- Email: [email protected]
Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome in patients undergoing surgery for ovarian cancer: A propensity score-matched analysis
YGYNO-977750; No. of pages: 7; 4C: Gynecologic Oncology xxx (xxxx) xxx
Contents lists available at ScienceDirect
Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno
Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome in patients undergoing surgery for ovarian cancer: A propensity score-matched analysis Hao Zhang 1, Xin Wu 1, Zheng Xu 1, Zhirong Sun, Minmin Zhu ⁎, Wankun Chen ⁎, Changhong Miao ⁎ Department of Anesthesiology, Fudan University Shanghai Cancer Center, Shanghai 200032, China Department of Oncology, Shanghai Medical College, Fudan University, Shanghai 200032, China
H I G H L I G H T S • Transfusion was associated with increased cancer recurrence and mortality in ovarian cancer patients. • Transfusion was associated with higher postoperative complication rates and longer hospitalization stays. • Transfusion was associated with more postoperative fluctuations in systemic inflammatory markers.
a r t i c l e
i n f o
Article history: Received 20 September 2019 Received in revised form 4 December 2019 Accepted 6 December 2019 Available online xxxx Keywords: Ovarian cancer Perioperative red blood cell transfusion Systemic inflammation Postoperative recovery Long-term outcome
a b s t r a c t Background. The impact of perioperative red blood cell transfusion (PRBCT) on cancer survival has remained controversial. Methods. We conducted a retrospective study in patients undergoing primary debulking surgery (PDS) for ovarian cancer between January 2013 and December 2017. The patients were divided into two groups based on whether they received PRBCT. Clinical characteristics were compared between groups. After propensity score matching, perioperative systemic inflammation-based scores, quality of recovery, postoperative outcomes, disease-free survival (DFS), and overall survival (OS) were compared between groups. Univariate and multivariable Cox proportional hazard models were used to evaluate the association between covariates and survival outcomes. Results. A total of 1037 patients were enrolled in this study, and 31.7% of patients received PRBCT. After propensity matching, there was no significant difference in the clinical characteristics between groups. Patients receiving PRBCT had more postoperative fluctuations in systemic inflammatory response-related indicators (P b 0.001), a higher incidence of postoperative grade II complications (28.4% vs. 14.8%), a longer length of stay (10.6 d vs. 6.2 d) and higher 30-day and total readmission rates (7.1% vs. 4.4% and 11.2% vs. 8.1%, P b 0.001, respectively) than patients who did not receive PRBCT. The OS and DFS rates 3 years after surgery were significantly lower in the patients receiving PRBCT than in patients not receiving PRBCT (58.9% vs. 74.5%, 39.6% vs. 52.3%). Conclusions. PRBCT was significantly associated with more fluctuations in systemic inflammatory indicators, a prolonged length of stay, higher postoperative complication rates and increased cancer recurrence and overall mortality in ovarian cancer patients undergoing PDS. © 2019 Published by Elsevier Inc.
1. Introduction
⁎ Corresponding authors at: Department of Anesthesiology, Fudan University Shanghai Cancer Center, Shanghai 200032, China. E-mail addresses: [email protected] (M. Zhu), [email protected] (W. Chen), [email protected] (C. Miao). 1 These authors contributed equally to this work.
Worldwide, the incidence of ovarian cancer ranks third among gynecologic tumors, and it is the second most common cause of death from gynecologic cancer [1]. Most patients suffering from ovarian cancer have progressed to an advanced stage at the time of diagnosis. The standard treatment for ovarian cancer is surgery supplemented with chemotherapy, targeted therapy and other comprehensive treatments [2,3]. Primary debulking surgery (PDS) is the cornerstone of ovarian
https://doi.org/10.1016/j.ygyno.2019.12.006 0090-8258/© 2019 Published by Elsevier Inc.
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006
2
H. Zhang et al. / Gynecologic Oncology xxx (xxxx) xxx
cancer treatment. Satisfactory tumor cytoreductive surgery is a favorable factor for prolonging survival in patients with ovarian cancer [3]. Ovarian cancer patients undergoing tumor reduction surgery often have anemia symptoms before surgery, and in order to achieve R0 resection, the surgical trauma is often substantial, which undoubtedly increases the risk of bleeding and perioperative anemia. It has been reported that perioperative blood transfusion rates in patients receiving PDS range from 19% to 56% [4,5]. For patients with severe anemia, blood transfusions can be lifesaving [6,7]. However, some studies in recent years have shown that the average length of hospital stay and the rates of postoperative infection, cancer recurrence, metastasis, and cancer-related mortality are significantly higher in cancer patients undergoing transfusion during surgery, and the cancer recurrence rate is also higher [8–10]. This may be because allogeneic blood cells are exogenous substances that induce immune responses in recipients and inhibit systematic anti-tumor immune function [11,12]. Some laboratory-based inflammatory indicators related to inflammatory and immune responses, such as the NLR (neutrophil-to-lymphocyte ratio), LMR (lymphocyte-to-monocyte ratio) and SII (systemic immune inflammation index), have been shown to be associated with prognosis in ovarian cancer [13–15]. Although there have been some studies on the long-term prognosis of perioperative transfusion in ovarian cancer patients, the conclusions are not the same [8,16]. This retrospective study was designed to compare the effects of perioperative blood transfusion on perioperative systemic inflammation indicators, postoperative recovery and long-term survival outcomes. 2. Materials and methods This study was approved by the Ethics Committee of Fudan University Shanghai Cancer Center (FUSCC), China. From January 2013 to December 2017, patients undergoing PDS for ovarian cancer were enrolled in this retrospective cohort. The exclusion criteria were emergency surgery, incomplete medical records, patients who underwent surgery for recurrent disease, treatment with anti-inflammatory agents or immunosuppressants for more than one month before surgery, patients with chronic inflammatory diseases and infections including autoimmune diseases, and preoperative neoadjuvant chemotherapy and radiotherapy. The data were extracted from the database of the FUSCC clinical information system. The medical information of each patient was reviewed and recorded, including demographic information, primary diagnosis, medical history, operative details (operative time, surgical complexity and estimated blood loss volumes), pathological details (histologic diagnosis, tumor size, tumor differentiation, and FIGO stage), preoperative Hgb level, preoperative CA125 levels, perioperative red blood cell transfusion volumes, perioperative systemic inflammation-based scores, and postoperative outcomes. The counts of neutrophils, lymphocytes, monocytes and platelets were recorded within 3 days before surgery (pre op), on the first day after surgery (POD1) and on the third day after surgery (POD3). Preoperative systemic inflammation-based scores, including the neutrophilto-lymphocyte ratio (NLR), the lymphocyte-to-monocyte ratio (LMR) and the systemic immune inflammation index (SII), were calculated based on these laboratory results. The SII was defined as follows: SII = neutrophil × platelet / lymphocyte. Surgical complexity of each patients was assigned according to a surgical complexity score system reflecting the complexity and number of surgical procedures as described by Aletti et al. [17]. The postoperative outcomes included postoperative complications (graded I–V according to the Clavien-Dindo classification system for surgical complications [18,19]), hospital length of stay (LOS), hospital readmission within 30 and 90 days, and 90-day mortality. Perioperative red blood cell transfusion (PRBCT) was defined as intraoperative transfusion and transfusion occurring during the
