Risk factors and outcomes of intraoperative blood transfusion in elderly patients undergoing gastrointestinal cancer surgery

Abstract

BACKGROUND

There is an ongoing debate regarding the relationship between intraoperative blood transfusions and patient outcomes. Unifying the results is difficult because of differences in surgery type, target population and postoperative observation indicators.

AIM

To evaluate the risk factors for intraoperative blood transfusion and its impact on postoperative outcomes in elderly gastrointestinal cancer patients.

METHODS

This was a retrospective single-center study of elderly patients (≥ 65 years old) who underwent elective abdominal surgery for gastrointestinal cancer with general anesthesia. Patients with chronic kidney disease and missing related data were excluded. The primary outcomes included acute kidney injury (AKI), myocardial injury, and respiratory complications during hospitalization. Multivariate logistic regression was performed to explore the exposure-outcome relationship.

RESULTS

A total of 967 patients were included in this study. A lower preoperative hematocrit level, longer operative time (> 300 minutes) and greater amount of blood loss were observed in 145 (15.0%) patients who received blood transfusions during surgery (P < 0.0005). Among these patients, the incidences of AKI, myocardial injury and respiratory complications were 8.3% (n = 12), 5.5% (n = 8), and 15.9% (n = 23), respectively, and these values were significantly greater. Multivariate analysis revealed that receiving a transfusion was an independent risk factor for AKI, myocardial injury and respiratory complications (all P < 0.05).

CONCLUSION

These results demonstrate that intraoperative blood transfusion increases the risk of poorer outcomes in elderly patients receiving gastrointestinal cancer surgery. These findings provide new ideas for improving the prognosis of elderly cancer patients.

Key Words: Blood transfusion; Acute kidney injury; Myocardial injury; Respiratory complications; Elderly patients; Gastrointestinal cancer

Core Tip: This study retrospectively analyzes large - sample clinical data of elderly gastrointestinal tumor patients. By comprehensively analyzing incidences of postoperative complications, it evaluates the impact of intraoperative blood transfusion on the postoperative outcomes of radical resection of major gastrointestinal tumors (surgery duration > 2 hours). The research fills the void of lacking relevant clinical data support for elderly patients in this field. It offers clinical guidance, and provides evidence for transfusion and prognosis management in elderly gastrointestinal tumor patients.

INTRODUCTION

With the rapid growth of the aging population in China, the need for surgery among elderly patients has increased dramatically[1]. Gastrointestinal cancer surgery, which is the most common operation performed in elderly patients, has attracted substantial attention because of its association with high amounts of blood loss, high rates of blood transfusion during surgery and a high incidence of adverse outcomes after surgery[2]. The relationship between intraoperative blood transfusions and patient outcomes is still not clear[3,4]. Although many related studies have been conducted in the past 10 years, it is difficult to draw consistent conclusions because of differences in surgery type, target population and postoperative observation indicators. An analysis of a large multicenter database suggested that blood transfusion during or after myomectomy is associated with increased 30-day postoperative morbidity[5]. However, another study revealed that perioperative blood transfusions did not affect disease-free survival or overall survival after curative resection of intrahepatic cholangiocarcinoma[6]. Therefore, our study focused on high-risk patients who received intraoperative blood transfusions, such as elderly patients who received gastrointestinal cancer surgery[2,7].

Elderly patients have more complications, lower circulatory capacity and poorer tolerance[8]. Some studies have suggested that blood transfusion can lead to transfusion-related immune regulation through various mechanisms, such as immune suppression or immune disorders, and the degree of its impact depends on the underlying condition of the patient[9,10]. The results of animal models have revealed that mice of different ages have different immunosuppressive reactivities[11]. These findings indicate that the effect of blood transfusion on the outcome of surgery depends on the age distribution[12]. Given the group’s ability to respond to the environment differently, elderly patients should be treated as a special group to evaluate the effect of intraoperative blood transfusion on the outcomes of this group[7]. Intraoperative blood transfusion has always been a popular topic in clinical research, but few studies have been conducted in elderly populations or have observed multisystem recovery.

This study was carried out in a large tertiary general hospital. After the effects of patient demographics and preoperative clinical status were excluded, the relationship between intraoperative blood transfusion and postoperative outcomes was explored in elderly patients who received gastrointestinal cancer surgery. The risk factors for intraoperative blood transfusion were further explored.

