Preoperative Albumin, Transferrin, and Total Lymphocyte Count as Risk Markers for Postoperative Complications After Total Joint Arthroplasty: A Systematic Review

Introduction: The purpose of this systematic review is to identify whether poor nutrition, as defined by the more commonly used markers of low albumin, low transferrin, or low total lymphocyte count (TLC), leads to more postoperative complications. We hypothesized that it may be possible to identify the levels of these laboratory values at which point total joint arthroplasty (TJA) may be ill advised. To this end, we analyzed the available literature regarding links between these three variables on postoperative complications after TJA. Methods: This systematic review was done in two parts: (1) In the first part, we reviewed the most commonly used malnutrition marker, albumin. (2) In the second part, we reviewed TLC and transferrin. We accessed PubMed, EMBASE, and Cochrane Library using relevant keywords to this study. The biostatistics were visualized using a random-effects forest plot. We compared data from all articles with sufficient data on patients with complications (ie, cases) and patients without complications (ie, noncases) among the two groups, malnourished and normal nutrition, from albumin, transferrin, and TLC data. Results: A meta-analysis of seven large-scale articles detailing the complications of albumin led to an all-cause relative risk increase of 1.93 when operating with hypoalbuminemia. This means that in the studies detailed enough to incorporate in this pooled analysis, operating on elective TJAs with low albumin is associated with a 93% increase in all measured complications. In the largest studies, analysis of transferrin levels for the most common complications revealed a relative risk increase of 2.52 when operating on patients with low transferrin levels. There were not enough subjects to do a biostatistical analysis in articles using TLC as the definition of malnutrition. Conclusion: The focus is on the trends rather than absolutes. As shown in Table 1, whether the albumin cutoff for albumin was 3.0 g/dL, 3.5 g/dL, or 3.9 g/dL, the trend remains the same. Because low albumin before TJAs tends to increase complications, it is recommended to incorporate albumin levels in preoperative workups. Many patients with hip and knee arthritis undergo months of conservative management (eg, physical therapy and corticosteroid injections) before considering surgery, and it would be wise to optimize their nutritional status in this period to minimize the risk of perioperative complications. The physician should use these data to provide informed consent of the increased risk to patients planning to undergo TJAs with elevated malnutrition markers. Because this research is retrospective in nature, albumin should be studied prospectively in hypoalbuminemic and normoalbuminemic patients and their postoperative outcomes should be measured. Regarding transferrin and TLC, future research should help elucidate their predictive value and determine the value of preoperatively optimizing them and their effect in mitigating postoperative complications.


Conclusion:
The focus is on the trends rather than absolutes. As shown in Table 1, whether the albumin cutoff for albumin was 3.0 g/dL, 3.5 g/dL, or 3.9 g/dL, the trend remains the same. Because low albumin before TJAs tends to increase complications, it is recommended to incorporate albumin levels in preoperative workups. Many patients with hip and knee arthritis undergo months of conservative management (eg, physical therapy and corticosteroid injections) before considering surgery, and it would be wise to optimize their nutritional status in this period to minimize the risk of perioperative complications. The physician should use these data to provide informed consent of the increased risk to patients planning to undergo TJAs with elevated malnutrition markers. Because this research is retrospective in nature, albumin should be studied prospectively in hypoalbuminemic and normoalbuminemic patients and their postoperative outcomes should be measured. Regarding transferrin and TLC, future research should help elucidate their predictive value and determine the value of preoperatively optimizing them and their effect in mitigating postoperative complications. P oor nutritional status before total joint arthroplasty (TJA) can lead to perioperative complications such as wound complications, 1 infection, [2][3][4] or even mortality. 5,6 Despite general agreement with this statement, the primary metric by which malnutrition is demonstrated is less clear. Several laboratory markers have been suggested as indicative of poor nutritional status. Among these, the more commonly used markers are albumin, 7 transferrin, 8 and total lymphocyte count (TLC). 1,9 The purpose of this systematic review is to identify whether poor nutrition, as defined by low albumin, low transferrin, or low TLC, leads to more postoperative complications. We hypothesized that it may be possible to identify the levels of these laboratory values at which point TJA may be ill advised. To this end, we analyzed the available literature regarding links between these three variables on postoperative complications after TJA.