postoperative hospital admission. Comparisons were made between patients who received perioperative RBC transfusion (group BT) and those who did not receive perioperative RBC transfusion (group NBT). The primary outcomes of interest were postoperative recovery parameters (incidence of postoperative complications, LOS, and readmission rates) and disease-free survival (DFS) and overall survival (OS). DFS was defined as the length of time from the date of surgery to the date of the first evidence of tumor recurrence, or December 31, 2018. OS was defined as the length of time from the date of surgery to the date of death or the last follow-up date. 3. Statistical methods In this study, we compared different clinicopathologic factors between the BT group and NBT group as well as descriptive statistics including the mean and standard deviation of continuous variables such as age, body mass index (BMI), operative time, estimated blood loss, preoperative Hgb, and preoperative tumor marker level. Frequency counts and percentages were calculated for categorical variables such as the American Society of Anesthesiologists physical status score (ASA), histologic diagnosis, FIGO stage, tumor size, tumor differentiation, surgical complexity, and postoperative chemotherapy. CA125 levels were expressed as median (25%, 75% quartiles). The Chi-square test was used to evaluate the association between two categorical variables. The Kaplan-Meier method was used for time-to-event analysis, including DFS and OS. The median time to the event in months with a 95% confidence interval (CI) was calculated. The log-rank test was used to evaluate the difference in the time-to-event end points between the patient groups. Univariate Cox proportional hazards models were fitted to evaluate the effects of continuous variables on the time-toevent outcomes. Multivariable Cox proportional hazard models were used for multivariate analysis to include important and significant covariates. Statistical software SAS version 9.3 (SAS, Cary, NC) and S-Plus version 8.2 (TIBCO Software Inc., Palo Alto, CA) were used for all the analyses. A P value b 0.05 was considered statistically significant. To adjust for selection bias in the retrospective observational study, a propensity score-matching (PSM) analysis was conducted. The following prognostic covariates were included in the multi-covariate logistic model to estimate the propensity scores: age, BMI, ASA, FIGO stage, tumor size, tumor differentiation, postoperative chemotherapy, preoperative CA125 levels, residual disease and estimated blood loss (EBL). The Greedy 5-to-1-digit match algorithm was used to match the baseline covariates so that the two study groups (BT or NBT) would have similar propensity scores [20]. Among the 1037 patients enrolled in the entire study population, the propensity score was calculated for 596 patients. Of those, 298 patients were in the BT group, and 298 were in the NBT group. A total of 298 patients in the BT group with non-missing values for the selected covariates were matched on a 1:1 ratio to the patients in the NBT group with non-missing values for the same covariates. The maximum standardized differences for all covariates were 9.6% in the post-matching cohort, suggesting a substantial reduction in bias between the two groups. 4. Results A total of 1037 patients who underwent PDS for ovarian cancer were enrolled in this study, with 329 in the BT group and 708 in the NBT group, of whom 31.7% received perioperative blood transfusion. The demographics and operative characteristics of all patients are shown in Table 1. There was no significant difference in histologic diagnosis, FIGO stage and tumor differentiation between groups (P N 0.05). The patients in the BT group were significantly older than those in the NBT group (54.0 ± 11.6 vs. 52.4 ± 12.4 yr, P = 0.043), and the BT group had more patients than the NBT group with an ASA score of III–IV (20.4% vs. 15.8%, P = 0.030); RBC transfusion was also associated with higher preoperative serum CA125 levels (1472 ± 1170 vs. 1141 ±
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006
H. Zhang et al. / Gynecologic Oncology xxx (xxxx) xxx
3
Table 1 Demographics and operative characteristics of patients in both groups before and after propensity score matching. Entire study population
Age (years) BMI (kg/m2) ASA I–II III–IV Histologic diagnosis Serous histology Non-serous histology FIGO stage I II III IV Tumor size b5 cm N5 cm Tumor differentiation Well Moderate Poor Surgical complexity Low Intermediate High Residual disease No visible residual disease ≤1 cm residual disease ≥1 cm residual disease Operation time (min) Estimated blood loss (ml) Ascites (ml) b200 N200 Pre Hgb (g/dl) Pre HCT (%) Pre CA125 (U/ml) Postop Chemotherapy
Matched covariates
BT group (n = 329)
NBT group (n = 708)
52.4 ± 12.4 28.7 ± 7.7
54.0 ± 11.6 29.2 ± 7.9
242 (73.6) 67 (20.4)
596 (84.2) 112 (15.8)
235 (71.4) 94 (28.6)
518 (73.1) 190 (26.9)
29 (8.8) 38 (11.6) 158 (48.0) 104 (31.6)
63 (9.9) 79 (11.1) 343 (48.4) 223 (30.6)
198 (60.1) 131 (39.9)
439 (62.0) 269 (38)
32 (9.7) 196 (59.7) 101 (30.6)
79 (11.0) 421 (59.6) 208 (29.4)
63 (19.1) 199 (60.5) 67 (20.4)
138 (19.5) 482 (68.1) 88 (17.2)
173 (52.5) 127 (38.5) 29 (9) 203 ± 56 536 ± 326
376 (53.1) 258 (36.5) 74 (10.4) 182 ± 48 421 ± 226
83 (25.2) 46 (14.0) 105 ± 15 29 ± 5 621(195–1073) 292 (89.1)
189 (26.8) 110 (15.6) 129 ± 12 33 ± 4 412 (52–1206) 644 (91.2)
p Value 0.043 0.339 0.030
BT group (n = 298)
NBT group (n = 298)
Standardized difference (%)
54.0 ± 13.1 28.8 ± 8.4
54.4 ± 12.0 29.6 ± 7.8
3.5 2.6 1.0
256 (85.9) 43 (14.1)
253 (84.8) 45 (15.2)
198 (66.4) 100 (33.6)
203 (68.1) 95 (32.9)
32 (10.7) 32 (10.7) 134 (45.0) 100 (33.6)
35 (11.7) 31 (10.4) 143 (48.0) 89 (29.9)
186 (62.4) 112 (37.6)
189 (63.4) 116 (36.6)
24 (8.2) 172 (57.6) 102 (34.2)
24 (7.9) 190 (63.7) 84 (28.4)
57 (19.2) 190 (63.9) 51 (15.9)
58 (19.6) 191 (64.3) 49 (16.1)
162 (54.3) 109 (36.5) 27 (9.2) 196 ± 42 545 ± 321
160 (53.6) 107 (35.8) 31 (10.6) 185 ± 52 418 ± 212
76 (23.2) 39 (13.1) 124 ± 15 29 ± 4 610 (186–943) 259 (86.9)
72 (24.5) 46 (14.6) 125 ± 15 32 ± 5 577 (56–836) 262 (87.8)
0.560
0.2
0.998
0.8
0.575
5.7
0.755
3.4
0.003
1.4
0.637
b0.001 b0.001 0.824
b0.001 b0.001 b0.001 0.265
2.3
1.5 3.6 5.2
3.5 3.1 9.6 5.2
Data are shown as the mean ± SD or n (%), or median (25%, 75% quartiles). BT = Blood transfusion; NBT = No blood transfusion; FIGO stage = Federation International of Gynecology and Obstetrics; BMI = Body mass index; ASA = American Society of Anesthesiologists; P ≤ 0.05 was considered statistically significant.
829 U/ml, P b 0.001), lower preoperative Hgb levels (105 ± 15 vs. 129 ± 12 g/dl, P b 0.001), higher surgical complexity (P = 0.003), a longer operation time (203 ± 56 vs. 182 ± 48 mins, P b 0.001), and higher estimated blood loss (536 ± 326 vs. 421 ± 226 ml, P b 0.001). Because of the significant differences in baseline characteristics between the two groups, we used the propensity matching score to reduce the imbalance. After matching, 298 pairs were formed. There was no significant difference in the demographic or operative characteristics between these two groups in the matched cohort (Table 1). In BT group, 65 patients (21.8%) received only intraoperative transfusion, 201 patients (67.4%) received both intraoperative and postoperative transfusion, and 32 patients (10.7%) received only postoperative transfusion, most transfusion (80.5%) occurred within 3 day after surgery; 47 (15.8%) patients received 1–4 units of PRBCT and 251 (84.2%) patients received N5 units of PRBCT perioperatively.
After matching, there were no significant differences in the preoperative NLR, LMR and SII between the two groups (P N 0.05, Fig. 1). Compared with pre op, after the operation, the NLR and SII in the two groups significantly increased, and the LMR was significantly decreased in all patients on POD1 and POD3 (P b 0.001, Fig. 1). The NLR and SII values were slightly lower on POD3 than on POD1 but were still higher than the pre op values (P b 0.05, Fig. 1). The NLR and SII values in the NBT group on POD1 and POD3 were significantly lower than those in the BT group (P b 0.001, Fig. 1), whereas the LMR was significantly higher in the NBT group than in the BT group (P b 0.001, Fig. 1). In terms of postoperative complications, the incidences of postoperative pneumonia, wound complications, intra-abdominal infection and postoperative deep venous thrombosis/pulmonary embolus (DVT/PE) were also significantly lower in the NBT group than in the BT group [mean incidence (%): 3.20 vs. 6.22; 10.45 vs. 15.44; 2.75 vs. 5.88 and
Fig. 1. Perioperative NLR, LMR and SII in both groups.
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006
4
H. Zhang et al. / Gynecologic Oncology xxx (xxxx) xxx
Fig. 2. Type of postoperative complications (A), Clavien-Dindo classification (B), length of hospital stays (C) and rates of readmission and mortality (D) in both groups.
2.24 vs. 4.41, respectively; P b 0.001, Fig. 2A]. The incidence of postoperative vomiting was slightly higher in the BT group than in the NBT group, but there was no significant difference between the two groups [mean incidence (%): 3.90 vs. 3.20; P = 0.420, Fig. 2A]. According to the Clavien-Dindo classification system for surgical complications, the incidence of grade I postoperative complications was comparable between the two groups, and the incidence of grade II complications was significantly lower in the NBT group than in the BT group [mean incidence (%): 1.9 vs. 3.4 and 14.8 vs. 28.4, respectively; P b 0.001, Fig. 2B]. There were no severe complications in either group (ClavienDindo grade N II). The LOS in the NBT group was significantly shorter than that in the BT group [mean time (days): 6.2 vs. 10.6, P b 0.001, Fig. 2C]. The NBT group also had significantly lower 30-day and total readmission rates than the BT group [mean incidence (%): 4.4 vs. 7.1 and
8.1 vs. 11.2, respectively; P b 0.001, Fig. 2D]. There was no significant difference in the 90-day readmission rate or the 90-day mortality rate between the two groups [mean incidence (%): 2.8 vs. 3.2 and 1.5 vs. 1.6, respectively; P b 0.001, Fig. 2D]. The median follow-up time was 4.3 years among the entire cohort. The Kaplan-Meier survival curves for the propensity-matched NBT and BT groups are displayed in Fig. 3. The shapes and Cox model analysis of the curves over time suggest that the OS rate 3 years after surgery was significantly lower in patients in the BT group than in patients in the NBT group (58.9% vs. 74.5%, P b 0.001, Fig. 3A), with a hazard ratio of 2.62 (95% CI, 2.02 to 4.08; P b 0.001). The DFS rate 3 years after surgery was also significantly lower in the BT group than in the NBT group (39.6% vs. 52.3%, P b 0.001; Fig. 3B), with a hazard ratio of 3.32 (95% CI, 2.36 to 6.06; P b 0.001). Further analysis of the relationship between
Fig. 3. Kaplan-Meier curves depicting the effect of blood transfusion on overall survival (A) and disease-free survival (B) in ovarian cancer patients.
