Effect of Perioperative Blood Transfusion on Postoperative Complications of Free-Flap Reconstruction for Oral Cancer: Analysis of Propensity Score-Matched Cohorts

Abstract

Background It has been confirmed in other fields that perioperative blood transfusion (PBT) will increase the incidence of complications after free-flap reconstruction and increase the risk of patients returning to the operating room within 48 hours after the initial operation. However, for head and neck tumors, whether PBT is related to postoperative complications is debatable. The aim of this study was to control the demographic and comorbidity characteristics of patients by propensity score matching (PSM) as well as to investigate the relationship between PBT and postoperative complications after oral cancer free-flap reconstruction.

Methods A total of 597 patients who underwent microvascular free tissue transfer in two top three hospitals in Southwest China from January 2015 to July 2023 were retrospectively reviewed. The study population was divided based on PBT within 24 hours of the start of the operation and to ensure homogeneity between groups by using the PSM. The primary outcomes were postoperative complications; secondary outcomes were to explore the intraoperative risk factors of PBT.

Results A total of 597 patients were included. Among them, 90 patients received a PBT. Those patients were successfully matched with 86 similar patients who did not receive a transfusion on a ratio of 1:1. This study did not find that patients receive a transfusion had a significantly higher risk of vascular pedicle thrombosis (p = 1.000), hematoma (p = 1.000), flap failure (p = 0.398), flap-related complications (p = 0.470), and other medical complications (p = 1.000). After controlling the preoperative confounding factors and adjusting the logistic regression model, it was concluded that the tumor location-mandible (odds ratio [OR] = 19.923, 95% confidence interval [CI]: 1.213–327.302, p = 0.036) and operation time (OR = 1.011, 95% CI: 1.008–1.014, p < 0.001) were the intraoperative risk factors for PBT.

Conclusion PBT is not associated with an increased probability of postoperative complications. Mandibular tumor may have a higher risk of PBT.

Keywords

blood transfusion - complication - free flap - perioperative - flap reconstruction

With the development of microsurgery, free tissue reconstruction has become the main treatment for oral and maxillofacial defects. Although the probability of successful free-flap transplantation is as high as 94 to 98.71%.[1] [2] [3] In clinical work, the occurrence of postoperative complications such as flap failure, hematoma, wound infection, lung infection, etc. are still the most concerned topic of every doctor. The reconstruction of oral and maxillofacial tissues often involves the removal of jaw bones, which requires longer operative times and higher intraoperative blood loss. Therefore, perioperative blood transfusion (PBT) is often required. Some studies have found that PBT after breast reconstruction will increase the probability of flap-related complications but will not reduce the risk of other postoperative medical complications.[4] The main reason may be related to the change of blood components during blood storage, which leads to vascular thrombosis, the release of cytokines, and the circulation overload caused by blood transfusion.[5] [6] [7] However, in the field of head and neck tumors, there are few studies on whether PBT is associated with postoperative complications such as flap complications and medical complications, and the conclusions are conflicting. Some studies believe that PBT are associated with flap complications,[4] [8] [9] [10] whereas others emphasize that PBT does not affect flap complications.[11] [12] [13] In addition to the small sample size and commonly evaluated single-center outcomes, these studies are often not limited to the head and neck, so the research results need to be further demonstrated. In this study, we used propensity score matching (PSM) to control the confounding factors of PBT and evaluated the impact of PBT on flap complications and medical complications in patients undergoing oral free-flap reconstruction.

Methods

Patients Select

This study included patients who underwent microvascular free tissue transplantation between January 2015 and July 2023 at two hospitals (Affiliated Hospital of Southwest Medical University and Zunyi Medical University Affiliated Stomatological Hospital) in Southwest China. Criteria for exclusion included preoperative severe cardiopulmonary dysfunction, blood and coagulation dysfunction, history of deep vein thrombosis, history of other malignant tumors, and incomplete electronic information data. Finally, 597 patients were included for analysis. This study had been ethically registered in Zunyi Medical University and the Affiliated Hospital of Southwest Medical University and was conducted according to the principles of the Declaration of Helsinki.

