Background

Total joint arthroplasty (TJA), encompassing total hip arthroplasty (THA) and total knee arthroplasty (TKA), is a surgical procedure that aims to relieve pain, improve function, and correct deformity [1]. With the aging of the population, the number of TJA surgeries is expected to increase by 174 to 673% by 2030 [2, 3], leading to greater blood loss and a higher demand for transfusion [4]. In China, the annual growth rate of primary THA is 25 to 30% [5], while in the United States, orthopedic surgery accounts for 10% of all red blood cell transfusions in a year, with 39% used for TJA patients [6]. Pennestrì reported that 12 to 22% of patients required transfusion in THA and TKA [7], while Menendez found that the transfusion rate of TKA was as high as 11%, and that of THA was as high as 15.9% [8]. Although the perioperative transfusion rate in TJA patients remains high, transfusion can increase the risk of perioperative complications [9,10,11]. Engoren revealed that blood transfusion in THA patients was linked with an increased risk of mortality within 90 days post-surgery [9]. Carson found that transfusion increased the risk of severe bacterial infection by 35% and the risk of pneumonia by 52% in THA patients [10]. In a large retrospective transfusion in noncardiac surgery increased by 29%, and the chances of pulmonary, septic, wound, or thromboembolic complications increased by 40–90% [11].

Additional allogeneic blood transfusions are an effective way to improve anemia status but raise concerns regarding viral transmission, possible immunosuppression and transfusion reactions [12]. Xiong highlighted preoperative anemia as a significant risk factor for perioperative transfusion in TJA patients [13]. Moreover, they found a direct correlation between the severity of preoperative anemia and the increased likelihood of requiring transfusion during the perioperative period [13]. Malnutrition is also considered a significant contributor to preoperative anemia among TJA patients [13]. Albumin, a key serum protein, is a widely utilized marker for assessing nutritional status [14]. Composing approximately 80% of the colloid osmotic pressure in plasma, albumin facilitates drug and endogenous compound transport, serves as an effective plasma buffer, possesses notable antioxidant capabilities, and supports microvasculature integrity [15, 16]. Currently, only a restricted number of studies have investigated the correlation between preoperative albumin levels and perioperative transfusion among TJA patients [17,18,19]. Cavazos noted that preoperative hypoalbuminemia increased the perioperative transfusion rate in TKA patients [17]. Hur further highlighted those patients with hypoalbuminemia had a fourfold higher probability of perioperative transfusion in unicompartmental knee arthroplasty (UKA) [18].

Numerous contemporary studies highlight the impact of preoperative low albumin levels and perioperative transfusions on postoperative complications of TJA. However, the association between serum albumin reduction and transfusion remains inadequately explored. In this study, we proposed the following hypothesis: the decrease in preoperative serum albumin levels is associated with a higher probability of requiring perioperative transfusion in TJA patients. Furthermore, we anticipated that a lower preoperative serum albumin level would lead to a higher risk of perioperative transfusion.

Materials and methods

Inclusion and exclusion criteria

Inclusion

All patients who received TJA from January 1, 2,017, to January 1, 2,022 in our hospital.

Exclusion

Patients who had no preoperative liver function test records; and patients with tumors or tuberculosis (Fig. 1).

Fig. 1
figure 1

Research scheme of PSM analysis of Low albumin group and Normal albumin group in TJA patients. AUC: areas under the curve; CI: confidence interval; TJA: total joint arthroplasty; P < 0.05 was considered statistically significant

Full size image

Research method

Receiver operating characteristic (ROC) curves were constructed to establish thresholds for serum albumin levels categorization. TJA patients with documented blood transfusions (primarily red blood cell transfusions) were assigned to the transfusion group, while those who did not receive transfusions during the perioperative period constituted the non-transfusion group. Subsequently, we conducted a multivariate binary logistic regression analysis to examine the results. This study has been approved by the Ethics Committee of Army Medical Center of PLA of the authors’ affiliated.

Data collection

We retrieved clinical information for our study from the electronic medical records system. Our data collection primarily focused on patients’ hospital identification numbers, sex, body mass index (BMI), age, and diagnoses (including osteoarthritis [OA], rheumatoid arthritis [RA], and bone fractures). Additionally, we documented any preexisting comorbidities among the patients, such as coronary heart disease (CHD), chronic bronchitis, diabetes, hypertension, cerebral infarction, cancer, renal failure, corticosteroid use, chronic obstructive pulmonary disease (COPD), alcohol consumption, smoking, and any major surgeries within the past 12 months. Furthermore, we conducted laboratory tests, including ABO blood typing (AB, A, O, B) and liver function (serum total protein, albumin, globulin) analyses, and recorded instances of blood transfusion during surgery.