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006
H. Zhang et al. / Gynecologic Oncology xxx (xxxx) xxx
5
Fig. 4. Kaplan-Meier curves depicting a dose-response relationship for overall survival (A) and disease-free survival (B) in ovarian cancer patients.
different volumes of blood transfusion and OS and DFS suggested that the OS rate 3 years after surgery was significantly lower in patients receiving both 1–4 RBC unit transfusions and N5 RBC unit transfusions than in the patients in the NBT group (60.7% vs. 74.5%, 58.1% vs. 74.5%, P b 0.001, respectively; Fig. 4A); however, there was no significant difference in OS between the two groups of patients receiving 1–4 RBC unit transfusions and N5 RBC unit transfusions (60.7% vs. 58.1%, P = 0.708; Fig. 4A). The DFS rate 3 years after surgery was also significantly lower in the patients receiving both 1–4 RBC unit transfusions and N5 RBC unit transfusions than in the patients in the NBT group (42.3% vs. 52.3%, 38.3% vs. 52.3%, P b 0.001, respectively; Fig. 4B); however, there was no significant difference in DFS between the two groups of patients receiving 1–4 RBC unit transfusions and N5 RBC unit transfusions (42.3% vs. 38.3%, P = 0.564; Fig. 4B). The OS and DFS were compared with blood transfusion and other variables separately in a univariate Cox model and subsequently in a multivariable Cox regression. In the univariate analysis, blood transfusion, age, ASA physical status (III–IV), non-serous histology, tumor size (N5 cm), FIGO stage (III–IV), poor tumor differentiation, higher surgical complexity, residual disease, longer operation time, higher estimated blood loss, and no postoperative chemotherapy were associated with a worse impact on OS and DFS (Table 2). Blood transfusion was associated with an HR of 1.54 (1.17 to 2.03) for OS and an HR of 1.61 (1.12 to 2.17) for DFS after multivariable analysis for known confounders (Table 3). Other variables associated with a significant increase in the hazard of death after multivariable analysis included non-serous histology, poor tumor differentiation, high surgical complexity, high estimated blood loss, increased volume of PRBCT, residual disease, and ascites (N200 ml) (Table 3). 5. Discussion The impact of PRBCT on survival and cancer recurrence/metastasis has remained a topic of controversy in oncological surgery [21,22]. In this single-center, retrospective cohort study, we found that PRBCT were strongly associated with worse short- (higher incidence of postoperative complications, a prolonged LOS, and higher readmission rates) and long-term outcomes after ovarian cancer surgery. In addition, the analysis showed that non-serous histology, poor tumor differentiation, high surgical complexity, high estimated blood loss, increased volume of PRBCT, residual disease and ascites were also independent risk factors of worse OS. PDS often requires multi-organ resections and large amounts of estimated blood loss that requires PRBCT [5,23,24]. In our study, 31.7% of patients received PRBCT. Not surprisingly, a large estimated blood loss was also associated with worse outcomes, suggesting that patients who may have bled more received a blood transfusion. Some studies demonstrated that perioperative blood transfusions may adversely affect outcomes in patients with cancer, favoring tumor growth and dissemination [25,26], but some other studies reached the opposite conclusion [27,28]. The literatures on PRBCT and prognosis in patients with ovarian cancer are also inconsistent [16,29]. For instance, Connor et al. demonstrated that PRBCT is associated with poor overall survival
in advanced epithelial ovarian cancer patients [8]. While Hunsicker et al. showed that perioperative transfusion of RBCs did not increase the risk of recurrence after ovarian cancer surgery [30]. Although, our results did not suggest a dose-response relationship between PRBCT Table 2 Univariate cox proportional hazard model for overall survival (OS) and disease-free survival (DFS). Variables
OS
DFS
HR
95% CI
Age, years BMI, kg/m2
1.02 1.00
1.00–1.04 0.020 0.97–1.03 0.810
P
HR
95% CI
0.98 1.01
0.96–1.00 0.009 0.98–1.04 0.569
P
ASA I–II III–IV
Reference – b0.001 Reference – b0.001 1.56 1.21–1.97 1.42 1.18–1.69
Histologic diagnosis Serous histology Reference – b0.001 Reference – 0.004 Non-serous 1.34 1.14–1.59 1.28 1.09–1.51 histology Tumor size b5 cm N5 cm
Reference – b0.001 Reference – b0.001 2.82 2.05–3.61 2.61 1.98–3.34
FIGO stage I–II III–IV
Reference – 0.013 1.61 1.18–2.18
Reference – 0.021 1.45 1.09–1.89
Tumor differentiation Well Reference – b0.001 Moderate 1.16 1.01–1.34 1.21 Poor 1.42 1.10–1.76 1.88
b0.001 1.04–1.42 1.42–2.31
Surgical complexity Low Reference – b0.001 Reference – b0.001 Intermediate 1.74 1.17–2.33 1.61 1.08–2.24 High 2.42 1.68–3.34 2.28 1.72–2.81 Residual disease No visible residual disease ≤1 cm residual disease ≥1 cm residual disease Operation time (min) Estimated blood loss (ml) Ascites, (ml) b200 N200
Reference –
b0.001 Reference –
b0.001
1.45
1.26–1.64
1.38
1.32–1.75
1.68
1.32–1.95
1.72
1.42–1.86
1.30
1.02–1.45 b0.001 1.27
1.01–1.51 0.002
1.82
1.41–2.18 b0.001 1.87
1.37–2.39 b0.001
Reference – 0.004 1.51 1.15–2.05
Reference – 0.012 1.42 1.08–1.93
Blood transfusion No Reference – b0.001 Reference – b0.001 Yes 1.12 1.03–1.39 1.27 1.09–1.48 Postop chemotherapy Yes Reference – 0.008 No 1.41 1.20–1.62
Reference – 0.014 1.57 1.14–2.11
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006
6
H. Zhang et al. / Gynecologic Oncology xxx (xxxx) xxx
Table 3 Multivariable Cox proportional hazard model for overall survival (OS) and disease-free survival (DFS). Variables
OS
DFS
HR
95% CI
Age, years
0.97
0.94–1.01 0.455
ASA score I–II III–IV Non-serous histology
Reference – 0.103 1.03 0.94–1.02 1.23 1.04–1.46 0.004
Reference – 0.690 1.00 0.99–1.01 1.26 1.07–1.51 0.007
Tumor size b5 cm N5 cm
Reference – 0.549 1.01 0.81–1.22
Reference – 0.104 1.63 0.94–3.01
FIGO stage I–II III–IV
Reference – 0.187 1.11 0.81–2.21
Reference – 0.205 0.97 0.94–1.01
Tumor differentiation Well Reference – 0.013 Moderate 1.45 1.08–2.07 Poor 1.73 1.16–2.56
Reference – 0.003 1.38 1.06–1.79 1.52 1.17–1.91
Surgical complexity Low Reference – 0.001 Intermediate 1.74 1.10–2.74 High 2.11 1.42–2.87
Reference – 0.004 1.46 1.09–1.96 2.27 1.57–3.01
Residual disease No visible residual disease ≤1 cm residual disease ≧1 cm residual disease Operation time (min) Estimated blood loss (ml) Ascites, (ml) b200 N200
Reference –
P
HR
95% CI
0.95
0.80–1.12 0.518
b0.001 Reference –
P
b0.001
1.33
1.21–1.68
1.42
1.29–1.85
1.92
1.65–2.12
1.64
1.56–1.87
1.01
0.87–1.32 0.451
0.99
0.71–1.34 0.387
1.78
1.35–2.18 0.001
1.84
1.46–2.23 0.001
Reference – 0.015 1.60 1.09–2.33
Reference – 0.014 1.52 1.09–2.13
Blood transfusion No Reference – 0.002 Yes 1.61 1.12–2.17
Reference – 0.002 1.54 1.17–2.03
Number of pRBC 0 1–4 5 to more
Reference – b0.001 Reference – b0.001 1.32 1.22–1.65 1.26 1.12–1.47 1.45 1.34–1.76 1.41 1.32–1.62
Chemotherapy Yes No
Reference – 0.375 1.22 0.79–1.81
Reference – 0.874 1.01 0.84–1.27
and OS and DFS, but the multivariable analysis showed that the increased volume of PRBCT was an independent risk factors of worse survival. This may be caused by the different number of patients between the high and low volume subgroup of PRBCT, and these two subgroups are not matched during comparison. Allogeneic blood transfusion induces profound biology alterations on the immune system, which has been indicated as a potential mechanism for increased risk of infections, post-transfusion graft-versushost disease, and cancer recurrence [11,31,32]. Systematic inflammation and immunosuppression contribute to the malignant progression of cancer [33]. Recent studies showed that the suppression perioperative anti-tumor immune function could facilitate the formation of new metastases and promote postoperative metastasis and recurrence [34,35]. Numerous studies have also shown that some markers of hematological components, such as the NLR, LMR and SII, can comprehensively indicate the balance of the host immune system and have been
considered prognosis-related indicators in malignancies [36–38]. An increased NLR or SII and a decreased LMR are associated with an adverse prognosis in ovarian cancer [13–15]. The results of our study show that patients who receive PRBCT have more postoperative fluctuations in these systemic inflammatory markers. This suggests that PRBCT was associated with strong suppression of immune function and/or a dysregulated inflammatory response as indicated by higher complication rates and longer hospitalization stays in those who receive blood transfusions [39]. In recent years, with the better understanding of perioperative pathophysiology, the perioperative management concept of major surgery has advanced greatly. Enhanced Recovery After Surgery (ERAS) programs are becoming the standard practice in many surgical specialties [40]. ERAS is a multidisciplinary, evidence-based approach to improve perioperative outcomes and accelerate postoperative recovery [41]. Implementation of ERAS strategy in gynecological surgery has been shown to be associated an improvement in clinical outcomes, including shortened LOS [42]. Previous study showed that PRBCT could delay the return of bowel function and increase the LOS [42]. Our results also suggested that RBCT were associated with higher incidence of postoperative complications, a prolonged LOS, and higher readmission rates. These results indicated that reducing PRBCT could be an important part of the ERAS strategy for gynecologic oncology. However, more studies are needed to further clarify the criteria for PRBCT, especially postoperative blood transfusion in the ERAS strategy [43]. Our study has several strengths and limitations. To reduce significant confounding, we conducted a propensity score matching analysis after which PRBCT were still associated with worse outcomes. Limitation included first that the study was retrospective and not randomized, and the data were derived from a single center. Second, even though we conducted a propensity score matching analysis, given the multiple tumor and perioperative factors associated with receiving PRBCT, we could not avoid the possibility of residual confounding due to unmeasured confounders. Third, we did not further assess the effect of other blood products and other perioperative variables that might alter immune responses and affect long-term outcomes after surgery. In conclusion, perioperative blood transfusion was significantly associated with more fluctuations in systemic inflammatory indicators, a prolonged LOS, higher postoperative complication rates and increased cancer recurrence and overall mortality in ovarian cancer patients undergoing PDS. Well-designed prospective studies are warranted to explore the relationship between allogeneic blood transfusion and longterm outcomes in patients undergoing cancer surgery. Funding This research was supported by the National Natural Science Foundation of China (NO. 81873948, 81871590, 81871591), the National Key Research and Development Program of China (No: 2018YFC2001904), Shanghai Shenkang Hospital Development Center Clinical Science and Technology Innovation Project (NO. SHDC12018105), the Key Technology and Development Program of Shanghai (NO. 17411963400), Science and Technology Innovation Action Plan (No. 16DZ1930304). Consent for publication Not applicable. Authors' contributions H.Z., W.C. and C.M. conceived and designed the study; H.Z., Z.X. and Z.S. collected the data; H.Z., M.Z. and W.C. interpreted and analysed the data; H.Z., X.W. and Z.X. was a major contributor in writing the manuscript; M.Z., W.C. and C.M. reviewed and edited the manuscript. All authors read and approved the final manuscript.