Data Collection

The following variables were collected, preoperative variables: gender, age, body mass index (BMI), history of smoking, history of drinking, history of preoperative chemoradiotherapy, a history of comorbidities (hypertension, diabetes mellitus, preoperative pulmonary disease, preoperative cardiovascular disease, preoperative cerebrovascular disease, preoperative pleural effusion, peripheral vascular disease), the American Society of Anesthesiologists (ASA) classification, preoperative red blood cell count, preoperative hemoglobin (Hgb), preoperative fasting blood glucose, tumor location; intraoperative variables: jawbone resection, titanium plate and nail placement, neck dissection, flap type, osseous free flap, operation time; outcome variables: vascular pedicle thrombosis, flap failure, hematoma, flap-related complications, wound infection, pulmonary infection, deep vein thrombosis, medical complications. In this study, flap-related complications were defined as vascular pedicle thrombosis, hematoma, and total or partial flap failure within 30 days after operation. Medical complications mainly include wound infection, pulmonary infection, and deep vein thrombosis. The study population was divided into two groups based on the use of PBT defined as transfusion within 24 hours of the start of the operation. Primary outcomes were the postoperative complications. Secondary outcomes were to identify perioperative independent risk factors for blood transfusion.

Statistical Analysis

In this study, patients were divided into two groups based on the presence of PBT. The test group were patients receiving PBT, and the control group consisted of patients who underwent free-flap reconstruction without PBT. A univariate analysis was used to compare the difference between two groups of characteristics. Binary variables were compared using the chi-square test or the Fisher's exact test as appropriate. Normally distributed continuous variables were compared with the Student's t-test, whereas non-normally distributed variables were compared using the Mann–Whitney's U test. A logical regression model was used to calculates propensity score with the following covariates that present significant differences between the groups: BMI, history of smoking, history of drinking, history of preoperative radiotherapy and chemotherapy, history of preoperative pulmonary disease, preoperative red blood cell count, preoperative Hgb. Each patient was then matched with controls on a 1:1 ratio with a narrow caliper (0.02) of propensity without replacement. After PSM, the McNemar chi-square was used to compare the proportions and the Wilcoxon signed rank test was used to compare the means to ensure the suitability and applicability of the process. The intraoperative variables and complications of all differences were included in the binary logistic regression model to observe the correlation between blood transfusion and postoperative complications as well as the independent risk factors of PBT. Statistical significance was defined as p-value less than 0.05. All analyses were performed using IBM SPSS Statistics 29 (IBM Corp., Armonk, NY).

Results

Patient Characteristics

A total of 597 patients were included, predominantly male; the median age was 58 years (interquartile range [IQR]: 50–66), and the median BMI was 22.69 (IQR: 20.88–25.00). A total of 45.9% had a history of smoking, 39.5% had a history of drinking, and 5.4% had a history of preoperative radiotherapy and chemotherapy. Tumors were mainly located in the tongue (33.5%), buccal mucosa (17.3%), and jawbone (16.9%). A total of 86.0% of ASA score was Grade II, and 90 patients (15.08%) received a PBT. There were significant differences between the PBT group and the nontransfusion group in BMI (p = 0.041), history of smoking (p = 0.001), history of drinking (p = 0.013), history of preoperative radiotherapy and chemotherapy (p < 0.001), and history of preoperative pulmonary disease (p = 0.046), preoperative red blood cells count (p = 0.028), and preoperative Hgb (p < 0.001; [Table 1]).

Table1 Demographic and clinical characteristics before propensity score matching

Characteristics[a]	All patients
	Total	Transfusion	No transfusion	p [b]
	(n = 597)	(n = 90)	(n = 507)
Male	429 (71.9%)	57 (63.3%)	372 (73.4%)	0.051
Age	58 [50, 66]	58 [50, 64]	56 [49, 66]	0.572
BMI	22.69 [20.88, 25.00]	22.11 [19.89, 24.22]	22.98 [21.11, 25.39]	0.041
Smoking	274 (45.9%)	27 (30.0%)	247 (48.7%)	0.001
Alcohol	236 (39.5%)	25 (27.8%)	211 (41.6%)	0.013
Preoperative radiotherapy and chemotherapy	32 (5.4%)	13 (14.4%)	19 (3.7%)	< 0.001
Diabetes	45 (7.5%)	7 (7.8%)	38 (7.5%)	0.925
Hypertension	129 (21.6%)	19 (21.1%)	110 (21.7%)	0.901
Pulmonary disease	91 (15.2%)	20 (22.2%)	71 (14.0%)	0.046
Pleural effusion	16 (2.7%)	4 (4.4%)	12 (2.4%)	0.441
Cardiovascular disease	197 (33.0%)	24 (26.7%)	173 (34.1%)	0.166
Cerebrovascular disease	39 (6.5%)	8 (8.9%)	31 (6.1%)	0.326
Peripheral vascular disease	88 (14.7%)	11 (12.2%)	77 (15.2%)	0.465
ASA				0.243
I	14 (2.9%)	4 (4.7%)	10 (2.5%)
II	419 (86.0%)	69 (80.2%)	350 (87.3%)
III	54 (11.1%)	13 (15.1%)	41 (10.2%)
Preoperative red blood cells (1012/L)	4.41 [4.11, 4.74]	4.40 [3.92, 4.64]	4.44 [4.18, 4.75]	0.028
Preoperative hemoglobin (g/L)	135 [126, 145]	133 [119, 135]	136 [130, 146]	< 0.001
FBG (mmol/L)	5.19 [4.70, 5.68]	5.34 [4.73, 5.71]	5.19 [4.70, 5.71]	0.419
Tumor location				0.146
Tongue	200 (33.5%)	30 (33.3%)	170 (33.5%)
Floor of mouth	78 (13.1%)	8 (8.9%)	70 (13.8%)
Gingiva	53 (8.9%)	11 (12.2%)	42 (8.3%)
Mandible	101 (16.9%)	22 (24.4%)	79 (15.6%)
Buccal mucosa	103 (17.3%)	10 (11.1%)	93 (18.3%)
Palatal mucosa	21 (3.5%)	3 (3.3%)	18 (3.6%)
Oropharynx	25 (4.2%)	2 (2.2%)	23 (4.5%)
Other parts	16 (2.7%)	4 (4.4%)	12 (2.4%)

