Colorectal cancer (CRC) is the third most common cancer and the most frequent cancer presenting in the elderly.¹ Each year there are almost 1 million new cases and 500,000 deaths from colorectal cancer worldwide.¹ Colorectal cancer is a common cause of anemia, contributing to a higher incidence of perioperative blood transfusions in these patients.² Peri-operative blood transfusions ranging from 10%−68% have been reported for patients undergoing colorectal cancer resection.³ Blood transfusions have been associated with tumor recurrence and poorer survival through immunomodulation, with many theories of the mechanisms presented in the current literature.

A literature review was conducted and 11 relevant studies analyzed (see Table 1).³⁻¹⁵ Current findings in the literature remain controversial. While most studies agree that nontransfused patients had better outcomes, the magnitude of the association of immunomodulation to blood transfusions, and whether the association is a causal one, remains speculative. Thus there is still impetus to establish whether any concrete causal relationship exists.

Table 1 Studies evaluating the association between survival, recurrence, and receiving blood transfusions during colorectal cancer resection

View Fullscreen

This is, to our knowledge, the largest study to date, which aims to determine whether receiving perioperative blood transfusions during curative colorectal cancer resection impacts overall survival, cancer-related survival, and cancer recurrence, as well as the independent outcome predictors.¹³

Materials and Methods

Statistical Analysis

Data from 1370 patients who underwent curative colorectal cancer surgery were analyzed using univariate and multivariate analyses to identify factors significantly associated with overall survival, cancer-related survival, and disease-free survival. All data was displayed as categoric variables only. Chi squared (Fisher exact and Pearson's) testing was used to compare transfused and nontransfused groups and predicted prognostic factors. A P value of <0.05 was considered statistically significant. The Kaplan-Meier model and log rank test was used for univariate analysis of the outcome measures. Five year and median survival probabilities were recorded. The Cox regression model was used for multivariate survival analysis. The logistic regression model was used to determine independent predictors of survival and recurrence. Forward and backward regression analysis of variables identified those with a P value <0.05 and these formed the provisional model. Remaining variables were re-entered into the model as potential confounders. A confounding variable was defined as one where the estimate changed by ≥10%. Factors found to be significant were assessed for 2-way interactions. Independent prognostic significance was presented as a relative risk (RR) with 95% confidence intervals. All statistical evaluations were conducted using the PASW statistics version 19.0 (2010 SPSS Inc; IM, Chicago, IL).

Results

Patients and tumor characteristics

Of the 1370 patients who received curative surgery for stage I to III colorectal cancer, 423 patients (30.9%) required transfusion and 947 patients (69.1%) did not. Characteristics of the transfused and nontransfused groups are listed and compared in Table 2, and shows that the two groups were well matched for demographics and cancer-related characteristics: 44.4% of transfused patients and 27.7% of nontransfused patients had 1 or more surgical complications (P < 0.001), and 24.3% of transfused patients and 13.8% of nontransfused patients had 1 or more medical complications (P < 0.001). More reoperations were required (13.2%) during the same admission for transfused patients compared to nontransfused patients (4.8%; Table 2).

Table 2 Characteristics of 947 nontransfused and 423 transfused patients—Chi squared test

View Fullscreen

Recurrence

Peri-operative blood transfusions were independently associated with earlier cancer recurrence. This result remained significant even when subclassifying patients according to whether they had a transfusion pre-operatively, intra-operatively or postoperatively. Overall peri-operatively, 63.8% of patients in the transfused group and 68.6% in the non-transfused group developed recurrence at 5 years (P = 0.024) [see Table 3]. Preoperative hemoglobin was not associated with poorer disease free survival. Multivariate analysis found that sex, stage, lymph node involvement, grade, urgency of procedure, reoperation, perioperative transfusion including postoperative and intraoperative transfusions, were also independently related to recurrence.

Table 3 Factors related to recurrence in patients with stage I to III colorectal cancer

View Fullscreen

Overall survival

On univariate analysis the overall survival was 35.2% in the transfused patients, and 43.8% in the nontransfused patients (P = 0.001; Table 4). Multivariate analysis found that age, sex, stage, lymph node involvement, grade, urgency of procedure, medical complications, reoperation, and intra-operative transfusion were all independently related to overall survival. Intraoperative transfusion recipients had a median survival time of 50 months while those who were not transfused or transfused at a different time had a median survival time of 78 and 77 months respectively. This was found to be statistically significant on multivariate analysis (RR = 1.31, P = 0.004). Pre-operative hemoglobin, peri-operative transfusion, postoperative transfusion, and number of units transfused were not found to be significantly associated with overall survival on multivariate analysis.