MATERIALS AND METHODS

Approval

The Institutional Review Board for Clinical Investigations at Nanjing Drum Tower Hospital (Drum Tower Hospital, Medical School of Nanjing University, No. 321 Zhongshan Road, Nanjing 210008, Jiangsu Province, China) approved the study and agreed to waive the need for informed consent (approval No. 2017-049-01). All patient data were anonymized to ensure that personal and electronic databases were encrypted to restrict access permissions. No additional interventions were performed on the patients during the research process, and data extraction and analysis were based on existing clinical records; therefore, there are no additional ethical risks.

Study design and cohort selection

We conducted a retrospective study of elderly patients (≥ 65 years) who received major elective abdominal surgery between 2018 and 2020. The inclusion criteria were patients who received elective gastrointestinal surgery under general anesthesia, stayed for at least 1 day, had a preoperative creatinine concentration measurement and had at least one postoperative creatinine measurement. The exclusion criteria were patients with chronic kidney disease (estimated preoperative glomerular filtration rate less than 60 mL/minute/1.73 m², nephrectomy or renal transplantation) and without cancer. Patients who did not successfully complete surgery or whose related data were missing were also excluded.

Data collection

The electronic medical record system was used to collect perioperative data, which included age, sex, comorbidities, American Society of Anesthesiologists (ASA) physical status score (classified as 1 or 2, 3, and 4 or 5), preoperative laboratory values and pathological tumor stage. Heart diseases include coronary heart disease and moderate to severe valve disease. Pulmonary diseases include chronic obstructive pulmonary disease, asthma, bronchiectasis and interstitial lung disease.

Intraoperative hemodynamic data were recorded by our electronic anesthesia information management system. The collected data included the operative site, duration of surgery, amount of intraoperative blood loss and whether blood transfusion is required. None of the data could be modified manually. The type of operation was divided into 3 categories according to the operative site (stomach, colon, or rectum). Intraoperative blood loss was standardized by body weight (mL/kg). Intraoperative red blood cell transfusion was initiated when hemoglobin fell below < 7 g/dL. Transfusion decisions for patients with hemoglobin levels between 7-10 g/dL were made through comprehensive clinical evaluation by physicians.

Outcomes

Consistent with the Acute Kidney Injury Network threshold, patients were considered to have acute kidney injury (AKI) if the postoperative creatinine concentration was either more than 1.5-fold greater or 0.3 mg/dL greater than the preoperative concentration[13]. The preoperative creatinine value was defined as the last measured value prior to surgery. The postoperative creatinine value that was used was the highest concentration measured within 7 days after surgery[14]. Myocardial injury after noncardiac surgery (MINS) was defined as an increase in either fourth-generation troponin T or creatinine kinase-myocardial band above the upper limit value 7 days after the operation. The upper limit was defined as 0.03 ng/mL for troponin T and 8.8 ng/mL for creatinine kinase-myocardial band[15,16]. Postoperative respiratory complications were defined as new respiratory diseases after surgery, including hypoxemia[17], pneumonia[18], respiratory failure[19], and pulmonary embolism[20]. Hypoxemia was defined as SpO₂ < 94% when inhaling air[21]. Postoperative pneumonia was diagnosed through a combination of clinical symptoms and chest computed tomography examination[22]. Postoperative respiratory failure was defined as PaO₂ < 60 mmHg when inhaling air[23]. Pulmonary embolism was diagnosed by computed tomography pulmonary angiography[24].

Statistical analysis

Continuous variables are presented as the means ± SDs or medians (25^th and 75^th percentiles) and were analyzed via Student’s t test or the Mann-Whitney U test, as appropriate. Categorical variables were analyzed via the χ² test or Fisher’s exact test and are presented as n (%). Statistically significant continuous variables were transformed to categorical variables by their cutoff points, which were defined by receiving operating characteristic curve analysis. Patients were classified into 4 categories according to their preoperative hematocrit levels: > 30.0%, 27.0% to 30.0%, 24.0% to 26.9%, and 15.0% to 23.9%[18]. Patients with a hematocrit level > 30.0% were considered the reference group.

Primary variables were selected based on previous researches and clinical relevance. Univariate and multivariate logistic regression models were applied to evaluate the associations between intraoperative blood transfusion and postoperative outcomes. Univariable predictors with P < 0.10 were initially included, then selected via multiple stepwise regression analysis (final model P < 0.05). Baseline characteristics were compared between patients who did and did not receive intraoperative blood transfusions. Univariate associations with P < 0.10 were included in further multivariate logistic regression analysis to determine the risk factors for intraoperative red blood cell transfusion. All retained variables had variance inflation factor < 10, indicating no severe collinearity. Finally, we quantified the discriminatory power and calibration ability of the final multivariate models by the c-index and Hosmer-Lemeshow statistic, respectively[25]. All of the analyses were conducted using the software Statistical Package for Social Sciences (SPSS version 23.0, Chicago, IL, United States).