Abbreviations: ASA, American Society of Anesthesiologists; BMI, body mass index; FBG, fasting blood glucose.

^a Data are depicted as numbers (proportions) or median (interquartile range).

^b Variables p < 0.05 are in bold.

Propensity Score Matching

The patients were matched using PSM on a ratio of 1:1; the 90 patients who received a blood transfusion were matched with 507 patients who did not, and finally 86 pairs were successfully matched. Following PSM, the two groups had no difference in demographic and clinical characteristics ([Table 2]). It had shown that there were differences in postoperative complications between the two groups, including wound infection (p < 0.001), flap failure (p = 0.047), flap-related complications (p = 0.042), and medical complications (p < 0.001). After PSM, the two groups showed that PBT were not related to postoperative complications ([Table 3]). Meanwhile, these postoperative complications were included in the logistic regression model, and the results were shown in [Table 4]. This study concluded that PBT would not lead to vascular pedicle thrombosis, hematoma, flap failure, postoperative lung infection, postoperative wound infection, postoperative deep vein thrombosis, and other complications.

Table 2 Demographic and clinical characteristics after propensity score matching

Characteristics[a]	Matched cohorts
	Total	Transfusion	No transfusion	p b
	(n = 172)	(n = 86)	(n = 86)
Male	115 (66.9%)	56 (65.1%)	59 (68.6%)	0.627
Age	59 [51, 66]	58 [50, 64]	58.5 [52, 67]	0.116
BMIc	22.17 [20.41, 24.09]	22.05 [19.73, 24.22]	22.34 [20.53, 23.90]	0.983
Smoking	55 (32.0%)	27 (31.4%)	28 (32.6%)	0.870
Alcohol	49 (28.5%)	25 (29.1%)	24 (27.9%)	0.866
Preoperative radiotherapy and chemotherapy	20 (11.6%)	10 (11.6%)	10 (11.6%)	1.000
Diabetes	11 (6.4%)	7 (8.1%)	4 (4.7%)	0.350
Hypertension	35 (20.3%)	19 (22.1%)	16 (18.6%)	0.570
Pulmonary disease	40 (23.3%)	18 (20.9%)	22 (25.6%)	0.470
Pleural effusion	9 (5.2%)	4 (4.7%)	5 (5.8%)	1.000
Cardiovascular disease	49 (28.5%)	22 (25.6%)	27 (31.4%)	0.398
Cerebrovascular disease	18 (10.5%)	8 (9.3%)	10 (11.6%)	0.618
Peripheral vascular disease	25 (14.5%)	11 (12.8%)	14 (16.3%)	0.516
ASAc				0.337
I	6 (3.7%)	4 (4.9%)	2 (2.5%)
II	126 (77.3%)	66 (80.5%)	60 (74.1%)
III	31 (19.0%)	12 (14.6%)	19 (23.5%)
Preoperative red blood cell (1012/L)	4.41 [4.05, 4.67]	4.41 [3.95, 4.66]	4.43 [4.11, 4.69]	0.273
Preoperative hemoglobin (g/L)	133 [121, 137]	134 [121, 136]	133 [122, 141]	0.615
FBGc (mmol/L)	5.19 [4.64, 5.68]	5.32 [4.72, 5.58]	5.08 [4.61, 5.70]	0.342
Tumor location				0.055
Tongue	71 (41.3%)	30 (34.9%)	41 (47.7%)
Floor of mouth	23 (13.4%)	8 (9.3%)	15 (17.4%)
Gingiva	18 (10.5%)	11 (12.8%)	7 (8.1%)
Mandible	27 (15.7%)	20 (23.3%)	7 (8.1%)
Buccal mucosa	13 (7.6%)	8 (9.3%)	5 (5.8%)
Palatal mucosa	8 (4.7%)	3 (3.5%)	5 (5.8%)
Oropharynx	6 (3.5%)	2 (2.3%)	4 (4.7%)
Other parts	6 (3.5%)	4 (4.7%)	2 (2.3%)

Abbreviations: ASA, American Society of Anesthesiologists; BMI, body mass index; FBG, fasting blood glucose.