Data analyses

We conducted statistical analysis using SPSS 26.0 software (IBM Corp., Armonk, NY, USA). For categorical data, we employed Chi-square tests or Fisher’s exact tests, expressing the results as percentages (%). To account for confounding factors, we established a logistic regression model using propensity score matching (PSM). The model used preoperative serum albumin as the dependent variable and included covariates such as BMI grade, age grade, sex, diagnosis, hypertension, diabetes, coronary heart disease, chronic obstructive pulmonary disease, chronic bronchitis, cerebral infarction, major surgeries within the last 12 months, renal failure, cancer, depression, corticosteroid use, smoking, drinking, and blood type, along with perioperative blood transfusion-related variables. We matched the normal serum albumin and the low serum albumin group at a ratio of 2:1, with a caliper value of 0.2. We assessed the PSM’s matching effectiveness using the standardized difference method. Subsequently, we conducted binary logistic regression analysis on both groups, computing the adjusted odds ratio (OR) and 95% confidence interval (CI) to explore the relationship between preoperative serum albumin levels, severity of low serum albumin, and perioperative blood transfusion, as well as transfusion volume in TJA patients. During our analysis, we addressed multicollinearity, particularly between low serum albumin, perioperative blood transfusion, and albumin reduction. The VIF between low serum albumin and severity of low serum albumin was 6.404, with a tolerance of 0.156. Although these values do not indicate severe multicollinearity (VIF > 10, tolerance < 0.05), they suggest a notable degree. Therefore, we conducted separate logistic regression analyses for these two variables to ensure reliable results. A significance level of P < 0.05 was used to denote statistical significance.

Results

Patients’ baseline characteristics

Based on the ROC curve constructed using preoperative serum albumin levels and the perioperative blood transfusion rate, we identified an area under the curve (AUC) of 0.601 (P < 0.001, 95% CI 0.573–0.629), with an albumin cutoff value of 37.3 g/L (Fig. 2). Among the 2,772 TJA patients, 2,462 patients were enrolled. The mean age of the TJA patients was 63.54 ± 11.71 years (Table 1). There were 924 men patients and 1,538 women patients. Among them, there were 1,131 patients with OA, 1,003 with RA, and 328 with fractures (Table 2). Notably, 939 (38.16%) of our patients exhibited serum albumin levels below 37.3 g/L.

Fig. 2
figure 2

Diagnostic performances of the serum albumin for predicting perioperative blood transfusion among TJA patients. OR: Odds ratio; PSM: propensity score matching; TJA: total joint arthroplasty; CI: confidence interval; P < 0.05 was considered statistically significant

Full size image
Table 1 Baseline characteristics of the patients before PSM
Full size table

Comparison of perioperative blood transfusion-related variables before and after PSM in TJA patients

Before PSM, we meticulously matched several variables, including BMI grade, age grade, sex, diagnosis, hypertension, diabetes, CHD, COPD, chronic bronchitis, cerebral infarction, major surgeries within the last 12 months, renal failure, cancer, depression, corticosteroid use, smoking, drinking, and blood type. Among the 18 indicators, significant differences emerged in BMI grade (P < 0.001), age grade (P < 0.001), hypertension (P = 0.001), renal failure (P = 0.001), and corticosteroid use (P < 0.001) across the groups. Subsequently, following successful PSM matching, we achieved 892 matched pairs in the low serum albumin group and 1401 matched pairs in the normal serum albumin group. Significantly, the P values for all 18 variables were above 0.05, suggesting no statistically significant differences between the two datasets. This outcome underscored the comparative balance achieved through PSM, demonstrating the comparability of the groups (Table 2).

Table 2 Distribution characteristics of covariates in TJA patients before and after PSM in low albumin group and normal albumin group
Full size table

Correlation between the extent of preoperative serum albumin reduction and perioperative blood transfusion in TJA patients

There were 568 cases of perioperative blood transfusion, accounting for 23.08% of TJA patients. In the normal serum albumin group, the transfusion rate was 18.9%, while in the low albumin group, it was 29.54%. The transfusion rate increased from 18.9% in the normal albumin group to 60% in the low albumin group (≤ 25 g/L). Both before and after PSM, as the preoperative serum albumin level decreased, the perioperative blood transfusion rate of TJA patients gradually increased (Table 3).

Table 3 The degree of low serum albumin and the incidence of perioperative blood transfusion in TJA patients before and after PSM
Full size table

The correlation between the severity of hpreoperative low albumin levels and perioperative blood transfusions in TJA patients

After PSM, we discovered the relationship between albumin values and perioperative blood transfusion, yielding an OR for albumin of 0.911 and a 95% CI of (0.888, 0.935), P < 0.001. As every 1 g increase in preoperative serum albumin levels, the likelihood of perioperative blood transfusion decreases by 0.911 times. Based on the cutoff value for serum albumin, we categorized the patients into the normal serum albumin group (≥ 37.3 g/L) and the low serum albumin group (< 37.3 g/L). Binary logistic regression analysis revealed that, for TJA patients, the perioperative transfusion rate in the low serum albumin group increased by 1.83 times in comparison to the preoperative normal serum albumin group, exhibiting a 95% CI of (1.50–2.23), P < 0.001 (Fig. 3).

