive bleeding was defined by 2 criteria: (a) postoperating room chest tube blood loss over 24 h more than or equal to 750 mL (chest tube drainage [CTD] ≥ 750) and (b) transfusion of any non-red blood cell (RBC) blood products. Logistic regression was used to estimate the association between excessive bleeding at the first and second cardiac procedures. The logistic regression models for CTD ≥ 750 in the second surgery determined that CTD ≥ 750 in the first surgery compared to CTD < 750 had an unadjusted odds ratio of 2.18 (P = 0.03) and an odds ratio of 2.42 (P = 0.03) when adjusted for age, sex, body surface area, preoperative anticoagulant use, cardiopulmonary bypass duration, and procedure type at second surgery. The logistic regression model for any non-RBC use in the second surgery determined that any non-RBC use in the first surgery compared with no non-RBC use had an unadjusted odds ratio of 2.32 (P = 0.02) and an odds ratio of 2.55 (P = 0.02) when adjusted for age, sex, body surface area, preoperative anticoagulant use, cardiopulmonary bypass duration, and procedure type at second surgery. We conclude that a history of excessive bleeding during the first operation is associated with more than two times increased risk for excessive bleeding in the second surgery.

In Brief

IMPLICATIONS: A history of excessive bleeding and blood transfusion in the first cardiac surgery is significantly associated with excessive bleeding and blood transfusion in a subsequent cardiac surgery.

There is a long-held belief that if a patient has a history of excessive bleeding with a surgical procedure such as tooth extraction, then they are more likely to have excessive bleeding with a subsequent surgical procedure. This is certainly true for the major known genetic deficiencies in coagulation, for example, the hemophilias, but these are well described and often well documented in the patient’s history. For patients without major known genetic defects in coagulation, this belief implies that there is either a genetic or environmental factor that predisposes the patient to excessive bleeding.

Recently, multiple genetic polymorphisms have been identified that may increase a patients risk for bleeding after cardiac surgery with cardiopulmonary bypass (CPB) (¹). Also, genetic polymorphisms, such as factor V Leiden, have been described that increase a patient’s risk for venous or arterial thrombosis (^2,3). There is a suggestion that the factor V Leiden may be associated with earlier saphenous vein graft occlusion after coronary artery bypass grafting (CABG) (⁴).

The etiology of excessive blood loss with cardiac surgery is multifactorial and may be related to medications, history of preexisting coagulopathy, surgical procedure, or other unrecognized factors (⁵). The significance of excessive blood loss is the subsequent need for the transfusion of allogeneic blood products (^6,7). This ultimately places the patient at risk for transmitted diseases, transfusion reactions, and other complications (⁸). Excessive blood loss may also lead to increased surgical closure time, anesthesia time, and overall expense.

Our hypothesis is that excessive bleeding in the first surgery associates with excessive bleeding in the second surgery. If this is true, it may imply that certain subpopulations of patients may have subtle, as yet undescribed, genetic deficiencies in coagulation that only become expressed during major surgical procedures with substantial blood loss, such as cardiac surgery with CPB.

Methods

After IRB approval, we performed a retrospective chart review of patients that underwent 2 cardiac surgeries with CPB during the study period between January 19, 1990 and June 25, 2002. Based on the criteria that the first cardiac surgery was the primary sternotomy, followed by another cardiac surgery, a minimum of 3 mo subsequent to the first surgery, 174 patients older than 16 yr were included in the study. Patients younger than 16 yr were excluded because pediatric patients undergoing cardiac surgery and CPB have many special considerations such as significant dilutional effects of CPB, immature coagulation systems, and a higher probability of cyanotic heart disease than adult patients (^9,10). All types of cardiac operations from any cardiac surgeon at our institution were included.

