Blood Transfusion is Associated With Increased Resource Utilisation, Morbidity, and Mortality in Cardiac Surgery

Bharathi H Scott, Frank C Seifert, Roger Grimson 
Departments of Anesthesiology and Surgery, SUNY at Stony Brook, New York, USA

Correspondence Address:
Bharathi H Scott
Department of Anesthesiology, SUNY at Stony Brook Health Sciences Center, L4-060 Stony Brook, NY 11794-8480
USA

Materials and Methods

This study was approved by the institutional review board for collecting data from the patient’s medical records. This is a retrospective observational study. The study population included all patients undergoing primary CABG at a tertiary care heart centre over a period of three years. There were a total of 1746 patients. All patients underwent CABG surgery through a median sternotomy incision. Aspirin was discontinued 7 d prior to elective surgery.

One day before surgery, the patients in the hospital were premedicated with morphine and scopolamine prior to their arrival in the operating room. Patients arriving on the day of surgery were premedicated with midazolam as deemed appropriate by the anaesthesiologist. Intraoperative monitors included the standard American Society of Anesthesiologists monitors along with 5-lead electrocardiogram with continuous automated ST segment analysis, continuous arterial pressure determination using a radial artery catheter and a pulmonary artery catheter. All patients received standard fast-track anaesthetic induction using a combination of 5-10 µg/kg fentanyl and 0.05-0.1 mg/kg midazolam. We used 0.1 mg/kg vecuronium or 1 mg/kg rocuronium for neuromuscular blockade. Anaesthesia was maintained with supplements of propofol, isoflurane or sevoflurane as deemed appropriate by the anaesthesiologist. Patients undergoing CABG under CPB received 300 units/kg of heparin, and the activated clotting time (ACT) was maintained above 400 s prior to institution of cardiopulmonary bypass. Patients undergoing off-pump coronary artery bypass (OPCAB) received 100 to 200 units/kg of heparin to maintain the ACT above 350 s. Both groups received additional heparin as deemed necessary to maintain the ACT at the desired levels. The left internal mammary artery was harvested as surgically indicated and the saphenous vein was used for additional conduits. Patients undergoing CABG under cardiopulmonary bypass (CPB) were cannulated with an aortic cannula and a two-stage venous cannula prior to institution of cardiopulmonary bypass. Membrane oxygenators were used in all patients. Both retrograde and antegrade intermittent cold (6 °C) blood cardioplegia solution was administered for myocardial protection. Systemic temperature was permitted to maintain at 30-32 °C during cardiopulmonary bypass. All distal and proximal anastomoses were constructed using a single cross-clamp technique. In patients undergoing OPCAB, special off-pump coronary retractors and stabilisers were utilised during distal coronary anastomosis (Guidant, Indianapolis, IN). Proximal anastomoses were constructed using multiple cross-clamp technique. All patients were actively rewarmed to 37 °C prior to weaning from CPB. All patients undergoing CABG under CPB received epsilon aminocaproic acid during surgery. No antifibrinolytics were used in the OPCAB patients. The guidelines we followed for blood transfusion was triggered by the hematocrit values, ventricular function and other associated comorbidities. The end points of blood transfusion were to maintain a hematocrit value over 25% in patients with normal haemodynamics and ventricular function and over 30% in patients with compromised haemodynamics and ventricular function. These guidelines applied to all patients.