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006
H. Zhang et al. / Gynecologic Oncology xxx (xxxx) xxx
Availability of data and materials The datasets analysed during the current study available from the corresponding author on reasonable request. Declaration of competing interest The authors declare that they have no competing interests. Acknowledgements The authors would like to thank Dr. Juan P. Cata, M.D. (Department of Anesthesiology and Perioperative Medicine, The University of Texas MD Anderson Cancer Center, Houston, USA) for his advice and comments for revising the manuscript. References [1] F. Bray, et al., Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries, CA Cancer J. Clin. 68 (2018) 394. [2] U.A. Matulonis, et al., Ovarian cancer, NAT REV DIS PRIMERS 2 (2016) 16061. [3] A. Karam, et al., Fifth Ovarian Cancer Consensus Conference of the Gynecologic Cancer InterGroup: first-line interventions, Ann. Oncol. 28 (2017) 711. [4] L.S. Prescott, et al., Perioperative blood transfusion in gynecologic oncology surgery: analysis of the National Surgical Quality Improvement Program Database, Gynecol. Oncol. 136 (2015) 65. [5] S.A. Ackroyd, et al., A preoperative risk score to predict red blood cell transfusion in patients undergoing hysterectomy for ovarian cancer, Am. J. Obstet. Gynecol. 219 (2018) 591. [6] A. Shander, M. Javidroozi, S. Ozawa, G.M. Hare, What is really dangerous: anaemia or transfusion? Br. J. Anaesth. 107 (Suppl. 1) (2011) i41. [7] B. Clevenger, S.V. Mallett, A.A. Klein, T. Richards, Patient blood management to reduce surgical risk, Br. J. Surg. 102 (1325) (2015) 1324. [8] J.P. Connor, A. O'Shea, K. McCool, E. Sampene, L.M. Barroilhet, Peri-operative allogeneic blood transfusion is associated with poor overall survival in advanced epithelial ovarian Cancer; potential impact of patient blood management on Cancer outcomes, Gynecol. Oncol. 151 (2018) 294. [9] P. Cybulska, C. Goss, W.P. Tew, R. Parameswaran, Y. Sonoda, Indications for and complications of transfusion and the management of gynecologic malignancies, Gynecol. Oncol. 146 (2017) 416. [10] C.T. Aquina, et al., Association among blood transfusion, sepsis, and decreased longterm survival after colon cancer resection, Ann. Surg. 266 (2017) 311. [11] J.P. Cata, H. Wang, V. Gottumukkala, J. Reuben, D.I. Sessler, Inflammatory response, immunosuppression, and cancer recurrence after perioperative blood transfusions, Br. J. Anaesth. 110 (2013) 690. [12] S. Hart, C.M. Cserti-Gazdewich, S.A. McCluskey, Red cell transfusion and the immune system, Anaesthesia 70 (Suppl. 1) (2015) 38e13. [13] T. Baert, et al., Influence of CA125, platelet count and neutrophil to lymphocyte ratio on the immune system of ovarian cancer patients, Gynecol. Oncol. 150 (2018) 31. [14] D. Nie, H. Gong, X. Mao, Z. Li, Systemic immune-inflammation index predicts prognosis in patients with epithelial ovarian cancer: a retrospective study, Gynecol. Oncol. 152 (2019) 259. [15] B.S. Kwon, et al., Prognostic value of preoperative lymphocyte-monocyte ratio in patients with ovarian clear cell carcinoma, J. Cancer 9 (2018) 1127. [16] B.L. Manning-Geist, et al., Infection, thrombosis, and oncologic outcome after interval debulking surgery: does perioperative blood transfusion matter? Gynecol. Oncol. 153 (2019) 63. [17] G.D. Aletti, S.C. Dowdy, K.C. Podratz, W.A. Cliby, Relationship among surgical complexity, short-term morbidity, and overall survival in primary surgery for advanced ovarian cancer, Am. J. Obstet. Gynecol. 197 (671) (2007). [18] P.A. Clavien, et al., The Clavien-Dindo classification of surgical complications: fiveyear experience, Ann. Surg. 250 (2009) 187.
7
[19] S. Cham, et al., Development and validation of a risk-calculator for adverse perioperative outcomes for women with ovarian cancer, Am. J. Obstet. Gynecol. 220 (2019) 571. [20] P.C. Austin, A comparison of 12 algorithms for matching on the propensity score, Stat. Med. 33 (2014) 1057. [21] W.B. Al-Refaie, H.M. Parsons, A. Markin, J. Abrams, E.B. Habermann, Blood transfusion and cancer surgery outcomes: a continued reason for concern, SURGERY 152 (2012) 344. [22] M. Horowitz, E. Neeman, E. Sharon, S. Ben-Eliyahu, Exploiting the critical perioperative period to improve long-term cancer outcomes, Nat. Rev. Clin. Oncol. 12 (2015) 213. [23] J.A. Gilreath, D.D. Stenehjem, G.M. Rodgers, Diagnosis and treatment of cancerrelated anemia, Am. J. Hematol. 89 (2014) 203. [24] Practice guidelines for perioperative blood management: an updated report by the American Society of Anesthesiologists Task Force on Perioperative Blood Management*, ANESTHESIOLOGY 122 (2015) 241. [25] M.J. Latif, et al., Perioperative blood transfusion has a dose-dependent relationship with disease recurrence and survival in patients with non-small cell lung cancer, J. Thorac. Cardiovasc. Surg. 157 (2019) 2469. [26] A.G. Acheson, M.J. Brookes, D.R. Spahn, Effects of allogeneic red blood cell transfusions on clinical outcomes in patients undergoing colorectal cancer surgery: a systematic review and meta-analysis, Ann. Surg. 256 (2012) 235. [27] R. Warschkow, et al., Perioperative blood transfusions do not impact overall and disease-free survival after curative rectal cancer resection: a propensity score analysis, Ann. Surg. 259 (2014) 131. [28] R. Kang, B.E. Seath, V. Huang, R.J. Barth, Impact of autologous blood transfusion on survival and recurrence among patients undergoing partial hepatectomy for colorectal cancer liver metastases, J. Am. Coll. Surg. 228 (2019) 902. [29] M. Orskov, M. Iachina, R. Guldberg, O. Mogensen, N.B. Mertz, Predictors of mortality within 1 year after primary ovarian cancer surgery: a nationwide cohort study, BMJ Open 6 (2016), e10123. [30] O. Hunsicker, et al., Transfusion of red blood cells does not impact progression-free and overall survival after surgery for ovarian cancer, Transfusion. 59 (2019) 3589. [31] K.A. Horvath, et al., Blood transfusion and infection after cardiac surgery, Ann. Thorac. Surg. 95 (2013) 2194. [32] P. Cybulska, C. Goss, W.P. Tew, R. Parameswaran, Y. Sonoda, Indications for and complications of transfusion and the management of gynecologic malignancies, Gynecol. Oncol. 146 (2017) 416. [33] F.R. Greten, S.I. Grivennikov, Inflammation and cancer: triggers, mechanisms, and consequences, IMMUNITY 51 (2019) 27. [34] M. Horowitz, E. Neeman, E. Sharon, S. Ben-Eliyahu, Exploiting the critical perioperative period to improve long-term cancer outcomes, Nat. Rev. Clin. Oncol. 12 (2015) 213. [35] J.G. Hiller, N.J. Perry, G. Poulogiannis, B. Riedel, E.K. Sloan, Perioperative events influence cancer recurrence risk after surgery, Nat. Rev. Clin. Oncol. 15 (2018) 205. [36] A.J. Templeton, et al., Prognostic role of neutrophil-to-lymphocyte ratio in solid tumors: a systematic review and meta-analysis, J. Natl. Cancer Inst. 106 (2014) u124. [37] T.F. Nishijima, H.B. Muss, S.S. Shachar, K. Tamura, Y. Takamatsu, Prognostic value of lymphocyte-to-monocyte ratio in patients with solid tumors: a systematic review and meta-analysis, Cancer Treat. Rev. 41 (2015) 971. [38] G. Jomrich, et al., High systemic immune-inflammation Iindex an adverse prognostic factor for patients with gastroesophageal adenocarcinoma, Ann. Surg. (2019) https://doi.org/10.1097/SLA.0000000000003370. [39] B. Bhaskar, J. Dulhunty, D.V. Mullany, J.F. Fraser, Impact of blood product transfusion on short and long-term survival after cardiac surgery: more evidence, Ann. Thorac. Surg. 94 (2012) 460. [40] O. Ljungqvist, M. Scott, K.C. Fearon, Enhanced recovery after surgery: a review, JAMA SURG 152 (2017) 292. [41] A.D. Altman, et al., Enhanced recovery after surgery: implementing a new standard of surgical care, CMAJ 191 (2019) E469. [42] L. Wijk, et al., International validation of enhanced recovery after surgery society guidelines on enhanced recovery for gynecologic surgery, Am. J. Obstet. Gynecol. 221 (231) (2019). [43] L.S. Prescott, et al., Choosing wisely: decreasing the incidence of perioperative blood transfusions in gynecologic oncology, Gynecol. Oncol. 153 (2019) 597.