Fig. 3
figure 3

Binary logistic regression analysis of preoperative low serum albumin and perioperative blood transfusion in TJA patients after PSM. OR: Odds ratio; PSM: propensity score matching; TJA: total joint arthroplasty; CI: confidence interval; P < 0.05 was considered statistically significant

Full size image

Patients were classified on the basis of serum albumin levels into four groups: ≥37.3 g/L, 30–37.3 g/L, 25–30 g/L, and ≤ 25 g/L. Post-PSM, among TJA patients, when compared to those in the preoperative serum albumin group ≥ 37.3 g/L, the transfusion rates during the perioperative period increased by 1.63 times (95% CI: 1.37–1.99, P < 0.001), 5.41 times (95% CI: 3.08–9.50, P < 0.001), and 6.43 times (95% CI: 1.80-22.96, P = 0.004), respectively, in the 30–37.3 g/L, 25–30 g/L, and ≤ 25 g/L serum albumin groups (Fig. 4).

Fig. 4
figure 4

Binary logistic regression analysis of the degree of preoperative low serum albumin and perioperative blood transfusion in TJA patients after PSM

Full size image

Discussion

We utilized ROC curves to identify the cutoff value of preoperative serum albumin at 37.3 g/L for perioperative blood transfusions in TJA patients. As every 1 g increase in preoperative serum albumin levels, the likelihood of perioperative blood transfusion decreases by 0.911 times among TJA patients. Furthermore, we found that, compared to the normal serum albumin group, the overall perioperative transfusion rate increased by 1.83 times in the low albumin group among TJA patients. Additionally, in TJA patients, the perioperative transfusion rate increased by 1.63 times, 5.41 times, and 6.43 times in the 30–37.3 g/L, 25–30 g/L, and ≤ 25 g/L serum albumin groups, respectively, compared to the preoperative ≥ 37.3 g/L serum albumin group (Fig. 4). In summary, our study indicates that TJA patients with lower preoperative serum albumin levels are at an increased risk of perioperative blood transfusion.

The connection between reduced serum albumin and perioperative blood transfusion

Traditionally, serum albumin levels < 35 g/L are set as the sole parameter for malnutrition [20], yet the cutoff value of 37.3 g/L in our study is higher than the traditional threshold of malnutrition. Black discovered that albumin levels below 35 g/L increase the risk of 90-day readmission by 1.5 times in TJA patients [21]. They suggested that an albumin threshold of 35 g/L might not identify some high-risk patients and proposed an optimal albumin threshold of 39.4 g/L for the 90-day readmission of TJA patients [21]. Furthermore, Kudsk established a preoperative albumin level < 42.5 g/L as an independent determinant for hospitalization and severe postoperative complications [22]. Magovern reported that there is an elevated risk of perioperative transfusions when the preoperative albumin falls below 40 g/L in patients with additional comorbidities [19]. Nelson observed that serum albumin levels below 30 g/L were associated with increased incidence rates of complications, with ORs ranging from 1.68 to 3.58 [20]. Our study further revealed that at serum albumin levels between 30 g/L and 37.3 g/L, the risk of perioperative transfusion increased by 1.63 times compared to that in the normal serum albumin group. Hence, surgeons need to be vigilant when the preoperative serum albumin drops below 37.3 g/L in TJA patients.

Cavazos identified preoperative albumin levels as a risk factor associated with transfusion rates in TKA patients [17]. However, they only proposed a relation between albumin and perioperative blood transfusions without exploring the specific connection between albumin levels and perioperative blood transfusions. Hur observed a fourfold increase in the likelihood of receiving transfusions for patients with hypoproteinemia (< 3.5 g/dL) [23] undergoing UKA surgeries [18]. Similarly, Nelson noted a 1.56 times higher likelihood of transfusion in TKA patients with low serum albumin [20]. Consistent with these findings, our study revealed that TJA patients with low albumin levels exhibited an overall 1.83 times higher risk of perioperative transfusions.

Most studies report an association between albumin levels and blood transfusion. However, the study by Itagaki et al., which involved 23 patients with severe limb injuries, found no significant association between albumin levels and perioperative blood transfusion [24]. Due to the small sample size and the specific focus on patients with acute severe trauma, the generalizability and reliability of these findings may be limited.