^a Data are depicted as numbers (proportions) or median (interquartile range).

Table 3 Comparison of patients' intraoperative characteristics and characteristics after propensity score matching

Characteristics[a]	All patients				Matched cohorts
	Total	Transfusion	No transfusion	p [b]	Transfusion	No transfusion	p
	(n = 597)	(n = 90)	(n = 507)		(n = 86)	(n = 86)
Operation time (min)	465 [400, 533]	543 [469, 633]	450 [390, 505]	< 0.001	560 [480, 640]	480 [414, 541]	< 0.001
Jawbone resection	334 (55.9%)	62 (68.9%)	272 (53.6%)	0.007	58 (67.4%)	47 (54.7%)	0.085
Titanium plate and nail placement	185 (31.0%)	31 (34.4%)	154 (30.4%)	0.442	30 (34.9%)	31 (36.0%)	0.873
Neck dissection	530 (88.8%)	76 (84.4%)	454 (89.5%)	0.158	73 (84.9%)	77 (89.5%)	0.361
Flap types				< 0.001			< 0.001
ATFF	271 (45.4%)	34 (37.8%)	237 (46.7%)		33 (38.4%)	47 (54.7%)
RFFF	209 (35.0%)	20 (22.2%)	189 (37.3%)		19 (22.1%)	30 (34.9%)
FFF	84 (14.1%)	15 (16.7%)	69 (13.6%)		13 (15.1%)	4 (4.7%)
Others	33 (5.5%)	21 (23.3%)	12 (2.4%)		21 (24.4%)	5 (5.8%)
Osseous free flap[c]	94 (15.7%)	18 (20.0%)	76 (15.0%)	0.229	16 (18.6%)	6 (7.0%)	0.022
Flap-related complications	113 (18.9%)	24 (26.7%)	89 (17.6%)	0.042	22 (25.6%)	18 (20.9%)	0.470
VPT	36 (6.0%)	6 (6.7%)	30 (5.9%)	0.783	5 (5.8%)	4 (4.7%)	1.000
Flap failure				0.047			0.398
Total	43 (7.2%)	12 (13.3%)	31 (6.1%)		10 (11.6%)	5 (5.8%)
Partial	43 (7.2%)	7 (7.8%)	36 (7.1%)		7 (8.1%)	7 (8.1%)
No	511 (85.6%)	71 (78.9%)	440 (86.8%)		69 (80.2%)	74 (86.0%)
Hematoma	28 (4.7%)	6 (6.7%)	22 (4.3%)	0.489	6 (7.0%)	6 (7.0%)	1.000
Medical-related complications	233 (39.0%)	51 (56.7%)	182 (35.9%)	< 0.001	49 (57.0%)	49 (57.0%)	1.000
Wound infection	147 (24.6%)	37 (41.1%)	110 (21.7%)	< 0.001	35 (40.7%)	27 (31.4%)	0.204
Pulmonary infection	92 (15.4%)	18 (20.0%)	74 (14.6%)	0.191	18 (20.9%)	23 (20.7%)	0.371
Deep vein thrombosis	53 (8.9%)	12 (13.3%)	41 (8.1%)	0.107	11 (12.8%)	12 (14.0%)	0.823

Abbreviations: ATFF, anterolateral thigh free flap; RFFF, radial forearm free flap; FFF, fibular free flap; VPT, vascular pedicle thrombosis.

^a Data are depicted as numbers (proportions) or median (interquartile range).

^b Variables p < 0.05 are in bold.

^c Osseous free flap, include fibular free flap and Iliac muscle flap in this study.