The sample size was calculated on the basis of our previous investigation and the authoritative literature. The incidence rates of MINS and AKI after noncardiac surgery (P) are 7.4% and 2.3%, respectively[26]. Approximately 400 eligible surgeries (N) were performed at this medical center. According to the formula n = 1.96² × P × (1 - P)/0.01²; n_c = n/[1 + (n - 1)/N], through screening, the sample size available for the primary analysis was 967, which was sufficient to meet the statistical needs of formula calculation. There is no prior publication of the data in this manuscript.

RESULTS

Effect of intraoperative blood transfusion on postoperative outcome

This study included 1021 patients, 54 of whom were noncancer patients or had preoperative renal function impairment. After exclusion, 967 patients were ultimately analyzed (Figure 1). In this study, the incidence rates of AKI (8.3% vs 2.7%, P = 0.001), MINS (5.5% vs 1.2%, P = 0.001) and respiratory complications (15.9% vs 7.5%, P = 0.001) significantly increased after intraoperative transfusion (Table 1). Supplementary Table 1 lists the results of the univariate analysis of postoperative complications. After adjusting for age and preoperative albumin level, intraoperative blood transfusion increased the risk of AKI by 2.975 times [adjusted odds ratio (aOR) = 2.975, 95% confidence interval (CI): 1.416-6.252; P = 0.004]. After adjusting for age, heart disease history, ASA classification, surgical site and preoperative hematocrit, intraoperative blood transfusions significantly increased the risk of MINS, especially during colon surgery (aOR = 5.109, 95%CI: 1.943-13.438; P = 0.001). After adjusting for confounding factors, such as age, history of hypertension, preoperative albumin and preoperative hematocrit, intraoperative blood transfusion was still a risk factor for postoperative pulmonary function impairment (aOR = 2.221, 95%CI: 1.225-4.025; P = 0.009) (Table 2). Independent multivariate models had good discriminatory ability (c-statistics of 0.711, 0.757, and 0.972, respectively) and calibration ability on the basis of the Hosmer-Lemeshow statistic (all P > 0.05) (Table 3).

Figure 1 Study flow chart.

Table 1 Comparison of hospital outcomes between patients with and without transfusion, n (%).

	Transfusion	Non-transfusion	P value
AKI	12 (8.3)	22 (2.7)	0.001
MINS	8 (5.5)	10 (1.2)	0.001
Respiratory complications	23 (15.9)	62 (7.5)	0.001

AKI: Acute kidney injury; MINS: Myocardial injury after noncardiac surgery.

Table 2 Multivariate logistic regression model for acute kidney injury, myocardial injury and respiratory complications.

Risk factor	OR (95%CI)	P value	aOR (95%CI)	P value
AKI
Blood transfusion (%)	3.277 (1.584-6.779)	0.001	2.975 (1.416-6.252)	0.004
Age, year	1.110 (1.055-1.168)	< 0.0005	1.101 (1.045-1.159)	< 0.0005
Preoperative albumin (g/dL)	0.924 (0.847-1.008)	0.075	Reference	0.514
MINS
Blood transfusion (%)	4.742 (1.839-12.226)	0.001	5.109 (1.943-13.438)	0.001
Age, year	1.086 (1.013-1.164)	0.020	Reference	0.063
Heart disease (%)	3.059 (1.069-8.750)	0.037	Reference	0.075
Site of procedure (%)		0.041		0.026
Stomach	Reference		Reference
Colon	2.912 (1.096-7.738)	0.032	3.437 (1.270-9.301)	0.015
Rectum	0.466 (0.057-3.812)	0.476	0.625 (0.075-5.211)	0.664
ASA score (%)	Reference	0.076	Reference	0.096
Respiratory complications
Blood transfusion (%)	2.311 (1.380-3.869)	0.001	2.221 (1.225-4.025)	0.009
Age, year	1.039 (1.002-1.077)	0.040	1.033 (0.996-1.072)	0.081
Preoperative hematocrit level (%)		0.037		0.265
> 30.0%	-		Reference
27.0%-30.0%	0.952 (0.399-2.273)	0.912	0.623 (0.251-1.594)	0.331
24.0%-26.9%	2.965 (1.415-6.216)	0.004	1.815 (0.799-4.122)	0.154
15.0%-23.9%	1.252 (0.284-5.512)	0.766	0.729 (0.155-3.430)	0.689

OR: Odds ratio; aOR: Adjusted odds ratio; CI: Confidence interval; ASA: American Society of Anesthesiologists; AKI: Acute kidney injury; MINS: Myocardial injury after noncardiac surgery.