The primary end-points characterizing excessive bleeding were defined as (a) chest tube blood loss over 24 h more than or equal to 750 mL (chest tube drainage [CTD] ≥ 750) and (b) the subsequent transfusion of any non-red blood cell (RBC) blood products in the operating room after CPB, including platelets, fresh frozen plasma (FFP), and cryoprecipitate. This threshold value of 750 mL corresponds to the approximate 75th percentile of CTD values at the second surgery and is consistent with values used in other studies. The objective of this study was to determine if those patients who experienced excessive bleeding in their first cardiac surgery were more likely to have excessive bleeding during a subsequent surgery. Separate analyses were conducted for each excessive bleeding variable, CTD ≥ 750, and any use of non-RBC blood products. Logistic regression was used to determine the association between excessive bleeding at the first and second cardiac procedures. A multivariable logistic regression model was developed to determine whether excessive bleeding during the first procedure was independently associated with excessive bleeding during the second procedure. Variables known or presumed to be associated with excessive bleeding were included as adjustors in the multivariable analysis. Odds ratios (ORs) with corresponding 95% confidence intervals and P values are reported. In all cases, P values ≤ 0.05 were considered statistically significant. Demographic, surgical, blood loss, and transfusion characteristics are reported within surgery (first versus second) across CTD groups (<750 versus ≥750). Continuous variables are reported as medians and interquartile ranges (IQR); categorical variables are reported as counts and percentages. The medications noted as preoperative anticoagulants included heparin, aspirin, Coumadin, and clopidogrel. Red cell mass was defined as blood volume multiplied by hematocrit, where blood volume was calculated with adjustment for height, weight, and sex (¹¹).

Results

There were 174 patients who had multiple cardiac operations during the specified time period of this study. None of the patients had a documented history of preexisting conditions affecting coagulation such as hemophilia, von Willebrand disease, or Glanzmann thrombasthenia. None of the patients had been exposed to glycoprotein(GP) IIb/IIIa receptor antagonists or clopidogrel before cardiac surgery.

Of the 174 patients (119 men and 55 women) who underwent more than one cardiac operation after the age of 16 yr, the median age at the time of the first operation was 63 yr (IQR, 43–70). For the first operation, the median CTD for the first 24 h was 670 mL (IQR, 401–945). There were 77 patients (44%) who were transfused one or more units of RBC (range, 1–10 U), 37 patients (21%) who received FFP (range, 2–12 U), 39 patients (22%) who received platelets (range, 1–18 U), and 11 patients (6%) who received cryoprecipitate (range, 1–25 U). Overall, there were 44 patients (25%) who received one or more non-RBC blood products.

At the time of the second surgery, the median age was 67 yr (IQR, 46–74). The median time from the first operation to the first subsequent procedure was 3.4 yr (range, 0.3–11.1 yr). For the second procedure, the median CTD for the first 24 h was 492 (IQR, 335–760). There were 93 patients (53%) who were transfused one or more units of RBC (range, 1–16 U), 59 patients (34%) who received FFP (range, 1–32 U), 68 patients (39%) who received platelets (range, 1–48 U), and 10 patients (6%) who received cryoprecipitate (range, 1–16 U). Overall, there were 72 patients (41%) who received one or more non-RBC blood products.

The demographic and surgical characteristics for the first and second surgeries presented across CTD groups are shown in Table 1. Compared with patients with CTD < 750, patients who had CTD ≥ 750 in the first surgery were older, more likely to be men, more likely to have had a CABG with or without a valve procedure, had a longer duration of CPB, lower temperature on CPB, and larger red cell mass. For the second surgery, the patients who had CTD ≥ 750 were more likely to have had a CABG with or without a valve procedure and longer duration of CPB.

Table 1: Demographic and Surgical Characteristics

The blood loss and transfusion characteristics for the first and second surgeries presented across CTD groups are shown in Table 2. Compared with patients with CTD < 750, patients who had CTD ≥ 750 in the first surgery also had more CTD at 4 and 12 h and a more RBC, platelet, FFP, and cryoprecipitate transfusions. For the second surgery, the patients who had CTD ≥ 750 also had more CTD at 4 and 12 h and more RBC, platelet, and FFP transfusions. The patients who had CTD ≥ 750 in the second surgery also had more CTD in the first surgery, and a larger percentage of patients had CTD ≥ 750 in the first surgery.