The reasons for increased perioperative transfusion risk due to lower preoperative albumin

Our patient population showed a 38.16% prevalence of serum albumin below 37.3 g/L, a common finding in individuals undergoing TJA, who often present high-risk characteristics related to nutritional imbalances. Generally, our TJA patients were older, had multiple comorbidities and reduced mobility. Reports indicate that malnutrition prevalence among TJA recipients varies widely, ranging from 8.5 to 50% [25, 26], with its occurrence escalating with advancing age [27]. Multiple factors contributed to the low albumin levels observed in our patients. First, chronic inflammation plays a significant role. Patients with OA and RA in our cohort often suffered from end-stage conditions, characterized by prolonged disease durations spanning several years to decades. These conditions typically involve systemically and locally detectable concentrations of inflammatory mediators [28,29,30]. Tumor necrosis factor (TNF) and interleukin-6-mediated inflammation might impact hepatic albumin gene transcription [31], while inhibitory substances linked to the inflammatory state, such as TNF and interleukin-1(IL-1), can disrupt albumin synthesis [32]. Additionally, inflammation can significantly increase albumin permeability at the blood-joint barrier, causing substantial albumin uptake [33], thereby contributing to reduced serum albumin levels. Second, distinct dietary habits, with a preference for carbohydrate-rich diets in Asian populations compared to meat-centric Western diets, could explain the lower preoperative serum albumin levels among Asians [34]. Last, in TJA patients with fractures, the cause of decreased albumin levels may also stem from posttraumatic metabolic activation, leading to a catabolic state that results in albumin depletion [35].

There are three key reasons for the impact of decreased serum albumin on the heightened risk of perioperative transfusions. First, as demonstrated by Rasmussen, reduced albumin levels can impair clotting ability, leading to increased bleeding during major noncardiac surgeries [36]. Second, albumin plays a crucial role in maintaining osmotic pressure inside and outside blood vessels. Decreased serum albumin levels may result in a reduced effective blood volume within the vessels, leading to a more significant decline in hemoglobin levels following the same amount of bleeding in affected patients [37]. Hence, individuals with lower albumin levels experience a faster drop in hemoglobin compared to those with normal albumin levels when faced with an equivalent blood loss scenario [38]. Finally, Röhrig discovered a modest to moderate correlation between albumin and hemoglobin levels, increasing the likelihood of hypoproteinemia patients developing anemia by 2.6 times [39]. Consequently, individuals with hypoproteinemia are at an increased risk of developing anemia, consequently elevating the likelihood of requiring perioperative transfusion [13].

The necessity of preoperative optimization of albumin in TJA patients

Patients receiving blood transfusions during TJA experience significantly longer hospital stays than those who do not receive transfusions [17]. In addition to the risk of viral transmission, transfusion reactions, and immunosuppression [12], transfusions can also lead to elevated postoperative mortality rates [9] and other adverse complications. Consequently, addressing patients’ preoperative nutritional status becomes imperative. Statz revealed that optimization of the preoperative nutritional status in THA patients decreased the likelihood of infection or complications by over twofold compared to that in THA patients without preoperative optimization [40]. Burgess emphasized the potential benefits of optimizing preoperative nutritional status in mitigating surgical stress responses, especially in malnourished, frail, or elderly individuals [41]. Johnson recommended that elective surgery be postponed for malnutrition (< 35 g/L) and that the patients should be referred to a nutritionist for optimization until the serum albumin exceeds 35 g/L [42]. Our findings suggest that preoperative albumin transfusions may benefit patients with hypoproteinemia. However, due to the high cost and potential risks associated with allergic reactions, renal dysfunction, and cardiac complications, it is wise to recommend albumin transfusions for only high-risk patients, particularly elderly individuals undergoing TJA [43, 44]. After nutritional support begins, the serum albumin concentration gradually returns to normal after three weeks [45]. Given that albumin has a half-life of 20 days, reflecting dietary intake from the previous 3 weeks [25], optimizing albumin levels through nutritional support one month prior to elective TJA surgery is a reasonable and feasible option for our cohort of end-stage OA patients undergoing nonemergency surgeries.

Nonetheless, this investigation has certain limitations. Firstly, its retrospective nature and the presence of incomplete data in certain cases present challenges. Despite our efforts to minimize sample loss during the matching process, a small proportion of cases (6.8%) were not included in the final analysis, which may impact the generalizability of our findings. Additionally, as a single-center study, further validation with a larger sample size is needed. Finally, the role of preoperative nutritional support for TJA patients and its potential benefits for minimizing perioperative blood transfusions require additional research.

Conclusion

In patients TJA, low preoperative serum albumin levels have been linked to a higher risk of perioperative blood transfusion. Moreover, a clear dose‒response relationship exists: the lower the preoperative serum albumin level, the higher the risk of requiring a blood transfusion during surgery. These findings suggest that preoperative serum albumin levels could serve as a predictive factor for blood transfusion risk in TJA patients.