Table 3 Effect of intraoperative blood transfusion on postoperative complications.

	Incidence (%)	aOR (95%CI)	P value	c-index	H-L statistic
AKI	8.3	2.975 (1.416-6.252)	0.004	0.711	5.360
MINS	5.5	5.109 (1.943-13.438)	0.001	0.757	2.829
Respiratory	15.9	2.221 (1.225-4.025)	0.009	0.972	6.189

aOR: Adjusted odds ratio; CI: Confidence interval; H-L: Hosmer-Lemeshow; AKI: Acute kidney injury; MINS: Myocardial injury after noncardiac surgery.

Patient characteristics

The baseline characteristics of the excluded patients were not significantly different from those of the patients who were included (data not displayed). The basic status of the elderly patients in the group was poor. A total of 80.67% (n = 780) of the patients had an ASA grade greater than II. Overall, 145 (15%) patients received intraoperative blood transfusions (Table 4).

Table 4 Baseline characteristics of blood transfusion and non-transfusion patients, n (%).

Characteristics	Transfusion (n = 145)	Non-transfusion (n = 822)	P value
Age, year	72 (68, 78)	71 (68, 76)	0.175
Males	101 (69.7)	595 (72.4)	0.500
Blood loss (mL/kg)	6.91 ± 6.40	3.26 ± 3.01	< 0.0005
Weight (kg)	61.75 ± 11.11	63.32 ± 10.13	0.089
Preoperative albumin (g/dL)	36.27 ± 3.42	38.22 ± 4.03	< 0.0005
Duration of surgery > 300 minutes	59 (40.7)	171 (20.8)	< 0.0005
Heart disease	18 (12.4)	93 (11.3)	0.702
Hypertension	66 (45.5)	353 (43.1)	0.583
Stroke	17 (11.7)	65 (7.9)	0.128
Pulmonary disease	6 (4.1)	60 (7.3)	0.164
Obesity	21 (14.5)	113 (13.7)	0.813
Cancer stage			0.016
I	14 (9.6)	170 (20.7)
II	48 (33.1)	233 (28.3)
III	72 (49.7)	350 (42.6)
IV	11 (7.6)	69 (8.4)
Preoperative hematocrit level			< 0.0005
> 30.0%	70 (48.3)	752 (91.5)
27.0%-30.0%	34 (23.4)	43 (5.2)
24.0%-26.9%	26 (17.9)	22 (2.7)
15.0%-23.9%	15 (10.3)	5 (0.6)
ASA score			0.325
2	32 (22.1)	155 (18.9)
3	107 (73.8)	646 (78.6)
4	6 (4.1)	21 (2.6)
Site of procedure			0.001
Stomach	436 (53.0)	99 (68.3)
Colon	235 (28.6)	34 (23.4)
Rectum	151 (18.4)	12 (8.3)

ASA: American Society of Anesthesiologists.

Factors related to intraoperative blood transfusion

Intraoperative blood transfusions were more common among patients who had greater blood loss (6.91 vs 3.26, P < 0.0005), lower preoperative albumin values (36.27 vs 38.22, P < 0.0005) and an operative length > 300 minutes (40.7% vs 20.8%, P < 0.0005). Moreover, patients who received blood transfusions typically had lower preoperative hematocrit levels and more advanced cancer stages (all P < 0.05) (Table 4). The predictive model of intraoperative blood transfusion is shown in Table 5. After excluding the effects of preoperative nutritional status, disease stage and surgical site, a longer duration of operation (> 300 minutes), greater blood loss and preoperative hematocrit level < 30.0% were still independent risk factors for receiving an intraoperative blood transfusion (P < 0.0005). On the basis of the Hosmer-Lemeshow statistic (P > 0.05), the discriminatory ability (c statistic = 0.883) and calibrating ability of the model were good.

Table 5 Risk factors for intraoperative blood transfusion.

