The Research Electronic Data Capture (REDCap) system is supported by Center for Translational Science Activities (Grant UL1 TR000135). This research was carried out with departmental Small Grant. The funders played no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.
Allogeneic blood transfusion induces immunosuppression, and concern has been raised that it may increase propensity for cancer recurrence; however, these effects have not been confirmed. We examined the association of perioperative transfusion of allogeneic blood long-term oncologic outcomes in patients with prostate cancer who underwent prostatectomy.
Study Design and Methods
We reviewed medical records of patients who underwent radical prostatectomy between 1991 and 2005 and received allogeneic nonleukoreduced blood. Each transfused patient was matched to two controls who did not receive blood: matching included age, surgical year, prostate-specific antigen level, pathologic tumor stages, pathologic Gleason scores, and anesthetic type. Primary outcome was systemic tumor progression, with secondary outcomes of prostate cancer death and all-cause mortality. Stratified proportional hazards regression analysis was used to assess differences in outcomes between the transfused and nontransfused group.
A total of 379 prostatectomy patients who were transfused and 758 nontransfused controls were followed for 9.4 and 10.2 years (median), respectively. In a multivariable analysis that took into account the matched study design and adjusted for positive surgical margins and adjuvant therapies, the use of allogeneic blood was not associated with systemic tumor progression (hazard ratio [HR], 0.88; 95% confidence interval [CI], 0.39-1.99; p = 0.76), prostate cancer–specific death (HR, 1.69; 95% CI, 0.44 to 6.48; p = 0.44), or all-cause death (HR, 1.20; 95% CI, 0.87 to 1.67; p = 0.27).
When adjusted for clinicopathologic and procedural variables transfusion of allogeneic blood was not associated with systemic tumor progression and survival outcomes.
Blood transfusions (BTs) were used liberally in the past to improve perioperative tissue oxygen delivery, but the potential harm of BT has progressively gained increased attention.[1, 2] The risks of BT can be attributed to errors in blood component transfusions, transfusion-transmitted infections, and immunomodulatory effects. Transfusion-related immunomodulation (TRIM) encompasses the potential immunosuppressive and proinflammatory effects of BT. The exact mechanisms and mediators remain speculative, but the three proposed include alloimmunization, tolerance, and immunosuppression. Specifically, allogeneic blood has been found to be the major cause of TRIM due to constituents that mediate immunosuppression, such as allogeneic mononuclear cells, soluble biologic response modifiers, and soluble human leukocyte antigen (HLA) Class I peptides.
Gantt raised a concern that TRIM may lead to an increased rate of cancer recurrence in patients with malignancies. Subsequently, multiple reports with conflicting conclusions on the effects of perioperative BT on cancer outcomes were published.6-9 Several studies investigated the role of TRIM in colorectal cancers with a recent meta-analysis of 55 studies concluding that perioperative allogeneic BT increased the risk of cancer-related and all-cause death after colorectal cancer surgery. Several retrospective studies examined the role of perioperative BT on oncologic outcomes after prostate cancer surgery.7,11-18
Many of the studies did not match on tumor pathology, which is important for comparing long-term outcomes. Finally, none of these studies accounted for potential effects of anesthetic techniques. It has been suggested that neuraxial analgesia may reduce cancer recurrence and tumor progression by preserving natural killer cell function. These effects are believed to be mediated through attenuation of stress and reduction in systemic opioid use. We recently reported in a retrospective study that the use of opioid-sparing neuraxial analgesia was associated with reduced prostate cancer progression.
Our primary aim was to examine the association between perioperative allogeneic nonleukoreduced BT and prostate cancer progression in patients who underwent radical prostatectomy for adenocarcinoma. We hypothesized that after controlling for clinicopathologic and procedural characteristics, allogeneic BT would not be significantly associated with progression of prostate cancer.
Materials and Methods
This study received approval from the institutional review board of Mayo Clinic (Rochester, MN). Consistent with Minnesota Statute 144.335 Subd. 3a. (d), we included only patients who had provided authorization for research use of their medical records (historically > 95% of Mayo Clinic patients). To identify population of interest in this study we used resources of our institutional prostatectomy registry for the period between January 1, 1991, and December 31, 2005. This period was selected to reduce the potential for including the more advanced prostate cancers that were common before prostate-specific antigen (PSA) was introduced into clinical practice in the late 1980s and also to allow for sufficient follow-up for observing outcomes. Since inception of Mayo Clinic's radical prostatectomy registry in 1966, more than 20,000 patients were recorded and followed, making it one of the largest registries of this kind in the world. Abstractors prospectively maintain the registry by following up annually with patients after surgery. After prostatectomy at the Mayo Clinic patients are typically asked to have physical examinations and serum PSA measurement quarterly for the initial 2 postoperative years, semiannually for an additional 2 years, and annually thereafter. We recorded biochemical recurrence (defined as postoperative PSA concentrations ≥ 0.4 ng/mL), local recurrence (clinical examination or prostate bed needle biopsy), and systemic progression (presence of metastases on imaging or on biopsies from locations other than the prostatic bed). The compliance of this follow-up is exceptionally high, and registry has less than 5% lost to follow-up rate. Our dedicated research nurse follows up on all the recurrences to document secondary treatments and anyone who fails to send in his or her yearly follow-up letter.