Table 2: Blood Loss and Transfusion

The logistic regression models for blood loss outcomes at the second surgery are presented in Table 3. The logistic regression model for CTD ≥ 750 in the second surgery determined that, compared to CTD < 750 at the first surgery, CTD ≥ 750 in the first surgery had an OR of 2.18 (P = 0.03) and an OR of 2.42 (P = 0.03) when adjusted for age, sex, body surface area, preoperative anticoagulant use, CPB duration, and procedure type at the second surgery. Compared with patients not receiving a non-RBC transfusion at the first surgery, those patients who received any non-RBC transfusion in the first surgery were more likely to receive any non-RBC transfusion in the second surgery (P = 0.02). There were 44 patients who received a non-RBC transfusion in their first surgery; of those patients, 25 (57%) received non-RBC transfusion in the second surgery. In the first surgery, there were 130 patients who did not receive a non-RBC transfusion; of those, 47 (36%) did receive a non-RBC transfusion in the second surgery. The logistic regression models for any non-RBC use in the second surgery determined that, compared with no non-RBC use in the first surgery, any non-RBC use in the first surgery had an OR of 2.32 (P = 0.02) and an OR of 2.55 (P = 0.02) when adjusted for age, sex, body surface area, preoperative anticoagulant use, CPB duration, and procedure type at the second surgery.

Table 3: Logistic Regression Models for Blood Loss Outcomes at Second Surgery

Discussion

We found that a history of excessive bleeding during the first operation is associated with a more than two times increased risk for excessive bleeding in the second surgery. This held true for two distinct definitions of excessive bleeding, as characterized by both CTD over 24 hours more than or equal to 750 mL and the administration of any non-RBC blood products. This is the first attempt to assess whether a patient’s likelihood of excessive bleeding with subsequent operations is associated with the amount of bleeding in the initial operation. There is one report by Heit et al. (¹²) (in abstract form) demonstrating that women undergoing hysterectomy for menorrhagia did not have more bleeding in subsequent surgical procedures than women without menorrhagia.

There is a growing body of knowledge of the genetics of the coagulation system, especially for genetic polymorphisms, that create a procoagulant state. The polymorphism that results in factor V Leiden is the most commonly known inherited risk factor for deep venous thrombosis (²). A recent prospective observational study of cardiac surgical patients found that patients with factor V Leiden had significantly small CTD than noncarriers (¹³). Those investigators also performed a multivariate regression analysis and determined that factor V Leiden was an independent contributor to CTD but not to the number of units of blood transfused. Finally, they found that factor V Leiden had a significant independent protective effect on the risk of receiving a transfusion during hospitalization.

Recently, Welsby et al. (¹) identified seven genetic polymorphisms associated with bleeding after cardiac surgery. These polymorphisms were GPIaIIa-52C>T and 807C>T, GPIbα 524C>T, tissue factor-603A>G, prothrombin 20210G>A, tissue factor pathway inhibitor-399C>T, and angiotensin converting enzyme deletion and insertion. These genetic factors seemed to be primarily independent of, and explain at least as much variation in, bleeding as clinical covariates. They found that combining genetic and clinical factors doubled their ability to predict bleeding after cardiac surgery, and they noted that accounting for genotype may be required when stratifying risk of bleeding after cardiac surgery. Our OR results of approximately 2 are consistent with the doubling of bleeding risk after cardiac surgery.

This investigation was performed as a preliminary study to identify a patient population that may have genetic polymorphisms that increase the risk of bleeding. We plan to perform subsequent studies to determine if there are new genetic polymorphisms, especially those found by Welsby et al. that may explain the differences found in this study. The primary defect in coagulation associated with CPB is thought to be a deficiency in platelet function (^14,15). Genetic variance of the platelet GP IIb/IIIa receptor has been demonstrated with important clinical implications in cardiovascular mortality and drug effects (^16–18). A study (¹⁹) has found a correlation between postoperative blood loss in cardiac surgical patients and genetic variance of the platelet GP Ib receptor. Another study has demonstrated, by multivariate analysis, that the number of tissue factor – 1208 insertion alleles and the total number of GP Ib repeats were independent contributors to the age at which patients require CABG surgery (²⁰).