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006
Contents lists available at ScienceDirect
Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno
Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome in patients undergoing surgery for ovarian cancer: A propensity score-matched analysis Hao Zhang 1, Xin Wu 1, Zheng Xu 1, Zhirong Sun, Minmin Zhu ⁎, Wankun Chen ⁎, Changhong Miao ⁎ Department of Anesthesiology, Fudan University Shanghai Cancer Center, Shanghai 200032, China Department of Oncology, Shanghai Medical College, Fudan University, Shanghai 200032, China
H I G H L I G H T S • Transfusion was associated with increased cancer recurrence and mortality in ovarian cancer patients. • Transfusion was associated with higher postoperative complication rates and longer hospitalization stays. • Transfusion was associated with more postoperative fluctuations in systemic inflammatory markers.
a r t i c l e
i n f o
Article history: Received 20 September 2019 Received in revised form 4 December 2019 Accepted 6 December 2019 Available online xxxx Keywords: Ovarian cancer Perioperative red blood cell transfusion Systemic inflammation Postoperative recovery Long-term outcome
a b s t r a c t Background. The impact of perioperative red blood cell transfusion (PRBCT) on cancer survival has remained controversial. Methods. We conducted a retrospective study in patients undergoing primary debulking surgery (PDS) for ovarian cancer between January 2013 and December 2017. The patients were divided into two groups based on whether they received PRBCT. Clinical characteristics were compared between groups. After propensity score matching, perioperative systemic inflammation-based scores, quality of recovery, postoperative outcomes, disease-free survival (DFS), and overall survival (OS) were compared between groups. Univariate and multivariable Cox proportional hazard models were used to evaluate the association between covariates and survival outcomes. Results. A total of 1037 patients were enrolled in this study, and 31.7% of patients received PRBCT. After propensity matching, there was no significant difference in the clinical characteristics between groups. Patients receiving PRBCT had more postoperative fluctuations in systemic inflammatory response-related indicators (P b 0.001), a higher incidence of postoperative grade II complications (28.4% vs. 14.8%), a longer length of stay (10.6 d vs. 6.2 d) and higher 30-day and total readmission rates (7.1% vs. 4.4% and 11.2% vs. 8.1%, P b 0.001, respectively) than patients who did not receive PRBCT. The OS and DFS rates 3 years after surgery were significantly lower in the patients receiving PRBCT than in patients not receiving PRBCT (58.9% vs. 74.5%, 39.6% vs. 52.3%). Conclusions. PRBCT was significantly associated with more fluctuations in systemic inflammatory indicators, a prolonged length of stay, higher postoperative complication rates and increased cancer recurrence and overall mortality in ovarian cancer patients undergoing PDS. © 2019 Published by Elsevier Inc.
1. Introduction
⁎ Corresponding authors at: Department of Anesthesiology, Fudan University Shanghai Cancer Center, Shanghai 200032, China. E-mail addresses: [email protected] (M. Zhu), [email protected] (W. Chen), [email protected] (C. Miao). 1 These authors contributed equally to this work.
Worldwide, the incidence of ovarian cancer ranks third among gynecologic tumors, and it is the second most common cause of death from gynecologic cancer [1]. Most patients suffering from ovarian cancer have progressed to an advanced stage at the time of diagnosis. The standard treatment for ovarian cancer is surgery supplemented with chemotherapy, targeted therapy and other comprehensive treatments [2,3]. Primary debulking surgery (PDS) is the cornerstone of ovarian
https://doi.org/10.1016/j.ygyno.2019.12.006 0090-8258/© 2019 Published by Elsevier Inc.
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006
2
H. Zhang et al. / Gynecologic Oncology xxx (xxxx) xxx
cancer treatment. Satisfactory tumor cytoreductive surgery is a favorable factor for prolonging survival in patients with ovarian cancer [3]. Ovarian cancer patients undergoing tumor reduction surgery often have anemia symptoms before surgery, and in order to achieve R0 resection, the surgical trauma is often substantial, which undoubtedly increases the risk of bleeding and perioperative anemia. It has been reported that perioperative blood transfusion rates in patients receiving PDS range from 19% to 56% [4,5]. For patients with severe anemia, blood transfusions can be lifesaving [6,7]. However, some studies in recent years have shown that the average length of hospital stay and the rates of postoperative infection, cancer recurrence, metastasis, and cancer-related mortality are significantly higher in cancer patients undergoing transfusion during surgery, and the cancer recurrence rate is also higher [8–10]. This may be because allogeneic blood cells are exogenous substances that induce immune responses in recipients and inhibit systematic anti-tumor immune function [11,12]. Some laboratory-based inflammatory indicators related to inflammatory and immune responses, such as the NLR (neutrophil-to-lymphocyte ratio), LMR (lymphocyte-to-monocyte ratio) and SII (systemic immune inflammation index), have been shown to be associated with prognosis in ovarian cancer [13–15]. Although there have been some studies on the long-term prognosis of perioperative transfusion in ovarian cancer patients, the conclusions are not the same [8,16]. This retrospective study was designed to compare the effects of perioperative blood transfusion on perioperative systemic inflammation indicators, postoperative recovery and long-term survival outcomes. 2. Materials and methods This study was approved by the Ethics Committee of Fudan University Shanghai Cancer Center (FUSCC), China. From January 2013 to December 2017, patients undergoing PDS for ovarian cancer were enrolled in this retrospective cohort. The exclusion criteria were emergency surgery, incomplete medical records, patients who underwent surgery for recurrent disease, treatment with anti-inflammatory agents or immunosuppressants for more than one month before surgery, patients with chronic inflammatory diseases and infections including autoimmune diseases, and preoperative neoadjuvant chemotherapy and radiotherapy. The data were extracted from the database of the FUSCC clinical information system. The medical information of each patient was reviewed and recorded, including demographic information, primary diagnosis, medical history, operative details (operative time, surgical complexity and estimated blood loss volumes), pathological details (histologic diagnosis, tumor size, tumor differentiation, and FIGO stage), preoperative Hgb level, preoperative CA125 levels, perioperative red blood cell transfusion volumes, perioperative systemic inflammation-based scores, and postoperative outcomes. The counts of neutrophils, lymphocytes, monocytes and platelets were recorded within 3 days before surgery (pre op), on the first day after surgery (POD1) and on the third day after surgery (POD3). Preoperative systemic inflammation-based scores, including the neutrophilto-lymphocyte ratio (NLR), the lymphocyte-to-monocyte ratio (LMR) and the systemic immune inflammation index (SII), were calculated based on these laboratory results. The SII was defined as follows: SII = neutrophil × platelet / lymphocyte. Surgical complexity of each patients was assigned according to a surgical complexity score system reflecting the complexity and number of surgical procedures as described by Aletti et al. [17]. The postoperative outcomes included postoperative complications (graded I–V according to the Clavien-Dindo classification system for surgical complications [18,19]), hospital length of stay (LOS), hospital readmission within 30 and 90 days, and 90-day mortality. Perioperative red blood cell transfusion (PRBCT) was defined as intraoperative transfusion and transfusion occurring during the
postoperative hospital admission. Comparisons were made between patients who received perioperative RBC transfusion (group BT) and those who did not receive perioperative RBC transfusion (group NBT). The primary outcomes of interest were postoperative recovery parameters (incidence of postoperative complications, LOS, and readmission rates) and disease-free survival (DFS) and overall survival (OS). DFS was defined as the length of time from the date of surgery to the date of the first evidence of tumor recurrence, or December 31, 2018. OS was defined as the length of time from the date of surgery to the date of death or the last follow-up date. 3. Statistical methods In this study, we compared different clinicopathologic factors between the BT group and NBT group as well as descriptive statistics including the mean and standard deviation of continuous variables such as age, body mass index (BMI), operative time, estimated blood loss, preoperative Hgb, and preoperative tumor marker level. Frequency counts and percentages were calculated for categorical variables such as the American Society of Anesthesiologists physical status score (ASA), histologic diagnosis, FIGO stage, tumor size, tumor differentiation, surgical complexity, and postoperative chemotherapy. CA125 levels were expressed as median (25%, 75% quartiles). The Chi-square test was used to evaluate the association between two categorical variables. The Kaplan-Meier method was used for time-to-event