For the period between January 1, 1991, and December 31, 2005, the prostatectomy registry, anesthesia database, and electronic medical records were used to identify all patients who received any type of neuraxial block (with or without general anesthesia) during hospitalization for radical retropubic prostatectomy, and these patients were 1:1 matched (age, date of surgery, preoperative PSA level, pathologic stage, and pathologic Gleason scores [n = 2555]) to patients who received general anesthesia only (n = 2555). Results from the primary analyses comparing oncologic outcomes between those who received neuraxial analgesia versus not are published elsewhere. For this study, we rereviewed this entire patient cohort (n = 5110) to identify those who received allogeneic BT during hospitalization. We excluded patients who received irradiated, leukoreduced, or washed allogeneic units since the number of patients who received these products was not large enough to warrant meaningful analyses. We also excluded patients who received only autologous blood products since these products have limited immunomodulatory potential and therefore are not pertinent to our scientific question.
For each patient who received allogeneic blood, we selected two control patients (1:2) who did not receive any blood product during the perioperative period. These patient sets were matched on age (±5 years), type of anesthesia (general anesthesia, general anesthesia supplemented by neuroaxial analgesia, and neuroaxial anesthesia only), date of surgery (±1 year), preoperative PSA level, pathologic stage (T2, T3/4, TxN+), and pathologic Gleason scores (≤6, 7, ≥8) using an optimal matching algorithm method.
Data were abstracted from the electronic and paper medical records and entered manually into the web-based Research Electronic Data Capture (REDCap) system (Version 3.6.7, Vanderbilt University, Nashville, TN). Patients' medical records were reviewed to obtain age, body mass index, preoperative PSA, American Society of Anesthesiologists physical status, and comorbidities, which were ascertained using previously described definitions including coronary artery disease, hypertension, pulmonary disease (obstructive or restrictive chronic lung disease), diabetes mellitus, cerebrovascular disease, peripheral vascular disease, severe kidney disease, and history of other cancers. The surgical, anesthetic, and blood bank records were reviewed to determine anesthetic technique and type and amount of blood products transfused. The number of units transfused and for each unit the exact product (red blood cells [RBCs], fresh-frozen plasma [FFP], platelets [PLTs], and cryoprecipitate), and source (autologous, allogeneic) as well as any other blood processing techniques used (leukoreduction, irradiation, or washing blood) was recorded. We also recorded the type and amount of opioids used intraoperatively, in the recovery room, and for the first 48 hours postoperatively. All intravenous (IV) and peroral opioids given were converted to IV morphine equivalents. Gleason (pathology) scores and tumor pathologic stage were determined at the time of surgery. Prostate tumor staging was performed according to the American Joint Committee on Cancer 1997 TNM cancer classification system. Prostate histopathology was reviewed by Mayo Clinic staff pathologists and urologists. Adjuvant therapy given in the form of androgen deprivation or radiotherapy was at the discretion of the surgeon and patient and was defined as any therapy given within 90 days of surgery. Positive surgical margins were noted from pathology reports. Systemic progression was defined from the presence of metastatic deposits on imaging (bone scan, computerized tomography, X-ray, or magnetic resonance imaging) or on biopsies from locations other than the prostatic bed. In general, vital status was determined from death certificates and/or from yearly correspondence with patients' personal physicians. The prostatectomy registry assigns cause of death based on a review of the death certificate as well as the medical records. The cause of death is attributed to prostate cancer if prostate cancer is listed as a cause in these source documents or if the death occurred within 30 days after prostatectomy (of note, in this study sample there were no patients who died within 30 days of surgery).
Characteristics of patients who received allogeneic BT and those who did not were summarized using descriptive statistics and compared between groups using the rank sum test or chi-square test, as appropriate. The primary outcome of interest was systemic tumor progression, with secondary outcomes being prostate cancer mortality and all-cause mortality. Survival curves were constructed using the Kaplan-Meier method. Survival analyses were performed using stratified proportional hazards regression, taking into account the 1:2 matched set design. Additional multivariable post hoc analyses were performed to compare groups after adjusting for positive surgical margins and adjuvant hormonal or radiotherapy. In addition to analyzing blood product transfusion using a dichotomous variable (any vs. none) we performed an analysis with allogeneic BT modeled using a categorical variable indicating the number of units transfused (none, 1, 2, 3, or more). Results are presented as point estimates and 95% confidence intervals (CIs). All tests were two-sided, with p values of not more than 0.05 considered significant. Statistical analyses were performed using statistical software (SAS, Version 9.1.3, SAS Institute, Cary, NC).