Risk factor	aOR (95%CI)	P value
Duration of surgery > 300 minutes (%)	2.738 (1.641-4.568)	< 0.0005
Blood loss (mL/kg)	1.230 (1.161-1.303)	< 0.0005
Preoperative hematocrit level (%)		< 0.0005
> 30.0%	Reference
27.0%-30.0%	12.040 (6.540-22.166)	< 0.0005
24.0%-26.9%	16.408 (8.115-33.177)	< 0.0005
15.0%-23.9%	36.584 (11.003-121.638)	< 0.0005

aOR: Adjusted odds ratio; CI: Confidence interval.

DISCUSSION

A total of 967 elderly patients who received elective gastrointestinal cancer surgery were enrolled in this study. The incidence of intraoperative blood transfusion was 15.0% (n = 145). We found that the duration of surgery, intraoperative blood loss and preoperative hematocrit level were all risk factors for receiving an intraoperative blood transfusion. Therefore, correction of preoperative anemia is necessary to reduce the need for intraoperative blood transfusions. Targeted interventions for preoperative anemia are carried out on the basis of the etiology and laboratory indicators, such as hemoglobin and the serum iron concentration[27]. Common intervention strategies include the use of iron supplements to treat iron deficiency anemia and folic acid to treat megaloblastic anemia[28]. Although there is no consistent conclusion regarding the correlation between cancer stage and intraoperative blood transfusion, most studies have shown that there is a positive correlation. However, in our multivariate analysis, advanced cancer stage was not found to be an independent risk factor for intraoperative blood transfusion.

Hypoalbuminemia (< 3.5 g/dL) reduces the stability of red blood cell membranes and shortens the lifespan of red blood cells[29]. Low nutritional status increases the need for blood transfusion, leading to a decrease in the hemoglobin threshold used to determine whether a patient requires a blood transfusion[30]. However, in this study, after the influence of confounding factors was removed, sex and preoperative nutritional status did not independently increase the risk of intraoperative blood transfusion.

Consistent with previous studies, our multivariate model revealed that age and blood transfusion were independently associated with AKI[31,32]. However, the results could not explain the relative importance due to the possibility of synergy between them. In this study, intraoperative blood transfusion increased the risk of AKI by 2.975-fold. Blood transfusion can cause ischemia-reperfusion injury, leading to microcirculatory disorders and inflammatory storms and increasing the risk of AKI[33]. Older age may further increase the correlation between intraoperative blood transfusion and AKI.

It has been reported that blood transfusions can prevent tachycardia caused by anemia to reduce the risk of myocardial injury[34]. In contrast, we were surprised to find that intraoperative blood transfusion increased MINS. Blood transfusion increases the risk of MINS by 5.109-fold. This result may be because blood transfusion activates the systemic inflammatory response by releasing proinflammatory mediators, leading to coronary endothelial dysfunction[35]. Moreover, the levels of complement C3a/C5a increase after blood transfusion, inducing the aggregation of neutrophils and platelets and promoting the formation of microthrombi[36]. In addition, after blood transfusion, the hemorheological properties of patients change, which affects the metabolism of myocardial cells and exacerbates myocardial oxidative stress[37]. Elderly patients have insufficient vascular endothelial function reserve, and they are more likely to suffer from coronary artery spasm induced by metabolic disorders associated with nitric oxide after blood transfusion, thus triggering myocardial injury with MINS[38].

There is controversy in the literature regarding the relationship between blood transfusion and respiratory complications. Some studies suggest that blood transfusion may increase the risk of acute respiratory distress syndrome through inflammatory reactions, whereas other studies support the benefits of correcting anemia and improving the supply of oxygen. This study suggests that intraoperative blood transfusion increases the risk to the respiratory system by 2.221-fold. During processing and storage, erythrocytes undergo ATP and 2,3-diphosphoglycerate depletion, resulting in a reduction in plastic deformation capacity[39-42]. These changes reduce their oxygen-carrying capacity[31]. The oxygen dissociation curve shifts to the left, and oxygen utilization decreases, exacerbating tissue hypoxia injury and damaging microvascular oxygenation. In addition, injecting red blood cells to remove NO (a vasodilator) results in the contraction of small vessels[43]. All of these factors ultimately lead to tissue ischemia[44,45]. Moreover, the inflammatory response increases capillary permeability, on the basis of which plasma colloid osmotic pressure and blood volume increase, followed by an increase in transfusion-associated circulatory overload[46,47]. Elderly patients seem to have a greater risk of circulatory overload, which is accompanied by more complications, less circulatory capacity and poorer tolerance[8].