analysis, including DFS and OS. The median time to the event in months with a 95% confidence interval (CI) was calculated. The log-rank test was used to evaluate the difference in the time-to-event end points between the patient groups. Univariate Cox proportional hazards models were fitted to evaluate the effects of continuous variables on the time-toevent outcomes. Multivariable Cox proportional hazard models were used for multivariate analysis to include important and significant covariates. Statistical software SAS version 9.3 (SAS, Cary, NC) and S-Plus version 8.2 (TIBCO Software Inc., Palo Alto, CA) were used for all the analyses. A P value b 0.05 was considered statistically significant. To adjust for selection bias in the retrospective observational study, a propensity score-matching (PSM) analysis was conducted. The following prognostic covariates were included in the multi-covariate logistic model to estimate the propensity scores: age, BMI, ASA, FIGO stage, tumor size, tumor differentiation, postoperative chemotherapy, preoperative CA125 levels, residual disease and estimated blood loss (EBL). The Greedy 5-to-1-digit match algorithm was used to match the baseline covariates so that the two study groups (BT or NBT) would have similar propensity scores [20]. Among the 1037 patients enrolled in the entire study population, the propensity score was calculated for 596 patients. Of those, 298 patients were in the BT group, and 298 were in the NBT group. A total of 298 patients in the BT group with non-missing values for the selected covariates were matched on a 1:1 ratio to the patients in the NBT group with non-missing values for the same covariates. The maximum standardized differences for all covariates were 9.6% in the post-matching cohort, suggesting a substantial reduction in bias between the two groups. 4. Results A total of 1037 patients who underwent PDS for ovarian cancer were enrolled in this study, with 329 in the BT group and 708 in the NBT group, of whom 31.7% received perioperative blood transfusion. The demographics and operative characteristics of all patients are shown in Table 1. There was no significant difference in histologic diagnosis, FIGO stage and tumor differentiation between groups (P N 0.05). The patients in the BT group were significantly older than those in the NBT group (54.0 ± 11.6 vs. 52.4 ± 12.4 yr, P = 0.043), and the BT group had more patients than the NBT group with an ASA score of III–IV (20.4% vs. 15.8%, P = 0.030); RBC transfusion was also associated with higher preoperative serum CA125 levels (1472 ± 1170 vs. 1141 ±
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006
H. Zhang et al. / Gynecologic Oncology xxx (xxxx) xxx
3
Table 1 Demographics and operative characteristics of patients in both groups before and after propensity score matching. Entire study population
Age (years) BMI (kg/m2) ASA I–II III–IV Histologic diagnosis Serous histology Non-serous histology FIGO stage I II III IV Tumor size b5 cm N5 cm Tumor differentiation Well Moderate Poor Surgical complexity Low Intermediate High Residual disease No visible residual disease ≤1 cm residual disease ≥1 cm residual disease Operation time (min) Estimated blood loss (ml) Ascites (ml) b200 N200 Pre Hgb (g/dl) Pre HCT (%) Pre CA125 (U/ml) Postop Chemotherapy
Matched covariates
BT group (n = 329)
NBT group (n = 708)
52.4 ± 12.4 28.7 ± 7.7
54.0 ± 11.6 29.2 ± 7.9
242 (73.6) 67 (20.4)
596 (84.2) 112 (15.8)
235 (71.4) 94 (28.6)
518 (73.1) 190 (26.9)
29 (8.8) 38 (11.6) 158 (48.0) 104 (31.6)
63 (9.9) 79 (11.1) 343 (48.4) 223 (30.6)
198 (60.1) 131 (39.9)
439 (62.0) 269 (38)
32 (9.7) 196 (59.7) 101 (30.6)
79 (11.0) 421 (59.6) 208 (29.4)
63 (19.1) 199 (60.5) 67 (20.4)
138 (19.5) 482 (68.1) 88 (17.2)
173 (52.5) 127 (38.5) 29 (9) 203 ± 56 536 ± 326
376 (53.1) 258 (36.5) 74 (10.4) 182 ± 48 421 ± 226
83 (25.2) 46 (14.0) 105 ± 15 29 ± 5 621(195–1073) 292 (89.1)
189 (26.8) 110 (15.6) 129 ± 12 33 ± 4 412 (52–1206) 644 (91.2)
p Value 0.043 0.339 0.030
BT group (n = 298)
NBT group (n = 298)
Standardized difference (%)
54.0 ± 13.1 28.8 ± 8.4
54.4 ± 12.0 29.6 ± 7.8
3.5 2.6 1.0
256 (85.9) 43 (14.1)
253 (84.8) 45 (15.2)
198 (66.4) 100 (33.6)
203 (68.1) 95 (32.9)
32 (10.7) 32 (10.7) 134 (45.0) 100 (33.6)
35 (11.7) 31 (10.4) 143 (48.0) 89 (29.9)
186 (62.4) 112 (37.6)
189 (63.4) 116 (36.6)
24 (8.2) 172 (57.6) 102 (34.2)
24 (7.9) 190 (63.7) 84 (28.4)
57 (19.2) 190 (63.9) 51 (15.9)
58 (19.6) 191 (64.3) 49 (16.1)
162 (54.3) 109 (36.5) 27 (9.2) 196 ± 42 545 ± 321
160 (53.6) 107 (35.8) 31 (10.6) 185 ± 52 418 ± 212
76 (23.2) 39 (13.1) 124 ± 15 29 ± 4 610 (186–943) 259 (86.9)
72 (24.5) 46 (14.6) 125 ± 15 32 ± 5 577 (56–836) 262 (87.8)
0.560
0.2
0.998
0.8
0.575
5.7
0.755
3.4
0.003
1.4
0.637
b0.001 b0.001 0.824
b0.001 b0.001 b0.001 0.265
2.3
1.5 3.6 5.2
3.5 3.1 9.6 5.2
Data are shown as the mean ± SD or n (%), or median (25%, 75% quartiles). BT = Blood transfusion; NBT = No blood transfusion; FIGO stage = Federation International of Gynecology and Obstetrics; BMI = Body mass index; ASA = American Society of Anesthesiologists; P ≤ 0.05 was considered statistically significant.
829 U/ml, P b 0.001), lower preoperative Hgb levels (105 ± 15 vs. 129 ± 12 g/dl, P b 0.001), higher surgical complexity (P = 0.003), a longer operation time (203 ± 56 vs. 182 ± 48 mins, P b 0.001), and higher estimated blood loss (536 ± 326 vs. 421 ± 226 ml, P b 0.001). Because of the significant differences in baseline characteristics between the two groups, we used the propensity matching score to reduce the imbalance. After matching, 298 pairs were formed. There was no significant difference in the demographic or operative characteristics between these two groups in the matched cohort (Table 1). In BT group, 65 patients (21.8%) received only intraoperative transfusion, 201 patients (67.4%) received both intraoperative and postoperative transfusion, and 32 patients (10.7%) received only postoperative transfusion, most transfusion (80.5%) occurred within 3 day after surgery; 47 (15.8%) patients received 1–4 units of PRBCT and 251 (84.2%) patients received N5 units of PRBCT perioperatively.
After matching, there were no significant differences in the preoperative NLR, LMR and SII between the two groups (P N 0.05, Fig. 1). Compared with pre op, after the operation, the NLR and SII in the two groups significantly increased, and the LMR was significantly decreased in all patients on POD1 and POD3 (P b 0.001, Fig. 1). The NLR and SII values were slightly lower on POD3 than on POD1 but were still higher than the pre op values (P b 0.05, Fig. 1). The NLR and SII values in the NBT group on POD1 and POD3 were significantly lower than those in the BT group (P b 0.001, Fig. 1), whereas the LMR was significantly higher in the NBT group than in the BT group (P b 0.001, Fig. 1). In terms of postoperative complications, the incidences of postoperative pneumonia, wound complications, intra-abdominal infection and postoperative deep venous thrombosis/pulmonary embolus (DVT/PE) were also significantly lower in the NBT group than in the BT group [mean incidence (%): 3.20 vs. 6.22; 10.45 vs. 15.44; 2.75 vs. 5.88 and
Fig. 1. Perioperative NLR, LMR and SII in both groups.
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006
4
H. Zhang et al. / Gynecologic Oncology xxx (xxxx) xxx
Fig. 2. Type of postoperative complications (A), Clavien-Dindo classification (B), length of hospital stays (C) and rates of readmission and mortality (D) in both groups.
2.24 vs. 4.41, respectively; P b 0.001, Fig. 2A]. The incidence of postoperative vomiting was slightly higher in the BT group than in the NBT group, but there was no significant difference between the two groups [mean incidence (%): 3.90 vs. 3.20; P = 0.420, Fig. 2A]. According to the Clavien-Dindo classification system for surgical complications, the incidence of grade I postoperative complications was comparable between the two groups, and the incidence of grade II complications was significantly lower in the NBT group than in the BT group [mean incidence (%): 1.9 vs. 3.4 and 14.8 vs. 28.4, respectively; P b 0.001, Fig. 2B]. There were no severe complications in either group (ClavienDindo grade N II). The LOS in the NBT group was significantly shorter than that in the BT group [mean time (days): 6.2 vs. 10.6, P b 0.001, Fig. 2C]. The NBT group also had significantly lower 30-day and total readmission rates than the BT group [mean incidence (%): 4.4 vs. 7.1 and
8.1 vs. 11.2, respectively; P b 0.001, Fig. 2D]. There was no significant difference in the 90-day readmission rate or the 90-day mortality rate between the two groups [mean incidence (%): 2.8 vs. 3.2 and 1.5 vs. 1.6, respectively; P b 0.001, Fig. 2D]. The median follow-up time was 4.3 years among the entire cohort. The Kaplan-Meier survival curves for the propensity-matched NBT and BT groups are displayed in Fig. 3. The shapes and Cox model analysis of the curves over time suggest that the OS rate 3 years after surgery was significantly lower in patients in the BT group than in patients in the NBT group (58.9% vs. 74.5%, P b 0.001, Fig. 3A), with a hazard ratio of 2.62 (95% CI, 2.02 to 4.08; P b 0.001). The DFS rate 3 years after surgery was also significantly lower in the BT group than in the NBT group (39.6% vs. 52.3%, P b 0.001; Fig. 3B), with a hazard ratio of 3.32 (95% CI, 2.36 to 6.06; P b 0.001). Further analysis of the relationship between
Fig. 3. Kaplan-Meier curves depicting the effect of blood transfusion on overall survival (A) and disease-free survival (B) in ovarian cancer patients.
