| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2003;126:755-760
© 2003 The American Association for Thoracic Surgery
General thoracic surgery |
a Departments ofDepartment of Thoracic Surgery, Medical University of Gdañsk, Gdañsk, Poland
b Department of Oncology and Radiotherapy, Medical University of Gdañsk, Gdañsk, Poland
c Department of Pneumonology, Medical University of Gdansk, Gdansk, Poland
Received for publication October 1, 2002; revisions received November 5, 2002; revisions received November 28, 2002; accepted for publication January 7, 2003.
* Address for reprints: Witold Rzyman, MD, PhD, Department of Thoracic Surgery, Medical University of Gdansk, 7 Dêbinki St, 80-211 Gdansk, Poland
wrzyman{at}wp.pl
 |
Abstract |
|---|
 Top Abstract Materials and methodsResultsDiscussionReferences |
|---|
OBJECTIVE: To determine the effect of perioperative allogeneic and autologous blood transfusions on survival in non–small cell lung cancer patients treated with curative pulmonary resection.
METHODS: Of 493 consecutive patients, 185 (37.5%) received allogeneic blood products and 145 (29.4%) received autologous blood products, whereas 163 patients (33.1%) received no blood products. Survival analysis included univariate log-rank test and multivariate Cox regression model.
RESULTS: Three-year survival probabilities in allogeneic, autologous, and nontransfused groups were 40%, 48%, and 61%, respectively, and the estimated 5-year survival probabilities were 34%, 38%, and 48%, respectively. In the univariate analysis there was a reduced survival in allogeneic versus nontransfused group (P < .01). In the multivariate analysis stage (P < .01), initial hemoglobin level (P = .01) and sedimentation rate (P = .03) remained significant factors, whereas the type of blood transfusion (allogeneic versus autologous) was no longer significant. There was no significant impact of transfusion in the multivariate analysis when patients in stage III or patients who underwent pneumonectomy were excluded.
CONCLUSION: Neither allogeneic nor autologous blood transfusion has independent, adverse survival impact in non–small cell lung cancer patients treated with radical pulmonary resection.
Due to ethical reasons, the impact of perioperative allogeneic blood transfusion on survival or the risk of relapse in cancer patients has never been assessed in a randomized prospective trial including random allocation to transfusion or no transfusion. Obviously, it would be unethical to transfuse blood without clinical indications or to withhold transfusion in patients who need it. Consequently, the entire knowledge on this issue is based on retrospective studies.
The immunosuppressive effects of allogeneic blood transfusion were analyzed in different types of tumors, most frequently in colorectal cancer. Landers and colleagues7 reviewed 36 retrospective studies that analyzed the influence of perioperative blood transfusion on the risk of relapse or survival in patients operated for different tumors. Fifteen studies showed no influence of transfusion and 21 demonstrated its detrimental effect. Of the 6 retrospective studies in lung cancer, 4 showed adverse effect of allogeneic blood transfusion8-11 and 2 did not.12,13
The current knowledge on the effects of autologous blood transfusion is relatively scarce. Transfusion of autologous, predeposited blood was not documented to cause significant immunosuppression in experimental models.14,15 Two studies in surgically treated colorectal cancer patients failed to show significant survival impact of allogeneic and autologous blood transfusion, but there was an increased incidence of postoperative infections in transfused patients.6,16,17 The prognostic impact of autologous blood transfusion has never been studied in surgically treated non–small cell lung cancer (NSCLC) patients.
This retrospective analysis assessed the prognostic relevance of allogeneic and autologous blood transfusions in a large series of NSCLC patients treated with curative pulmonary resection.
|
Materials and methods |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionReferences |
|---|
|
Methods
Patients who donated their own blood were scheduled to deposit 2 units of blood (450 mL/unit) between 6 and 3 days before surgery. The level of hemoglobin concentration above 12 g/dL was considered a minimal value for predeposition, whereas hemoglobin level below 8 g/dL accompanied by the evidence of inadequate tissue oxygenation (tachycardia, low blood pressure, etc) was the indication for perioperative autologous or allogeneic blood transfusion. In this analysis, only blood transfusions during hospitalization were considered. No patient received blood transfusion preoperatively. All patients received prophylactic antibiotics and anticoagulative agents. Transfusion material included whole blood or packed red blood cells. All autologous blood that was not retransfused to the donor was discarded. The operations were performed under general anesthesia with the same routine rules for all the patients. The surgical approach was standardized: the anatomical lung resection was performed first, followed by systemic complete dissection of hilar and mediastinal lymph nodes. Operations were performed by 8 surgeons: 4 specialized thoracic surgeons and 4 surgeons training to be thoracic surgeons due to the educational character of the department. Pathological assessment was done in 1 laboratory, and then pathological lung cancer staging status was established. In principle, adjuvant irradiation or chemotherapy was not administered; however, a few patients received radio- or chemotherapy according to individual decisions of the treating physicians. After the discharge from the hospital, all patients were followed up in the outpatient clinic every 3 months for 2 years and twice a year later on. Follow-up ranged from 26 to 73 months (median 46 months). Death certificates from provincial registry were used to confirm the death.