The effect of intraoperative blood transfusions on postoperative outcomes may be directly attributable to red blood cells or an indirect manifestation of intraoperative hypoperfusion or acute ischemia and hypoxia[48]. Our multivariate model indicated that blood loss, length of operation and preoperative anemia were risk factors for intraoperative blood transfusion, so we cannot completely exclude the potential impact of these factors on postoperative outcomes. In the present study, intraoperative blood transfusion was an independent risk factor for AKI, MINS and respiratory complications. The incidence of complications related to blood transfusion during hospitalization is low, and one of the main reasons for this low incidence is a lag in diagnosis[46]. As a retrospective study, none of our patients were diagnosed with transfusion-related lung or kidney injury, while our results were consistent with those of previous studies.

Because the preoperative conditions of patients are difficult to change, doctors could reduce the probability of intraoperative blood transfusion by using some of the factors under comprehensive consideration. On the basis of our findings, we suggest a series of measures to reduce the need for intraoperative blood transfusion, including strengthening clinical techniques, improving operative procedures, shortening the operative time and controlling bleeding. It was also necessary to actively correct preoperative anemia. Preoperative use of erythropoietin has been reported to ameliorate preoperative anemia and thus reduce the risk of intraoperative blood transfusion. However, it is still unclear whether this drug should be used for preoperative pathological anemia in cancer patients.

If there is a need to improve anemia rapidly when the cancer has progressed to a very serious stage, in addition to transfusion support, pharmacological hemoglobin-elevating therapies can be considered, such as erythropoiesis-stimulating agents and intravenous iron, or symptomatic supportive treatments, such as oxygen therapy and glucocorticoids, for immediate intervention[49]. Recent studies have highlighted the therapeutic potential of hemoglobin stabilizers (e.g., hypoxia-inducible factor stabilizers) and alternative strategies to transfusion (e.g., artificial oxygen carriers) in cancer-related anemia. However, further research is needed to explore their value in emergency settings[5].

The results of this study provide new directions for future research on transfusion strategies for elderly patients. Further research can explore the mechanisms underlying the observed associations, such as the role of transfusion-related inflammatory responses or microvascular changes. In addition, prospective studies or randomized controlled trials can help establish causal relationships and provide information for transfusion guidelines suitable for elderly patients. Regarding academic positioning, this study is a continuation and deepening of previous work. Previous studies focused on the postoperative outcomes of elderly patients undergoing major abdominal surgeries but did not conduct in-depth analyses of this special subgroup of elderly cancer patients, who should be studied independently as a special group owing to their pathological and physiological characteristics.

There are several inherent limitations of retrospective studies, including differences in anesthesia management practices, lack of uniform blood transfusion standards and failure to explain all potential confounding variables. First, we collected the baseline characteristics of the patients as much as possible. Second, compared with the results of other studies, the homogeneity of the patient group was reduced, resulting in deviations from individual differences; thus, the sample size was limited. Third, this was a single-center study. The bias in selecting patients and surgical types limits our results from being extended to other groups and medical centers. Furthermore, we cannot completely exclude the potential impact of intraoperative and postoperative circulatory changes (including the use of vasoactive drugs). Moreover, the transfusion units used in this study were not recorded. Although the lack of data on transfusion units may affect the analysis of dose-response relationships, the core conclusion of this study (the association between transfusion and outcomes) remains reliable. Subsequent studies will prospectively record blood transfusion volumes (accurate to units) and use standardized transfusion report templates to address this limitation. Finally, the causal relationship between intraoperative blood transfusion and postoperative outcomes cannot be determined. Prospective clinical trials are needed to further prove whether restricting intraoperative blood transfusion actually reduces the incidence of complications during hospitalization. The goal is to develop better individualized intraoperative blood transfusion protocols[50]. Meanwhile, future prospective studies would benefit from adopting the Clavien-Dindo classification to enable more granular and internationally comparable analysis of complication severity. Incorporating time-to-event (survival) analysis in prospective designs will also provide critical insights into the temporal dynamics of postoperative complications.

CONCLUSION

In conclusion, in the present study, with many patients, we found that intraoperative blood transfusion increased the risk of poorer outcomes in elderly patients (aged ≥ 65 years) who underwent gastrointestinal cancer surgery. Improvements in preoperative anemia, improvements in clinical techniques and the development of accepted guidelines are necessary to reduce the need for intraoperative blood transfusions.