All demographic and clinical data were collected on the standardized data collection forms as required by the New York State cardiac surgical reporting system for predicting the risk-adjusted mortality rate according to the New York State department of health. This included age, gender, weight, body surface area (BSA), use of intraaortic balloon pump, left ventricular ejection fraction, use of cardiopulmonary bypass pump, number of distal coronary grafts, comorbid conditions such as diabetes, hypertension, congestive heart failure (CHF), renal failure, chronic obstructive pulmonary disease (COPD), previous myocardial infarction and postoperative bleeding requiring reoperation. The data pertaining to pre- and post-operative hematocrit values, the time to postoperative tracheal extubation, PRBC transfusion, ICULOS, PLOS and mortality and morbidity were obtained from the medical records of the patients. Extubation time is defined as the time from leaving the operating room to the removal of the endotracheal tube. ICULOS is defined as the time of admission to the ICU until the time of discharge to the intermediate care or step down unit. PLOS is defined as the time from the day of surgery until the day of discharge from the hospital. All patients were assessed for extubation within the first hour of arrival in the ICU and placed on our standard extubation protocol. The data pertaining to postoperative complications such as stroke, renal failure and 30-day mortality were also collected.

Statistical methods

Data are summarized and described using means, standard deviations and proportions. The demographic, clinical and outcome variables were compared between the transfused and nontransfused patients using the Wilcoxon rank sum test. The P values for these comparisons are reported. A P value of P 72 h of mechanical ventilation ( P P P P values for partial correlation coefficients between transfusion and time to extubation, ICULOS, PLOS and mortality were P values to be [3],[4] Our data also showed that 3.6% of the transfused patients remained intubated for longer than 72 h; however, in the nontransfused group, no patient remained intubated for longer than 72 h. This is an important utilisation factor because early tracheal extubation has been shown to decrease the cost of surgery. [5] We examined ICULOS as the subsequent resource. We believe that this resource generally follows a trend similar to the time to extubation. Our data shows that the transfused patients had significantly longer ICULOS similar to that in extubation. This fact is in agreement with those from previous reports. [3],[4],[5],[6],[7] With regard to PLOS, the transfused patients had significantly longer PLOS as compared to their nontransfused cohorts. Our results are in agreement with those of previous reports. [3],[6],[7] Our study shows that the incidence of postoperative complications was significantly higher in the transfused group. They had a significantly higher incidence of postoperative renal failure, stroke, infection and GI complications [3],[8],[9] and were reexplored for postoperative bleeding more often. This data is in agreement with previously reported studies. [8],[10] We examined the 30-day mortality and found that patients who received blood transfusion had significantly higher incidence of 30-day mortality. No deaths occurred in the nontransfused group. Previous studies have examined mortality in the transfused CABG patients and a majority of these reports have documented the deleterious effects of transfusion on long-term mortality. Our results are in agreement with those of previous studies. [7],[8],[10],[11],[12] We further examined the relationship between the number of units of PRBCs transfused and PLOS. We found that PLOS increased as the number of units of blood administered increased in a linear fashion. The length of stay increased by approximately 0.82 d for every unit of blood transfused. This is in agreement with a previously published report examining the relation between PLOS and number of units of PRBCs transfused. [6]

It is important to note that our transfused patients were older, smaller and had lower preoperative hematocrit values. The transfused group comprised significantly more females. All these are the known risk factors for transfusion. [2],[13],[14] The transfused group also had significantly more preoperative comorbidities. They had a significantly higher incidence of preoperative diabetes, hypertension, renal failure, CHF, COPD and PVD. They also had a lower preoperative hematocrit value. This is similar to what has been reported in previous studies. [8],[10] The question is “is it the higher incidence of preoperative comorbidities that is responsible for the increased incidence of postoperative morbidity and mortality or is transfusion a marker for the severity of illness.” Our analysis shows that transfusion is an independent predictor of increased resource utilisation because even in patients with no preoperative morbidities, transfusion correlated with increased duration of intubation, ICULOS and PLOS. Among these patients, PLOS increased with the number of units of PRBCs transfused. Some of the previous reports have encountered the same question and have concluded that transfusion is an independent predictor of early and late mortality and increased ICULOS and PLOS. [3],[4],[6] Our results are in agreement with these previous reports.

In conclusion, our data show that patients who received transfusion had significantly longer time to extubation, ICULOS and PLOS. They also had significantly higher incidence of the 30-day mortality. We found that PLOS increased with the number of units of PRBCs transfused.

References