Identifying a subset of patients who are at increased risk for bleeding and for transfusion of non-RBC products would allow additional studies to identify the cause of their bleeding. This could result in a significant improvement in patient care and the use of limited blood resources. However, an unintended consequence of this study may be an increase in transfusion behavior because of a patient’s history of bleeding. There are ranges of blood conservation strategies to prevent bleeding; however, some are expensive (²¹). Determining which patients are at highest risk for bleeding would help direct the correct conservation therapy to these high-risk patients increasing efficiency and care. One of the conservation therapies is prophylactic use of antifibrinolytic therapy. A risk of antifibrinolytic therapy is thrombosis. A study by Moor et al. (⁴) suggested that patients with factor V Leiden may be at increased risk of early saphenous vein thrombosis. Further, factor V Leiden has been implicated as a confounding factor in graft thrombosis with antifibrinolytic drugs (²²). Identification of those patients who would most benefit from antifibrinolytic therapy could hopefully reduce the risk of thrombosis with the use of these drugs.

There are multiple limitations to our study. It is possible, although unlikely, that the anesthesiologists caring for the patients in the second surgery were aware of the presence or absence of excessive bleeding in the first surgery. This may have affected their transfusion behavior in the second surgery, but it should not have affected the amount of blood loss in the second surgery. Patients tend to stay with the same surgeon for their subsequent surgery. There were 11 surgeons and 28 anesthesiologists who participated in the procedures studied. We note that 77% of patients had the same surgeon at both first and second procedures, whereas only 18% had the same anesthesiologist at the first and second procedures. We were limited in the number of adjustor variables we could include in our multivariable model because of our sample size, so we could not consider the effect of anesthesiologist and surgeon identity on transfusion behavior. If the surgeons differed significantly in their transfusion behavior, this may explain the association seen in this study. We have studied our surgeons’ transfusion behaviors in previous continuous improvement studies and have not found statistically significant differences. Use of antifibrinolytic drugs was not found to be significantly associated with excessive bleeding in the second surgery. This may be explained by the observed trend of longer CPB duration in those patients receiving antifibrinolytic drugs (data not shown). Thus, the longer duration of CPB may have negated the protective effects of the antifibrinolytic drugs with respect to blood loss. Finally, our definition of excessive CTD was a small value (750 mL). We analyzed the data using more than 1000-mL CTD as a definition of excessive bleeding. There were 39 (22%) patients who had CTD >1000 mL after their first surgery and 31 (18%) patients who had CTD >1000 mL after their second surgery. Although statistical power was limited by the small number of patients with CTD >1000 mL after their second surgery, univariate and multivariable logistic regression analyses were performed to assess whether CTD >1000 mL at the first surgery was predictive of CTD >1000 mL at the second surgery. From these analyses, the variable coefficient for CTD >1000 mL was not found to be statistically significant; however, the magnitude of the association from both univariate (OR = 1.87; 95% confidence interval, 0.80–4.41; P = 0.15) and multivariate (OR = 1.91; 95% confidence interval, 0.75–4.89) analyses was similar to that observed when CTD >750 mL is used to define excessive bleeding.

Given the retrospective nature of this investigation, we cannot determine whether there are truly intrinsic patient or procedural factors that explain the apparent association of excessive bleeding with multiple operations or if our results may be secondary to environmental factors, such as herbal remedies, for which we were not able to account. We hope to contact these patients and obtain blood samples to prospectively investigate whether a genetic polymorphism or polymorphisms was responsible for their excessive bleeding. Another possible explanation for our results is that previous exposure to allogeneic platelet transfusion may expose the patients to platelet human leukocyte antigens. One study demonstrated an increase in blood transfusions in liver transplant patients who had alloimmunization from previous platelet transfusions and increased panel reacting antibodies (²³). Given these limitations, we are restricted in our conclusions.

A literature search revealed that this is most likely the first and only study to suggest that excessive bleeding during the first cardiac surgery is associated with excessive bleeding in subsequent cardiac surgeries. One possible explanation for these results may be that there is a subset of the patient population with an inherent bleeding disorder. This disorder would be subclinical and not detected with standard coagulation tests but becomes active in response to coagulopathy associated with cardiac surgery requiring CPB. We found that a history of excessive bleeding during the first operation was associated with a more than two times increased risk for excessive bleeding in the second surgery.