Statistical analysis
Survival probability and its 95% confidence interval (CI) were computed using the Kaplan and Meier method, with the date of the operation set as zero time. Postoperative 30-day deaths were included in the survival analysis. The log-rank statistic was used for the univariate statistical comparisons of survival curves. The Cox proportional hazard model was used to calculate hazard ratio and its 95% CI for particular variables in univariate and multivariate model. In the multivariate model, backward-stepwise regression was used and insignificant variables were removed from the model based on the Wald’s test. Additionally, a forward-stepwise model was employed to assure that the construction of the model would not influence the results.
|
Results |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionReferences |
|---|
Most of the analyzed variables, except for sedimentation rate, hemoglobin level, and lung cancer staging status, were similarly distributed in the analyzed groups (Table 1).
Median survival for autologous, allogeneic, and nontransfused groups were 2.9, 1.7, and 4.4 years, respectively (Figure 1). Three-year survival probabilities (±95 CI) were 48% (39-56%), 40% (33-47%), and 61% (53-68%), respectively, and 5-year survival probabilities were 38% (29-48%), 34% (27-41%), and 48% (37-59%), respectively.
|
|
There was a significant difference in survival between the entire group of transfused patients versus nontransfused patients (P = .017) in univariate analysis, but during the backward-stepwise regression in Cox model this variable was excluded as nonsignificant. The same results were obtained with forward-stepwise model, indicating that the type of the model does not influence the results.
The resections in NSCLC patients with stage IIIA and IIIB tumors are often considered as palliative management. Since the possible detrimental effect of blood transfusion is of particular importance for patients with early-stage disease, we have repeated the analysis on 314 patients with stage I and II tumors. In the univariate analysis in this subset there was no 5-year survival difference between the nontransfused and autologous group (P = .20), nontransfused and allogeneic group (P = .70), or autologous and allogeneic group (P = .09).
Finally, to eliminate the extent of resection as a cofactor, we excluded 168 patients who underwent pneumonectomy. This analysis, restricted to 325 patients who had either lobectomy or sublobar resection (wedge resections were excluded), showed significant survival difference in univariate analysis between nontransfused and allogeneic groups (P < .01). There was, however, no significant difference between nontransfused and autologous (P = .65) and autologous and allogeneic groups (P = .06). In the multivariate analysis including these 325 patients, only hemoglobin level and stage of disease remained significant (both P < .01), whereas prognostic impact of allogeneic transfusion was lost.
|
Discussion |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionReferences |
|---|
The data on the prognostic impact of allogeneic blood transfusion in patients operated for NSCLC are limited and provide contradictory results (Table 3). So far, there have been only 2 published studies including more than 300 patients.11,12 Often, a limited number of known prognostic factors have been included in the multivariate analysis. Furthermore, only 3 studies8,10,11 included information on initial hemoglobin or hematocrit levels—one of the most important factors determining the need of transfusion and linked to the disease stage and prognosis. In our study, initial hemoglobin level was analyzed and found to predict survival in both univariate and multivariate models. The statistical tests used to analyze the effect of blood transfusion have also varied among particular studies (Table 3). The most reliable test—adjusted hazard ratio and corresponding 95% confidence interval—has only occasionally been employed.
|
We demonstrated an adverse effect of allogeneic blood transfusion in univariate analysis but not in the multivariate model. There was also no independent effect of allogeneic transfusion when all patients with stage III tumors or patients who underwent pneumonectomy were excluded from analysis. Thus, allogeneic blood transfusion did not prove to be an independent negative predictor of survival when other prognostic variables were taken into account.
Neither in univariate nor in multivariate analysis did we find any significant survival differences between autologous versus allogeneic or between autologous versus nontransfused patients. The latter observation corresponds to the results of the only 2 studies addressing the impact of autologous blood transfusion,16,17 both performed in colorectal cancer.
In retrospective analysis the impact of a given variable on survival may be difficult to assess due to other confounding covariates. Thus, it is important to include as many known prognostic factors as possible and to employ multivariate rather than univariate statistical models. Nevertheless, a selection bias may occur even if all attempts are taken to balance the distribution of all known prognostic factors.
It has been suggested that a randomized clinical trial with adequate power to detect a relatively minor survival difference might give a definite answer to this relevant clinical problem.19 However, such a study does not seem feasible for a number of reasons. A major intraoperative or postoperative blood loss requiring allogeneic blood transfusion cannot be determined before randomization. Furthermore, unnecessary exposure to allogeneic blood may not be acceptable due to ethical reasons. Another approach, although not free of some potential biases of retrospective studies, is an appropriately designed case-control study.
The blood transfusion rate accompanying pulmonary resection varies in particular series from 36% to 55%.8-13 In our series the frequency of transfusions (67%) was higher than in most studies including operable NSCLC. This outcome is due to several factors including mainly frequent preoperative blood donation, relatively high incidence of squamous cell carcinoma typically located centrally and necessitating more extensive resection, relatively high percentage of stage III disease, and routinely performed complete mediastinal lymph node dissection.
In conclusion, our study failed to support the hypothesis that blood transfusion, be it allogeneic or autologous, has a detectable detrimental effect on survival. However, we will continue to avoid blood transfusions in cancer patients whenever possible due to other potential risks carried by this procedure. In our opinion, the indications for transfusion of allogeneic or autologous blood should be limited to low hemoglobin or hematocrit levels, combined with the clinical symptoms of inadequate tissue oxygenation.
|
References |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. Thomas, P. Michelet, F. Barlesi, X. Thirion, C. Doddoli, R. Giudicelli, and P. Fuentes Impact of blood transfusions on outcome after pneumonectomy for thoracic malignancies Eur. Respir. J., March 1, 2007; 29(3): 565 - 570. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