Table 4 Multivariate regression analysis of postoperative complications

Complications	Univariate analysis	p [a]	Multivariate analysis	p
	OR (95% CI)		OR (95% CI)
VPT	0.881 (0.356, 2.180)	0.783	0.587 (0.194, 1.779)	0.346
Hematoma	0.635 (0.250, 1.613)	0.340	1.234 (0.277, 5.505)	0.783
Flap failure		0.054
Total	2.399 (1.177, 4.889)	0.016	1.908 (0.452, 8.049)	0.379
Partial	1.205 (0.516, 2.812)	0.666	0.971 (0.216, 4.361)	0.969
No	Ref		Ref
Flap-related complications	0.586 (0.348, 0.985)	0.044	1.252 (0.290, 5.406)	0.764
Wound infection	2.520 (1.575, 4.031)	< 0.001	2.072 (0.934, 4.599)	0.073
Pulmonary infection	0.684 (0.386, 1.212)	0.193	1.239 (0.579, 2.652)	0.581
Deep vein thrombosis	1.749 (0.880, 3.474)	0.111	1.374 (0.613, 3.081)	0.440
Medical-related complications	0.428 (0.272, 0.675)	< 0.001	1.207 (0.492, 2.957)	0.681

Abbreviations: CI, confidence interval; OR, odds ratio; VPT, vascular pedicle thrombosis.

^a Variables p < 0.05 are in bold.

Perioperative Risk Factors of Blood Transfusion

Following PSM, the variables that were significantly different between groups, including osseous free flap (p = 0.022), tumor location (p = 0.055), jawbone resection (p = 0.085), flap type (p < 0.001), and operation time (p < 0.001) were included in the logistic regression equation, and by adjusting the model, tumor location-mandible (odds ratio [OR] = 19.923, p = 0.036) and operation time (OR = 1.011, p < 0.001) were independent risk factors for PBT. Flap type, osseous free flap, and jawbone resection were not related to PBT after adjusting the model ([Table 5]).

Table 5 Multivariate logistic regression analysis of intraoperative risk factors of perioperative blood transfusion

Factors	(Nonadjusted model)[b]	p [a]	Model-I	p	Model-II	p
	OR (95% CI)		OR (95% CI)		OR (95% CI)
Jawbone resection	1.913 (1.185, 3.089)	0.008	1.544 (0.732, 3.258)	0.254	NA
Tumor location		0.166				0.185
Tongue	0.529 (0.160, 1.751)	0.297	6.264 (0.937, 41.875)	0.058	6.575 (0.787, 54.947)	0.082
Floor of mouth	0.343 (0.089, 1.319)	0.119	1.873 (0.261, 13.460)	0.533	2.280 (0.241, 21.541)	0.472
Gingiva	0.786 (0.212, 2.918)	0.719	5.982 (0.821, 43.561)	0.077	6.923 (0.729, 65.766)	0.092
Mandible	0.835 (0.245, 2.848)	0.774	19.959 (1.646, 242.042)	0.019	19.923 (1.213, 327.302)	0.036
Buccal mucosa	0.323 (0.087, 1.191)	0.09	4.163 (0.570, 30.419)	0.160	4.374 (0.459, 41.639)	0.199
Palatal mucosa	0.50 (0.095, 2.645)	0.415	4.715 (0.381, 58.366)	0.227	5.901 (0.390, 89.313)	0.200
Oropharynx	0.261 (0.042, 1.635)	0.151	Ref		Ref
Other parts	Ref		NA		NA
Flap types		< 0.001		< 0.001		0.994
ATFF	0.082 (0.037, 0.182)	< 0.001	0.042 (0.012, 0.145)	< 0.001	Ref
RFFF	0.06 (0.026, 0.141)	< 0.001	0.046 (0.013, 0.162)	< 0.001	1.038 (0.518, 2.080)	0.916
FFF	0.124 (0.05, 0.306)	< 0.001	0.256 (0.050, 1.325)	0.104	NA	1.000
Other	Ref		Ref		NA
Osseous free flap	1.418 (0.801, 2.510)	0.231	0.033 (0.003, 0.385)	0.007	0.000	1.000
Operation time	1.009 (1.007, 1.012)	< 0.001	1.011 (1.007, 1.014)	< 0.001	1.011 (1.008, 1.014)	< 0.001

Abbreviations: ATFF, anterolateral thigh free flap; CI, confidence interval; FFF, fibular free flap; OR, odd ratio; NA, not applicable; RFFF, radial forearm free flap.

^a Variables p < 0.05 are in bold.

^b NA: not involved, Ref: reference variable of classification covariate, No-adjusted model: single factor regression model without adjustment. Model I: adjust for jaw resection, tumor location, flap types, osseous free flap, and operation time. Model II: The adjustment variables were flap types, tumor location, and operation time.