Table 4 Factors related to overall survival in patients with stage I to III CRC

View Fullscreen

Cancer-related survival

On univariate analysis the cancer-related survival in transfused and nontransfused patients was 69.7% and 73% respectively (P = 0.051; Table 5). Multivariate analysis found that sex, stage, lymph node involvement, grade, urgency of procedure, medical complications, and intra-operative transfusion were independently associated with cancer-related survival. Pre-operative hemoglobin, peri-operative transfusion, postoperative transfusion, age, and re-operation, were not found to be significantly associated with cancer-related survival on multivariate analysis.

Table 5 Factors related to cancer related survival in patients with stage I to III colorectal cancer

View Fullscreen

Discussion

Transfusing patients peri-operatively for curative colorectal cancer resection is often done based on immediate clinical and biochemical indicators, with the aim of resuscitation and decreasing peri-operative mortality and morbidity. The indications for blood transfusions have been well established in the literature and include acute or subacute hemorrhage, severe anemia, hemolytic crisis, as well as applying good clinical judgment.^2,7,16 However, the impact of the transfusion on the patient's cancer-related prognosis might not be considered or discussed with the patient.

This is one of the largest studies to date and supports a growing body of evidence suggesting that peri-operative transfusions potentially have a negative impact on cancer related outcomes, particularly cancer recurrence. This finding is important, as alternative resuscitation methods and earlier preoperative optimization may need to be considered more carefully in patients with colorectal cancer.¹⁶ Closer followup protocols may need to be considered in transfused patients.

Blood transfusions have been associated with tumor recurrence and poorer survival possibly due to immunologic factors and have been termed transfusion-related immunomodulation.^17,18 The current literature offers numerous theories as to the exact mechanism of this immunomodulatory effect. Figure 1 summarizes the cell-mediated processes and humoral components involved. Most studies conclude that host down-regulation of immune function by blood transfusion allows cancer cells to escape the immune surveillance system causing micrometastases and recurrence, which then impacts the clinical course of the disease.¹⁹

View Figure

• Download as Powerpoint
• Download Figure

The studies listed in Table 1 have evaluated the effects of perioperative blood transfusions on patients with colorectal cancer,³⁻¹⁵ 9 studies^{5,7−11,13−15} assessed the effect of blood transfusions on cancer recurrence and 5 studies^5,11,13−15 found that it was significant. However, only 7 of these studies^{5,7−10,14,15} used multivariate analysis and of these studies, 3 found an independent association^5,14,15 (Table 6). Eleven studies^{3–10,12–14} tested the impact of blood transfusions on overall survival. Of these, 6 studies^{4–6,9,10,14} found an independent association (Table 1).

Table 6 Studies relating to transfusion and recurrence

View Fullscreen

We found that peri-operative blood transfusions were significantly associated with poorer overall survival on univariate alone. Its loss of significance on multivariate analysis suggests that the impact on survival is likely a result of confounding factors. This suggests that the phenomenon could be due to the circumstances that necessitate transfusion rather than the transfusion itself. Tatter et al (1992) found a more significant association as the number of units transfused increased; however, this study did not find any significant “dose-related association.”¹⁵ Nine studies evaluated the effect of blood transfusion on cancer-related survival and 4 of these found an association; however, only 2 of these were multivariate analysis.^{3−5,7,08,10,12−14}

Complications were higher in the transfused group, and this is consistent with most other studies.^3,5,7−9 The fact that some studies have found postoperative infections to increase with postoperative transfusions only, has led to suggestions that the effect of transfusions is short term as well as long term with recurrence.^4,5,7,8,12 One study found that preoperative transfusion did not increase wound infections after elective colorectal resection; however, this may not apply to this study's sample as preoperative transfusion was associated with lower preoperative hemoglobin and overall poorer physiologic reserve.⁵ Mystner et al (2000) found that transfusions were not significantly associated with reduced survival, but were significantly associated with recurrence, and attributed this mainly to the association between transfusions and postoperative infections.⁸ A fatal infectious complication is anastomotic failure, which can also lead to poorer survival and recurrence and has been included in this study as an unmeasured outcome as part of the surgical complications.

Limitations of this study include retrospective data and inclusion of patients between a 20-year period during which time blood product processing has markedly changed. Blood products are now mostly separated rather than pooled, have better infection screening, and superior cross-matching technology. Changing surgical technique and standards of practice over time must also be considered. Randomization is not possible between transfused and nontransfused patients and therefore selection bias will always exist. Survival outcomes based on the type of blood product given were not analyzed as some patients would have received whole blood during the earlier years of this study.

Conclusion

In conclusion, peri-operative blood transfusion in this study was found to increase the likelihood of colorectal cancer recurrence but was not found to be associated with overall survival or cancer-related survival. This study contributes to a body of evidence that blood transfusions may be associated with poorer outcomes and recurrence, and is significant enough to indicate caution. Where the aim is curative resection of colorectal cancer, this study emphasizes the importance of administering blood products with discretion and after considering the short and long-term risks and benefits.