Methods
This systematic review was done in two parts: (1) In the first part, we reviewed the most commonly used malnutrition marker, albumin. (2) In the second part, we reviewed TLC and transferrin. During the first part, we accessed PubMed, EMBASE, and Cochrane Library using search phrases with the following keywords: "albumin," "prealbumin," "prealbumin," "total joint arthroplasty," "total joint replacement," "total hip arthroplasty," "total hip replacement," "total knee arthroplasty," "total knee replacement," "infection," "complication," "read-mission," "readmit, "postoperative," and "perioperative." This yielded 312 results in PubMed, 15 results in EMBASE, and 88 results in Cochrane Library. These titles were reviewed by independent reviewers. Pertinent English-language articles were considered for inclusion in the final review, as were all duplicates. Screening of the initial titles yielded 23 PubMed articles, five EMBASE articles, and one Cochrane Library article. Of these 29 titles, three were duplicates. One additional article was identified during the full-text article retrieval and was included in the systematic review, yielding a total of 30 articles related to albumin. We reviewed the following items in the 29 obtainable articles: year of publication, primary outcome, albumin cutoff level used to define hypoalbuminemia, patient source, sample size, follow-up    The incidence of complications was higher in malnourished patients than in nonmalnourished patients, regardless of whether they were obese. Renal complications were the most common complication experienced by malnourished patients and occurred at markedly higher rates than for nonmalnourished patients. Age was not a notable factor in developing malnutrition, but the incidence increased steadily through age 70. Kamath et al 15 Unplanned postoperative intensive care unit admission Patients with low albumin had a higher risk of unplanned postoperative intensive care unit admission. Kamath et al 6 Mortality, superficial wound infection, and other Patients in the low serum albumin group were statistically more likely to develop deep SSI, organ space SSI, and other complications. Kim et al 16 Incidence of acute kidney injury, hospital stay, and overall mortality Low albumin within two postoperative days was an independent risk factor for acute kidney injury and increased length of hospital stay in patients undergoing TKA. Lavernia et al 17 Complications, resource consumption, length of stay, and number of in-hospital medical or surgical consults obtained Patients with low albumin levels had higher charges, higher severity of illness, and longer length of stay. Marín et al 9 Delayed wound healing Preoperative lymphocyte count of less than 1500 cells/mm 3 was associated with a three times greater frequency of healing complications, whereas preoperative serum albumin and transferrin levels had no notable predictive value. Mednick et al 18 Readmission The risk of readmission after THA increased with growing preoperative comorbidity burden. It specifically increased in patients with a body mass index of greater than or equal to 40 kg/m 2 , a history of corticosteroid use, and low preoperative serum albumin and in patients with postoperative SSI, a thromboembolic event, and sepsis. Morey et al 19 Wound complications (ie, drainage, hemarthrosis, skin necrosis, and dehiscence) or PJI Findings called into question the values of serum albumin level and TLC as a surrogate of malnutrition for predicting wound complications after TKA. Nelson et al 4 Mortality, superficial wound infection, and other Morbid obesity was not independently associated with most perioperative complications measured by the ACS-NSQIP and was associated only with increases in progressive renal insufficiency, superficial SSI, and sepsis among the 21 perioperative variables measured. Low serum albumin was associated with increased mortality and multiple additional major perioperative complications after TKA. Low serum albumin, more so than morbid obesity, was associated with major perioperative complications.
ACS-NSQIP = the American College of Surgeons National Surgical Quality Improvement Program; OR = odds ratio; PJI = periprosthetic joint infection; SSI = surgical site infection; THA = total hip arthroplasty; TLC = total lymphocyte count. Table 2 is a continuation of Table 1 and lists the outcomes measured in and conclusions derived from each of the studies.
duration, study design, complications evaluated, and relative risk (RR) of complications among patients with hypoalbuminemia. The resulting studies are listed in Tables 1-4 and Table 9.
We used a similar approach in the second part of the study, substituting the keywords "transferrin" and "total lymphocyte count." There were 125 relevant article titles in PubMed, 108 relevant article titles in EMBASE, and three in Cochrane Library, totaling 236 relevant titles based on the question of interest. The same independent reviewers reviewed the titles, which yielded 66 relevant article ab-stracts. Of these 66, 54 full-text articles were selected after abstract evaluation. Including duplicates, 29 unique fulltext articles were chosen for the final systematic review. On retrieval of the articles, we reviewed the following variables: primary outcome, secondary outcome, transferrin cutoff level used  24 Length of stay Preoperative serum albumin was the only preoperative serum test associated with length of stay. Albumin was inversely related with length of stay. Walls et al 25 Mortality, superficial incisional SSI, and other Hypoalbuminemia was a notable risk factor for mortality and major morbidity among THA patients, whereas morbid obesity was only associated with an increased risk of superficial SSI. Yi et al 26 Septic revision and aseptic revision The presence of one or more laboratory parameters suggestive of malnutrition (ie, low albumin, low transferrin, and low lymphocyte count), although common in both normal weight and overweight patients, was independently associated with both chronic PJI and the development of an acute postoperative infection after an aseptic revision arthroplasty.
ACS-NSQIP = the American College of Surgeons National Surgical Quality Improvement Program; OR = odds ratio; PJI = periprosthetic joint infection; SSI = surgical site infection; THA = total hip arthroplasty; TLC = total lymphocyte count. Table 2 is a continuation of Table 1 and lists the outcomes measured in and conclusions derived from each of the studies.
Preoperative Albumin, Transferrin, and TLC to define malnutrition, albumin cutoff level used to define hypoalbuminemia, TLC cutoff level used to define malnutrition, patient source, sample size, study design, and the number of cases and noncases among patients classified as normal nutrition and malnutrition. The resulting studies are listed in Tables 5-9.
The biostatistics were visualized using a random-effects forest plot. We compared data from all studies with sufficient data on patients with complications (ie, cases) and patients without complications (ie, noncases) among the two groups, normal albumin and hypoalbuminemia, to calculate a pooled RR combining the number of patients from all studies. A random-effects forest plot was also used to visualize information from the two studies with sufficient transferrin data.