Discussion

At present, in the field of head and neck cancer, there is controversy regarding whether PBT leads to postoperative complications after free-flap reconstruction. Zhao et al found that blood transfusion would cause postoperative wound complications and reoperation[9]; Puram et al found that blood transfusion was related to wound dehiscence, myocardial infarction, congestive heart failure, respiratory distress, and pneumonia in patients with head and neck free flaps[8]; Szakmany et al believed that blood transfusion significantly increased the mortality of patients with head and neck free flaps.[14] In recent years, studies have yielded conflicting results. PBT may not increase the risk of vascular pedicle thrombosis and affect the flap survival,[11] [15] but it may increase the occurrence of other postoperative complications.[4] [10] The latest study concluded that PBT may not increase the occurrence of medical complications through the change of blood transfusion strategy.[16] The impact of PBT on postoperative complications is still controversial. We designed this study and used PSM for the first time to explore the correlation between PBT and postoperative complications. To compare the differences between the PBT group and the nontransfusion group and to select the variables with differences between the groups for PSM, the matching variables included BMI, history of smoking, history of drinking, history of preoperative radiotherapy and chemotherapy, history of preoperative pulmonary disease, preoperative red blood cells count, and preoperative Hgb. Previous studies had shown that the patient's smoking and comorbidity affected the postoperative complications of patients with head and neck cancer reconstruction.[17] [18] [19] [20] This study matched the basic demographic and comorbidity characteristics of patients who received PBT and attempted to control the baseline confounding factors related to PBT that might affect postoperative complications, so as to achieve more objective results.

Karamanos et al described a higher incidence of reoperation within 48 hours after the first surgery in patients undergoing preoperative blood transfusion, mainly due to flap-related complications such as returning to the operating room to repair the anastomotic site and wound dehiscence.[4] Blood transfusion increases the risk of wound dehiscence, but the study focuses on breast reconstruction patients. Some scholars suggested that the occurrence of flap-related complications (thrombosis, hematoma, wound dehiscence) may be related to the longer storage time of red blood cells (> 14 days), which reduces the ability of free circulation, and thus decrease the microvascular flow and produces hypoxia that leads to thrombosis. It may also be related to possible changes in the coagulation system due to prolonged storage.[21] The author's previous research found that PBT may be related to the formation of vascular pedicle thrombosis in oral flap reconstruction. Torres et al found that red blood cell infusion did not cause vascular pedicle thrombosis. The research subjects in this study were not limited to the head and neck, and the observation time for vascular pedicle thrombosis was only 7 days.[11] However, there are also instances of thrombosis in clinical practice beyond 7 days. In this study, we focused on oral free-flap reconstruction and vascular pedicle thrombosis over a 30-day period. We used the PSM method for the first time to explore the correlation between PBT and postoperative complications of oral free-flap reconstruction. We concluded that PBT would not increase the occurrence of flap-related complications (thrombosis, flap failure, hematoma) and medical complications (wound infection, lung infection, deep vein thrombosis) in patients with oral cancer after reconstruction.

In this study, there were statistically significant differences between the two groups of patients in wound infection (p < 0.001), flap failure (p = 0.047), medical-related complications (p < 0.001), and flap-related complications (p = 0.044). We included postoperative complications into the regression equation. Multivariate analysis showed that PBT was not related to postoperative complications. Furthermore, after matching preoperative variables and comorbidities through PSM, it also showed that PBT would not lead to vascular pedicle thrombosis, hematoma, postoperative wound infection, postoperative lung infection, and other postoperative complications. It might be that we had formulated strict inclusion and exclusion criteria to exclude patients with severe liver and kidney dysfunction, hematological disorders, and coagulation disorders. PSM was used to match the preoperative red blood cells, Hgb, and comorbidities characteristics of the two groups of patients, which to some extent balanced the blood flow dynamics between the groups, made the blood viscosity consistent between the groups, and weakened the possibility of differences between the groups caused by changes in the fibrinolytic system.[22] Kim et al studied 674 patients that included head and neck and breast free flaps and believed that blood transfusion was not related to wound complications.[19] It was consistent with our research results is that PBT would not lead to postoperative wound infection (p = 0.204) and it was inconsistent with our study is that it only focuses on patients with head and neck cancer.