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006
H. Zhang et al. / Gynecologic Oncology xxx (xxxx) xxx
5
Fig. 4. Kaplan-Meier curves depicting a dose-response relationship for overall survival (A) and disease-free survival (B) in ovarian cancer patients.
different volumes of blood transfusion and OS and DFS suggested that the OS rate 3 years after surgery was significantly lower in patients receiving both 1–4 RBC unit transfusions and N5 RBC unit transfusions than in the patients in the NBT group (60.7% vs. 74.5%, 58.1% vs. 74.5%, P b 0.001, respectively; Fig. 4A); however, there was no significant difference in OS between the two groups of patients receiving 1–4 RBC unit transfusions and N5 RBC unit transfusions (60.7% vs. 58.1%, P = 0.708; Fig. 4A). The DFS rate 3 years after surgery was also significantly lower in the patients receiving both 1–4 RBC unit transfusions and N5 RBC unit transfusions than in the patients in the NBT group (42.3% vs. 52.3%, 38.3% vs. 52.3%, P b 0.001, respectively; Fig. 4B); however, there was no significant difference in DFS between the two groups of patients receiving 1–4 RBC unit transfusions and N5 RBC unit transfusions (42.3% vs. 38.3%, P = 0.564; Fig. 4B). The OS and DFS were compared with blood transfusion and other variables separately in a univariate Cox model and subsequently in a multivariable Cox regression. In the univariate analysis, blood transfusion, age, ASA physical status (III–IV), non-serous histology, tumor size (N5 cm), FIGO stage (III–IV), poor tumor differentiation, higher surgical complexity, residual disease, longer operation time, higher estimated blood loss, and no postoperative chemotherapy were associated with a worse impact on OS and DFS (Table 2). Blood transfusion was associated with an HR of 1.54 (1.17 to 2.03) for OS and an HR of 1.61 (1.12 to 2.17) for DFS after multivariable analysis for known confounders (Table 3). Other variables associated with a significant increase in the hazard of death after multivariable analysis included non-serous histology, poor tumor differentiation, high surgical complexity, high estimated blood loss, increased volume of PRBCT, residual disease, and ascites (N200 ml) (Table 3). 5. Discussion The impact of PRBCT on survival and cancer recurrence/metastasis has remained a topic of controversy in oncological surgery [21,22]. In this single-center, retrospective cohort study, we found that PRBCT were strongly associated with worse short- (higher incidence of postoperative complications, a prolonged LOS, and higher readmission rates) and long-term outcomes after ovarian cancer surgery. In addition, the analysis showed that non-serous histology, poor tumor differentiation, high surgical complexity, high estimated blood loss, increased volume of PRBCT, residual disease and ascites were also independent risk factors of worse OS. PDS often requires multi-organ resections and large amounts of estimated blood loss that requires PRBCT [5,23,24]. In our study, 31.7% of patients received PRBCT. Not surprisingly, a large estimated blood loss was also associated with worse outcomes, suggesting that patients who may have bled more received a blood transfusion. Some studies demonstrated that perioperative blood transfusions may adversely affect outcomes in patients with cancer, favoring tumor growth and dissemination [25,26], but some other studies reached the opposite conclusion [27,28]. The literatures on PRBCT and prognosis in patients with ovarian cancer are also inconsistent [16,29]. For instance, Connor et al. demonstrated that PRBCT is associated with poor overall survival
in advanced epithelial ovarian cancer patients [8]. While Hunsicker et al. showed that perioperative transfusion of RBCs did not increase the risk of recurrence after ovarian cancer surgery [30]. Although, our results did not suggest a dose-response relationship between PRBCT Table 2 Univariate cox proportional hazard model for overall survival (OS) and disease-free survival (DFS). Variables
OS
DFS
HR
95% CI
Age, years BMI, kg/m2
1.02 1.00
1.00–1.04 0.020 0.97–1.03 0.810
P
HR
95% CI
0.98 1.01
0.96–1.00 0.009 0.98–1.04 0.569
P
ASA I–II III–IV
Reference – b0.001 Reference – b0.001 1.56 1.21–1.97 1.42 1.18–1.69
Histologic diagnosis Serous histology Reference – b0.001 Reference – 0.004 Non-serous 1.34 1.14–1.59 1.28 1.09–1.51 histology Tumor size b5 cm N5 cm
Reference – b0.001 Reference – b0.001 2.82 2.05–3.61 2.61 1.98–3.34
FIGO stage I–II III–IV
Reference – 0.013 1.61 1.18–2.18
Reference – 0.021 1.45 1.09–1.89
Tumor differentiation Well Reference – b0.001 Moderate 1.16 1.01–1.34 1.21 Poor 1.42 1.10–1.76 1.88
b0.001 1.04–1.42 1.42–2.31
Surgical complexity Low Reference – b0.001 Reference – b0.001 Intermediate 1.74 1.17–2.33 1.61 1.08–2.24 High 2.42 1.68–3.34 2.28 1.72–2.81 Residual disease No visible residual disease ≤1 cm residual disease ≥1 cm residual disease Operation time (min) Estimated blood loss (ml) Ascites, (ml) b200 N200
Reference –
b0.001 Reference –
b0.001
1.45
1.26–1.64
1.38
1.32–1.75
1.68
1.32–1.95
1.72
1.42–1.86
1.30
1.02–1.45 b0.001 1.27
1.01–1.51 0.002
1.82
1.41–2.18 b0.001 1.87
1.37–2.39 b0.001
Reference – 0.004 1.51 1.15–2.05
Reference – 0.012 1.42 1.08–1.93
Blood transfusion No Reference – b0.001 Reference – b0.001 Yes 1.12 1.03–1.39 1.27 1.09–1.48 Postop chemotherapy Yes Reference – 0.008 No 1.41 1.20–1.62
Reference – 0.014 1.57 1.14–2.11
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006
6
H. Zhang et al. / Gynecologic Oncology xxx (xxxx) xxx
Table 3 Multivariable Cox proportional hazard model for overall survival (OS) and disease-free survival (DFS). Variables
OS
DFS
HR
95% CI
Age, years
0.97
0.94–1.01 0.455
ASA score I–II III–IV Non-serous histology
Reference – 0.103 1.03 0.94–1.02 1.23 1.04–1.46 0.004
Reference – 0.690 1.00 0.99–1.01 1.26 1.07–1.51 0.007
Tumor size b5 cm N5 cm
Reference – 0.549 1.01 0.81–1.22
Reference – 0.104 1.63 0.94–3.01
FIGO stage I–II III–IV
Reference – 0.187 1.11 0.81–2.21
Reference – 0.205 0.97 0.94–1.01
Tumor differentiation Well Reference – 0.013 Moderate 1.45 1.08–2.07 Poor 1.73 1.16–2.56
Reference – 0.003 1.38 1.06–1.79 1.52 1.17–1.91
Surgical complexity Low Reference – 0.001 Intermediate 1.74 1.10–2.74 High 2.11 1.42–2.87
Reference – 0.004 1.46 1.09–1.96 2.27 1.57–3.01
Residual disease No visible residual disease ≤1 cm residual disease ≧1 cm residual disease Operation time (min) Estimated blood loss (ml) Ascites, (ml) b200 N200
Reference –
P
HR
95% CI
0.95
0.80–1.12 0.518
b0.001 Reference –
P
b0.001
1.33
1.21–1.68
1.42
1.29–1.85
1.92
1.65–2.12
1.64
1.56–1.87
1.01
0.87–1.32 0.451
0.99
0.71–1.34 0.387
1.78
1.35–2.18 0.001
1.84
1.46–2.23 0.001
Reference – 0.015 1.60 1.09–2.33
Reference – 0.014 1.52 1.09–2.13
Blood transfusion No Reference – 0.002 Yes 1.61 1.12–2.17
Reference – 0.002 1.54 1.17–2.03
Number of pRBC 0 1–4 5 to more
Reference – b0.001 Reference – b0.001 1.32 1.22–1.65 1.26 1.12–1.47 1.45 1.34–1.76 1.41 1.32–1.62
Chemotherapy Yes No
Reference – 0.375 1.22 0.79–1.81
Reference – 0.874 1.01 0.84–1.27
and OS and DFS, but the multivariable analysis showed that the increased volume of PRBCT was an independent risk factors of worse survival. This may be caused by the different number of patients between the high and low volume subgroup of PRBCT, and these two subgroups are not matched during comparison. Allogeneic blood transfusion induces profound biology alterations on the immune system, which has been indicated as a potential mechanism for increased risk of infections, post-transfusion graft-versushost disease, and cancer recurrence [11,31,32]. Systematic inflammation and immunosuppression contribute to the malignant progression of cancer [33]. Recent studies showed that the suppression perioperative anti-tumor immune function could facilitate the formation of new metastases and promote postoperative metastasis and recurrence [34,35]. Numerous studies have also shown that some markers of hematological components, such as the NLR, LMR and SII, can comprehensively indicate the balance of the host immune system and have been
considered prognosis-related indicators in malignancies [36–38]. An increased NLR or SII and a decreased LMR are associated with an adverse prognosis in ovarian cancer [13–15]. The results of our study show that patients who receive PRBCT have more postoperative fluctuations in these systemic inflammatory markers. This suggests that PRBCT was associated with strong suppression of immune function and/or a dysregulated inflammatory response as indicated by higher complication rates and longer hospitalization stays in those who receive blood transfusions [39]. In recent years, with the better understanding of perioperative pathophysiology, the perioperative management concept of major surgery has advanced greatly. Enhanced Recovery After Surgery (ERAS) programs are becoming the standard practice in many surgical specialties [40]. ERAS is a multidisciplinary, evidence-based approach to improve perioperative outcomes and accelerate postoperative recovery [41]. Implementation of ERAS strategy in gynecological surgery has been shown to be associated an improvement in clinical outcomes, including shortened LOS [42]. Previous study showed that PRBCT could delay the return of bowel function and increase the LOS [42]. Our results also suggested that RBCT were associated with higher incidence of postoperative complications, a prolonged LOS, and higher readmission rates. These results indicated that reducing PRBCT could be an important part of the ERAS strategy for gynecologic oncology. However, more studies are needed to further clarify the criteria for PRBCT, especially postoperative blood transfusion in the ERAS strategy [43]. Our study has several strengths and limitations. To reduce significant confounding, we conducted a propensity score matching analysis after which PRBCT were still associated with worse outcomes. Limitation included first that the study was retrospective and not randomized, and the data were derived from a single center. Second, even though we conducted a propensity score matching analysis, given the multiple tumor and perioperative factors associated with receiving PRBCT, we could not avoid the possibility of residual confounding due to unmeasured confounders. Third, we did not further assess the effect of other blood products and other perioperative variables that might alter immune responses and affect long-term outcomes after surgery. In conclusion, perioperative blood transfusion was significantly associated with more fluctuations in systemic inflammatory indicators, a prolonged LOS, higher postoperative complication rates and increased cancer recurrence and overall mortality in ovarian cancer patients undergoing PDS. Well-designed prospective studies are warranted to explore the relationship between allogeneic blood transfusion and longterm outcomes in patients undergoing cancer surgery. Funding This research was supported by the National Natural Science Foundation of China (NO. 81873948, 81871590, 81871591), the National Key Research and Development Program of China (No: 2018YFC2001904), Shanghai Shenkang Hospital Development Center Clinical Science and Technology Innovation Project (NO. SHDC12018105), the Key Technology and Development Program of Shanghai (NO. 17411963400), Science and Technology Innovation Action Plan (No. 16DZ1930304). Consent for publication Not applicable. Authors' contributions H.Z., W.C. and C.M. conceived and designed the study; H.Z., Z.X. and Z.S. collected the data; H.Z., M.Z. and W.C. interpreted and analysed the data; H.Z., X.W. and Z.X. was a major contributor in writing the manuscript; M.Z., W.C. and C.M. reviewed and edited the manuscript. All authors read and approved the final manuscript.