Of the 5110 patients included in the initial cohort, there were 836 (16.4%) who received one or more units of RBCs. Of these, 346 were excluded because they received only autologous blood and 113 were excluded because they received products which were washed, leukoreduced, or irradiated. The remaining 379 patients received allogeneic BT in the perioperative period. Of these patients 93 (24%) received 1 unit, 215 (57%) received 2 units, and 71 (19%) received 3 or more units of blood. The median (range) number of allogeneic units transfused during hospitalization was 2 (1-17) units and the majority (70%) were transfused postoperatively. Besides receiving at least 1 allogeneic unit, 37 of the transfused patients also received autologous blood. In addition, 10 received other non-RBC products (3 FFP and PLTs, 3 FFP only, and 4 PLTs only).
Table 1 summarizes the demographics and clinicopathologic characteristics for the 379 patients who received allogeneic BT and the 758 matched controls who did not receive any blood products. No significant differences were found between the two groups for any major comorbidities or risk factors. As expected, given the groups were matched on type of anesthesia and use of neuraxial block, perioperative opioids use was similar between those who received allogeneic blood versus not (Table 2).
Table 1. Demographic and clinicopathologic characteristics of patients undergoing radical prostatectomy with and without perioperative allogeneic BT
Transfused (n = 379)
Nontransfused (n = 758)
* These conditions were recorded only in the patients receiving medical treatments. Data are number (%) and median (25-75 percentiles). p values are from rank-sum test or chi-square test as appropriate.
ASA = American Society of Anesthesiologists; BMI = body mass index.
Table 2. Type of anesthesia and amount of systemic opioid administered expressed in IV morphine equivalent (mg) in patients undergoing radical prostatectomy with and without perioperative allogeneic BT
Transfused (n = 379)
Nontransfused (n = 758)
aNeuraxial analgesia with spinal or epidural analgesia.
bPerioperative opioids used included morphine, fentanyl, hydrocodone, hydromorphone, meperidine, oxycodone, oxymorphone, propoxyphene, and sufentanil, and their administration was expressed in IV morphine equivalents. Data are number (%) and median (25-75 percentiles). p values are from rank-sum test.
The median (interquartile range) follow-up periods after prostatectomy in the transfused and nontransfused groups were 9.4 (7.3-14) and 10.2 (7.7-13.8) years, respectively. Among both groups 217 (74 transfused, 143 nontransfused) patients died during the follow-up period of which 23 (six transfused, 17 nontransfused) died due to prostate cancer, and 37 (11 transfused, 26 nontransfused) experienced systemic progression (metastases). Figure 1 shows the cumulative percentage of patients experiencing these events according to study group (allogeneic transfusion vs. not).
Stratified proportional hazards regression taking into account the 1:2 matched set design was used to assess differences in oncologic outcomes between the transfused and nontransfused groups. In analyses that did not adjust for any additional covariates, allogeneic BT was not associated with systemic progression (HR, 0.77; 95% CI, 0.35-1.70; p = 0.52), prostate cancer death (HR, 0.82; 95% CI, 0.28-2.43; p = 0.72), or all-cause death (HR, 1.17; 95% CI, 0.85-1.61; p = 0.35). Similarly, in multivariable models which included covariates for positive surgical margins and adjuvant therapies we found no association between BT and oncologic outcomes (systemic progression, HR 0.88, 95% CI 0.39-1.99; p = 0.76; prostate cancer death, HR 1.69, 95% CI 0.44-6.48, p = 0.44; or all-cause death, HR 1.20, 95% CI 0.87-1.67, p = 0.27). In supplemental analysis with the number of allogeneic units transfused modeled categorically (0 vs. 1 vs. 2 vs. >2 units) we found no association between the number of transfused units and systemic progression (p = 0.93), prostate cancer death (p = 0.75), or death from any cause (p = 0.61).
The main finding of our study was a lack of an association between allogeneic BT during radical retropubic prostatectomy for prostate cancer and systemic tumor progression, prostate cancer death, or all-cause death. These findings are in agreement with several previous studies;7,11-15 however, our study is the first to account for immunomodulatory effects of anesthetic technique.