During the reconstruction of oral free flap, due to the longer operative times and higher intraoperative blood loss, PBT is often unavoidable. With continuous development, PBT has experienced changes from free blood transfusion strategy (Hgb < 9–10 g/dL, hematocrit [Hct] < 26–30%) to restricted blood transfusion strategy (Hgb < 7–8 g/dL, Hct < 21–25%).[23] [24] Studies have confirmed that the restrictive blood transfusion strategy will not increase the risk of flap failure and complications.[13] Our study supports the results of this study. Our blood transfusion rate was 15.08%, which is lower than the 18.07 to 58% reported in previous studies.[12] [13] [25] The two centers from which we collected data (the Affiliated Hospital of Southwest Medical University and Dental Hospital Affiliated to Zunyi Medical University) adopted a restrictive blood transfusion strategy. Doctors carefully considered blood transfusion according to the actual situation of patients and hemoglobin (Hgb < 7.0 g/dL). Our research results show that it may be slightly lower than the restrictive blood transfusion strategy and will not affect the survival of oral free flaps, increasing the risk of medical complications. This not only supports previous studies on carefully discussing blood transfusion strategies, but also provides new thinking for clinical blood transfusion strategies.

The study concluded that tumor location-mandible (OR = 19.923, p = 0.036) and operation time (OR = 1.011, p < 0.001) were identified as independent risk factors for PBT. This indicates that mandibular tumor may have a higher risk of PBT. Of course, other tumors that invade the mandible cannot be excluded. After jawbone resection, osseous free flaps need to be prepared; the operative time and blood loss increase during the operation, which increases the possibility of blood transfusion. There is no consensus on the type of flap related to PBT. Nguyen et al believed that the type of flap was not related to PBT,[12] and Zhao et al believed that osseous free flap (OR = 1.434, p = 0.01) was an independent risk factor for PBT.[26] In our study, we did not find a significant correlation between flap type, osseous free flap, and jawbone resection with PBT. It is possible that these variables are only intermediate factors within the surgical process of jawbone tumor treatment, which indirectly affect the occurrence of PBT. Of course, when considering risk factors, the impact of preoperative related risk factors and the depth of tumor invasion on PBT cannot be ignored. This study has some limitations: first, this is a retrospective study with some data are missing, and there may be information bias. Second, fistula formation and flap infection were recorded as wound infection, which prevents separate analysis of these complications. Third, a dichotomy of some variables, such as smoking, drinking, and comorbidities, into “yes” or “no” without consideration of the time period and disease severity, which may also lead to bias. Furthermore, our data come from two different centers. Although both of them are top teaching hospitals in Southwest China, and the surgeons are the discipline leaders, both of them are equipped with the best nursing team, there are slight differences between the two centers in surgical techniques and perioperative program processing, which may cause bias.

Conclusion

Restrictive PBT strategy will not increase the occurrence of postoperative flaps and medical complications, and mandibular tumor may have a higher risk of PBT.

Conflict of Interest

None declared.

Acknowledgment

None.

Authors' Contributions

Y.C. and M.T. contributed to the conception of the study. Y.C. and Y.L. contributed significantly to analysis and manuscript preparation. H.W. and X.P. performed the data analyses. D.G. and L.Z. helped perform the analyses with constructive discussions. All authors read and approved the final manuscript.

Availability of Data and Materials

The datasets generated and analyzed during the current study are not publicly available due to ownership of data but available from the corresponding author on a reasonable request.

Ethics Approval

This study is a retrospective study, without the informed consent of the patients but has passed the ethical permission of the Affiliated Hospital of Southwest Medical University and Zunyi Medical University. This research was conducted in accordance with international guidelines and the ethical standards outlined in the Declaration of Helsinki.

^* Yu Chen and Yinfu Lei contributed equally to this study.