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006
H. Zhang et al. / Gynecologic Oncology xxx (xxxx) xxx
Availability of data and materials The datasets analysed during the current study available from the corresponding author on reasonable request. Declaration of competing interest The authors declare that they have no competing interests. Acknowledgements The authors would like to thank Dr. Juan P. Cata, M.D. (Department of Anesthesiology and Perioperative Medicine, The University of Texas MD Anderson Cancer Center, Houston, USA) for his advice and comments for revising the manuscript. References [1] F. Bray, et al., Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries, CA Cancer J. Clin. 68 (2018) 394. [2] U.A. Matulonis, et al., Ovarian cancer, NAT REV DIS PRIMERS 2 (2016) 16061. [3] A. Karam, et al., Fifth Ovarian Cancer Consensus Conference of the Gynecologic Cancer InterGroup: first-line interventions, Ann. Oncol. 28 (2017) 711. [4] L.S. Prescott, et al., Perioperative blood transfusion in gynecologic oncology surgery: analysis of the National Surgical Quality Improvement Program Database, Gynecol. Oncol. 136 (2015) 65. [5] S.A. Ackroyd, et al., A preoperative risk score to predict red blood cell transfusion in patients undergoing hysterectomy for ovarian cancer, Am. J. Obstet. Gynecol. 219 (2018) 591. [6] A. Shander, M. Javidroozi, S. Ozawa, G.M. Hare, What is really dangerous: anaemia or transfusion? Br. J. Anaesth. 107 (Suppl. 1) (2011) i41. [7] B. Clevenger, S.V. Mallett, A.A. Klein, T. Richards, Patient blood management to reduce surgical risk, Br. J. Surg. 102 (1325) (2015) 1324. [8] J.P. Connor, A. O'Shea, K. McCool, E. Sampene, L.M. Barroilhet, Peri-operative allogeneic blood transfusion is associated with poor overall survival in advanced epithelial ovarian Cancer; potential impact of patient blood management on Cancer outcomes, Gynecol. Oncol. 151 (2018) 294. [9] P. Cybulska, C. Goss, W.P. Tew, R. Parameswaran, Y. Sonoda, Indications for and complications of transfusion and the management of gynecologic malignancies, Gynecol. Oncol. 146 (2017) 416. [10] C.T. Aquina, et al., Association among blood transfusion, sepsis, and decreased longterm survival after colon cancer resection, Ann. Surg. 266 (2017) 311. [11] J.P. Cata, H. Wang, V. Gottumukkala, J. Reuben, D.I. Sessler, Inflammatory response, immunosuppression, and cancer recurrence after perioperative blood transfusions, Br. J. Anaesth. 110 (2013) 690. [12] S. Hart, C.M. Cserti-Gazdewich, S.A. McCluskey, Red cell transfusion and the immune system, Anaesthesia 70 (Suppl. 1) (2015) 38e13. [13] T. Baert, et al., Influence of CA125, platelet count and neutrophil to lymphocyte ratio on the immune system of ovarian cancer patients, Gynecol. Oncol. 150 (2018) 31. [14] D. Nie, H. Gong, X. Mao, Z. Li, Systemic immune-inflammation index predicts prognosis in patients with epithelial ovarian cancer: a retrospective study, Gynecol. Oncol. 152 (2019) 259. [15] B.S. Kwon, et al., Prognostic value of preoperative lymphocyte-monocyte ratio in patients with ovarian clear cell carcinoma, J. Cancer 9 (2018) 1127. [16] B.L. Manning-Geist, et al., Infection, thrombosis, and oncologic outcome after interval debulking surgery: does perioperative blood transfusion matter? Gynecol. Oncol. 153 (2019) 63. [17] G.D. Aletti, S.C. Dowdy, K.C. Podratz, W.A. Cliby, Relationship among surgical complexity, short-term morbidity, and overall survival in primary surgery for advanced ovarian cancer, Am. J. Obstet. Gynecol. 197 (671) (2007). [18] P.A. Clavien, et al., The Clavien-Dindo classification of surgical complications: fiveyear experience, Ann. Surg. 250 (2009) 187.
7
[19] S. Cham, et al., Development and validation of a risk-calculator for adverse perioperative outcomes for women with ovarian cancer, Am. J. Obstet. Gynecol. 220 (2019) 571. [20] P.C. Austin, A comparison of 12 algorithms for matching on the propensity score, Stat. Med. 33 (2014) 1057. [21] W.B. Al-Refaie, H.M. Parsons, A. Markin, J. Abrams, E.B. Habermann, Blood transfusion and cancer surgery outcomes: a continued reason for concern, SURGERY 152 (2012) 344. [22] M. Horowitz, E. Neeman, E. Sharon, S. Ben-Eliyahu, Exploiting the critical perioperative period to improve long-term cancer outcomes, Nat. Rev. Clin. Oncol. 12 (2015) 213. [23] J.A. Gilreath, D.D. Stenehjem, G.M. Rodgers, Diagnosis and treatment of cancerrelated anemia, Am. J. Hematol. 89 (2014) 203. [24] Practice guidelines for perioperative blood management: an updated report by the American Society of Anesthesiologists Task Force on Perioperative Blood Management*, ANESTHESIOLOGY 122 (2015) 241. [25] M.J. Latif, et al., Perioperative blood transfusion has a dose-dependent relationship with disease recurrence and survival in patients with non-small cell lung cancer, J. Thorac. Cardiovasc. Surg. 157 (2019) 2469. [26] A.G. Acheson, M.J. Brookes, D.R. Spahn, Effects of allogeneic red blood cell transfusions on clinical outcomes in patients undergoing colorectal cancer surgery: a systematic review and meta-analysis, Ann. Surg. 256 (2012) 235. [27] R. Warschkow, et al., Perioperative blood transfusions do not impact overall and disease-free survival after curative rectal cancer resection: a propensity score analysis, Ann. Surg. 259 (2014) 131. [28] R. Kang, B.E. Seath, V. Huang, R.J. Barth, Impact of autologous blood transfusion on survival and recurrence among patients undergoing partial hepatectomy for colorectal cancer liver metastases, J. Am. Coll. Surg. 228 (2019) 902. [29] M. Orskov, M. Iachina, R. Guldberg, O. Mogensen, N.B. Mertz, Predictors of mortality within 1 year after primary ovarian cancer surgery: a nationwide cohort study, BMJ Open 6 (2016), e10123. [30] O. Hunsicker, et al., Transfusion of red blood cells does not impact progression-free and overall survival after surgery for ovarian cancer, Transfusion. 59 (2019) 3589. [31] K.A. Horvath, et al., Blood transfusion and infection after cardiac surgery, Ann. Thorac. Surg. 95 (2013) 2194. [32] P. Cybulska, C. Goss, W.P. Tew, R. Parameswaran, Y. Sonoda, Indications for and complications of transfusion and the management of gynecologic malignancies, Gynecol. Oncol. 146 (2017) 416. [33] F.R. Greten, S.I. Grivennikov, Inflammation and cancer: triggers, mechanisms, and consequences, IMMUNITY 51 (2019) 27. [34] M. Horowitz, E. Neeman, E. Sharon, S. Ben-Eliyahu, Exploiting the critical perioperative period to improve long-term cancer outcomes, Nat. Rev. Clin. Oncol. 12 (2015) 213. [35] J.G. Hiller, N.J. Perry, G. Poulogiannis, B. Riedel, E.K. Sloan, Perioperative events influence cancer recurrence risk after surgery, Nat. Rev. Clin. Oncol. 15 (2018) 205. [36] A.J. Templeton, et al., Prognostic role of neutrophil-to-lymphocyte ratio in solid tumors: a systematic review and meta-analysis, J. Natl. Cancer Inst. 106 (2014) u124. [37] T.F. Nishijima, H.B. Muss, S.S. Shachar, K. Tamura, Y. Takamatsu, Prognostic value of lymphocyte-to-monocyte ratio in patients with solid tumors: a systematic review and meta-analysis, Cancer Treat. Rev. 41 (2015) 971. [38] G. Jomrich, et al., High systemic immune-inflammation Iindex an adverse prognostic factor for patients with gastroesophageal adenocarcinoma, Ann. Surg. (2019) https://doi.org/10.1097/SLA.0000000000003370. [39] B. Bhaskar, J. Dulhunty, D.V. Mullany, J.F. Fraser, Impact of blood product transfusion on short and long-term survival after cardiac surgery: more evidence, Ann. Thorac. Surg. 94 (2012) 460. [40] O. Ljungqvist, M. Scott, K.C. Fearon, Enhanced recovery after surgery: a review, JAMA SURG 152 (2017) 292. [41] A.D. Altman, et al., Enhanced recovery after surgery: implementing a new standard of surgical care, CMAJ 191 (2019) E469. [42] L. Wijk, et al., International validation of enhanced recovery after surgery society guidelines on enhanced recovery for gynecologic surgery, Am. J. Obstet. Gynecol. 221 (231) (2019). [43] L.S. Prescott, et al., Choosing wisely: decreasing the incidence of perioperative blood transfusions in gynecologic oncology, Gynecol. Oncol. 153 (2019) 597.
Please cite this article as: H. Zhang, X. Wu, Z. Xu, et al., Impact of perioperative red blood cell transfusion on postoperative recovery and long-term outcome i..., Gynecologic Oncology, https://doi.org/10.1016/j.ygyno.2019.12.006