To date, there are 10 retrospective reports that examined association between perioperative BT and prostate cancer recurrence and mortality after prostatectomy (see details of previous reports in Table S1, available as supporting information in the online version of this paper).7,11-18,27 Six demonstrated no significant association,7,11-15 and three suggested that perioperative transfusion resulted in increased cancer recurrence or mortality.16-18 In one study extent of operative blood loss rather than the type of blood transfused was associated with biochemical recurrence after prostatectomy.
One small prospective observational study compared patients undergoing radical retropubic prostatectomy who received allogeneic versus autologous or no blood and found no difference in cancer recurrence and overall survival between the two groups. Three larger retrospective studies reported similar findings to ours.[11, 13, 15] The first included 1785 patients (1345 patients received allogeneic blood) and found no difference in progression-free survival, cancer-specific death, and overall mortality between those who received blood versus not. The second examined 1412 patients (801 received blood) and found no association between the use of allogeneic or autologous BT and cancer recurrence and overall survival. The third examined 1291 patients (205 received blood) and found no association between autologous BT and biochemical recurrence after prostate cancer surgery.
The three small studies that demonstrated adverse effect of BT on prostate cancer recurrence and mortality were performed before PSA era. During this time it was more common for prostate cancer to present with urinary obstruction due to locally progressed cancer; thus most of the patients in these studies underwent transurethral resection of prostate (TURP) to assist in symptomatic management. Heal and colleagues retrospectively studied 262 patients and found a 2.8-fold increased risk of prostate cancer death in transfused patients. However, in this study the surgical approach varied, transfused patients were more likely to have a high Gleason score, and a high percentage of patients were in receipt of primary or secondary hormonal or radiation therapy, making the role of transfusion in prostate oncologic outcomes difficult to interpret. Davies and colleagues reviewed outcomes after 71 TURPs performed over 14 years and found higher prostate cancer recurrence (72% vs. 21%) and lower 5-year survival (17% vs. 66%) in patients who were transfused compared to nontransfused patients. Of note, only 18 patients in this study were transfused, and the study primarily raised the question regarding “wisdom of using TURP instead of radical prostatectomy in patients with prostate cancer (Assessor's comment, Peeling WB).” Finally, the third review included 246 patients with prostate cancer who underwent TURP and reported that transfusion of nonautologous blood had a negative effect on survival.
In the majority of available studies patient comorbidities were not reported; therefore, it is difficult to conclude whether outcomes on cancer recurrence and mortality are secondary to BT itself or if BT was rather a surrogate marker for clinically important variables that may affect oncologic prognosis. For example, Oefelein and coworkers found that the operative blood loss, but not the type of transfusion (autologous or allogeneic), was associated with decreased recurrence-free survival after prostatectomy, and they concluded that events leading to transfusion are more significant for prognosis than the immunologic effects of transfusion itself. In our supplemental analysis of prostate cancer oncologic outcomes, the number of transfused blood units during prostatectomy was not associated with outcomes. This was also found in a previous report from Mayo Clinic.
Strengths of the study
This study has a relatively large cohort of patients who are matched on relevant patient and procedural prognostic determinants of prostate cancer progression and mortality. By using matching that included anesthetic technique we achieved optimal matching, which accounted for the use of systemic opioids, which is considered to be an immunosuppressant perioperative element. As a result of this matching approach, the two groups were comparable in important baseline characteristics. Within the limits of retrospective design, this approach limits bias caused by confounding variables, although it certainly does not eliminate them entirely. Our median follow-up of approximately 10 years is sufficient for observing most important oncologic outcomes.
Limitations of the study
The major limitation of this study is its retrospective design. First, this study was conducted over a span of 15 years and is therefore susceptible to possible congruency issues related to practice changes over time, namely, improvements in surgical techniques (possible less mechanical tumor dissemination) and changes in transfusion practices, all of which may affect the outcomes of interest. We addressed this concern by matching the cases and controls by the year of surgery. On the other hand, by including the patients from the earlier PSA screening period there is a possibility that more aggressive prostate cancers were included. The impact of interventions like BT on oncologic outcomes may be dependent on the stage (or aggressiveness) of the tumor. Although we matched on tumor stage and the date of surgery, we do not have adequate statistical power to perform subgroup analyses to assess whether BT is associated with oncologic outcomes in specific patient populations. Another limitation is that since the period of this study, transfusion practice changed, with many of the blood units administered in later period being leukoreduced. Because our study includes only nonleukoreduced RBC transfusions, we cannot comment on the potential association between leukoreduced RBC transfusion and long-term prostate cancer outcomes.
In conclusion, in our study, which included well-matched patients according to potential risks for tumor recurrence, we found no significant association between administration of allogeneic BT during radical prostatectomy and subsequent occurrence of metastatic disease, cancer-specific death, and all-cause mortality.
Conflict of Interest
The authors report no conflicts of interest or funding sources.