• References

• 1 Crawley MB, Sweeny L, Ravipati P. et al. Factors associated with free flap failures in head and neck reconstruction. Otolaryngol Head Neck Surg 2019; 161 (04) 598-604
• 2 Ferrari S, Copelli C, Bianchi B. et al. Free flaps in elderly patients: outcomes and complications in head and neck reconstruction after oncological resection. J Craniomaxillofac Surg 2013; 41 (02) 167-171
• 3 Wang SJ, Shen SY, Lin B, Wang F, Yang HY. Factors affecting postoperative sleep quality of patients undergoing flap transfer for head and neck reconstruction. Oral Oncol 2022; 127 (187) 105804
• 4 Karamanos E, Shah AR, Kim JN, Wang HT. Impact of blood transfusion in free flap breast reconstruction using propensity score matching. J Reconstr Microsurg 2021; 37 (04) 315-321
• 5 Bosboom JJ, Klanderman RB, Migdady Y. et al. Transfusion-associated circulatory overload: a clinical perspective. Transfus Med Rev 2019; 33 (02) 69-77
• 6 Semple JW, Rebetz J, Kapur R. Transfusion-associated circulatory overload and transfusion-related acute lung injury. Blood 2019; 133 (17) 1840-1853
• 7 Goel R, Tobian AAR, Shaz BH. Noninfectious transfusion-associated adverse events and their mitigation strategies. Blood 2019; 133 (17) 1831-1839
• 8 Puram SV, Yarlagadda BB, Sethi R. et al. Transfusion in head and neck free flap patients: practice patterns and a comparative analysis by flap type. Otolaryngol Head Neck Surg 2015; 152 (03) 449-457
• 9 Zhao EH, Nishimori K, Brady J. et al. Analysis of risk factors for unplanned reoperation following free flap surgery of the head and neck. Laryngoscope 2018; 128 (12) 2790-2795
• 10 Grill FD, Wasmaier M, Mücke T. et al. Identifying perioperative volume-related risk factors in head and neck surgeries with free flap reconstructions - an investigation with focus on the influence of red blood cell concentrates and noradrenaline use. J Craniomaxillofac Surg 2020; 48 (01) 67-74
• 11 Torres Fuentes CE, Rodríguez Mantilla IE, Cáceres DNG, Camargo Gonzalez DF. Red blood cell transfusion and its relationship with pedicle thrombosis in microvascular free flaps. J Reconstr Microsurg 2022; 38 (05) 402-408
• 12 Nguyen A, Shin H, Saint-Cyr M, Verheyden C. Blood loss and transfusion rates in microsurgical head and neck reconstruction. Plast Reconstr Surg Glob Open 2018; 6 (11) e1988
• 13 Skoog H, Chisolm P, Altonji SJ. et al. Moving to a more restrictive transfusion protocol: outcomes in head and neck free flap surgery. Am J Otolaryngol 2022; 43 (01) 103268
• 14 Szakmany T, Dodd M, Dempsey GA. et al. The influence of allogenic blood transfusion in patients having free-flap primary surgery for oral and oropharyngeal squamous cell carcinoma. Br J Cancer 2006; 94 (05) 647-653
• 15 Lin PC, Kuo PJ, Kuo SCH, Chien PC, Hsieh CH. Risk factors associated with postoperative complications of free anterolateral thigh flap placement in patients with head and neck cancer: analysis of propensity score-matched cohorts. Microsurgery 2020; 40 (05) 538-544
• 16 Sanati-Mehrizy P, Massenburg BB, Rozehnal JM, Ingargiola MJ, Hernandez Rosa J, Taub PJ. Risk factors leading to free flap failure: analysis from the National Surgical Quality Improvement Program Database. J Craniofac Surg 2016; 27 (08) 1956-1964
• 17 Sweeny L, Curry JM, Crawley MB. et al. Age and comorbidities impact medical complications and mortality following free flap reconstruction. Laryngoscope 2022; 132 (04) 772-780
• 18 Crippen MM, Patel N, Filimonov A. et al. Association of smoking tobacco with complications in head and neck microvascular reconstructive surgery. JAMA Facial Plast Surg 2019; 21 (01) 20-26
• 19 Kim MJ, Woo KJ, Park BY, Kang SR. Effects of transfusion on free flap survival: searching for an optimal hemoglobin threshold for transfusion. J Reconstr Microsurg 2018; 34 (08) 610-615
• 20 Miller H, Bush K, Delancy M. et al. Effect of preoperative radiation on free flap outcomes for head and neck reconstruction: an updated systematic review and meta-analysis. J Plast Reconstr Aesthet Surg 2022; 75 (02) 743-752
• 21 Lee HK, Kim DH, Jin US, Jeon YT, Hwang JW, Park HP. Effect of perioperative transfusion of old red blood cells on postoperative complications after free muscle sparing transverse rectus abdominis myocutaneous flap surgery for breast reconstruction. Microsurgery 2014; 34 (06) 434-438
• 22 Jiang P, Wang Z, Yu X. et al. Effects of long-term high-altitude exposure on fibrinolytic system. Hematology 2021; 26 (01) 503-509
• 23 Franchini M, Marano G, Mengoli C. et al. Red blood cell transfusion policy: a critical literature review. Blood Transfus 2017; 15 (04) 307-317
• 24 Carson JL, Stanworth SJ, Roubinian N. et al. Transfusion thresholds and other strategies for guiding allogeneic red blood cell transfusion. Cochrane Database Syst Rev 2016; 10 (10) CD002042
• 25 Kolbenschlag J, Schneider J, Harati K. et al. Predictors of intraoperative blood transfusion in free tissue transfer. J Reconstr Microsurg 2016; 32 (09) 706-711
• 26 Zhao Y, Li X, Wang Y. et al. Maximum surgical blood order schedule for flap reconstruction in oral and maxillofacial cancer patients. BMC Oral Health 2022; 22 (01) 322

Publication History

Received: 19 February 2023

Accepted: 18 December 2023

Article published online:
03 April 2024