Results
When determining which studies had adequate detail regarding the numbers  Table 3, which is the addition of all patients from Table 2. c The complication is the number of patients having a PJI after revision for an aseptic indication. Amount of albumin w/aseptic revisions = 3802 in Figure 1, which separates low albumin from norm albumin. The amount of normal albumin and low albumin w/complications is written in the last paragraph of the results section and is graphed out in Figure 3 of Bohl et al. 10  Table 3 is a numerical representation of Figure 1. a Pooled RR using DerSimonian and Laird Random effects model.
of cases and noncases, we found that seven studies had sufficient data for inclusion in the pooled analysis ( Figure  1 and Tables 3 and 4). For each of the seven included studies, we chose only the complication of interest, "major complications" or "any complication," if presented. When malnutrition was defined as low transferrin, Roche et al 27  The study by Huang et al 14 is the only study that describes malnutrition as low albumin or transferrin in the second half of the study. It found an increased length of stay (1.7 days), renal complications (OR: 2.85), and any complication (OR: 2.42).
The study by Yi et al 26 is the only study that describes malnutrition as low albumin or transferrin or TLC in the second half of the study. It found an increased risk of chronic septic failure (OR: 2.13) and acute postoperative infection complicating an aseptic revision arthroplasty (OR: 5.9).
The study by Lavernia et al, 17 the only study that describes malnutrition as low TLC in the second half of the study, found increased risks with increased cost/charges (P-value: 0.004), increased anesthesia time (P-value: 0.02), increase in surgical time (P-value: 0.002), increased in-hospital consults (P-value: 0.004), and increased length of stay (0.3 days).
The study by Morey et al, 19 the only study that describes malnutrition as  Table 6 Outcomes and Conclusions of Transferrin Studies

Roche et al 27 Postoperative infection, wound complications, concomitant infection with wound complications, and infection after wound complications
Patients with lower values of nutritional markers had higher incidences, and hence, odds ratios of complications.
Huang et al 14 Multiple complications and length of stay Malnutrition had higher rates of postoperative complications and length of stay. Table 6 is a continuation of Table 5 and describes the outcomes and conclusions derived from each of the studies.  Table 7 is a numerical representation of Figure 2. The information from Figure 2 and Table 7 were derived from the data in Table 8.
Preoperative Albumin, Transferrin, and TLC low albumin or TLC in the second half of the study, found the complication with increased risk was a decrease in range of motion in function score according to the American Knee Society scale (P-value: 0.009, amount of decrease in degrees: 1.5). Meta-analysis of seven largescale studies detailing the complications of albumin led to an allcause RR increase of 1.93 when operating with hypoalbuminemia. This means that in the studies detailed enough to incorporate in this pooled analysis, operating on elective TJAs with low albumin is associated with a 93% increase in all measured complications. In the largest studies, analysis of transferrin levels for the most common complications revealed a RR increase of 2.52 when operating on patients with low transferrin levels. There were not enough subjects to do a biostatistical analysis in articles using TLC as the definition of malnutrition.

Discussion
Most of the articles use 3.5 g/dL as the cutoff for hypoalbuminemia, with the range varying from 3.0 to 3.9 g/dL (Table 1). In the studies isolating only albumin, hypoalbuminemia is associated with increased risk for postoperative complications. Based on this systematic review, sufficient evidence is not available to make a statement regarding the risk for postoperative complications in  patients with malnutrition as defined by low transferrin or low TLC. As shown in Table 9, one article isolated TLC and one article isolated transferrin as markers for malnutrition, with a maximum study cohort size of 3111. The other articles included albumin as a component of their definitions and did not separate out the patients with only low albumin, which makes it difficult to identify whether, in their patient cohorts, low transferrin alone or low TLC alone led to increased risks of complications. Although there is not enough conclusive evidence to state that transferrin or TLC levels alone warrant delaying an elective TJA, that does not mean low transferrin or TLC are not present when there is low albumin. The articles studying albumin along with TLC or transferrin do not describe the relationship between them and treat them only as separate, isolated cohorts. Therefore, we are not able to state that low TLC or transferrin levels tend to accompany low albumin levels. In addition, only one 19 of the 27 unique articles isolating albumin claims that albumin is not a reliable test of choice for identifying malnutrition. Despite the variability in methodologies, with certain studies selecting from the American College of Surgeons National Surgical Quality Improvement Program 3,5,10,12 and certain studies 26 being performed by a single surgeon on a team to limit variability, albumin seems to lead to an increased risk of postoperative complications, including mortality, unplanned readmissions, and increased length of stay.

Conclusion
The focus is on the trends rather than absolutes. As shown in Table 1, whether the albumin cutoff for Graph showing the forest plot for the role of malnutrition in albumin articles based on studies with enough subjects. The (1) and (2) next to Kamath were what was used to identify them in the data analysis and does not correlate with the references at the bottom. albumin was 3.0 g/dL, 3.5 g/dL, or 3.9 g/dL, the trend remains the same. Because low albumin before TJAs tends to increase complications, it is recommended to incorporate albumin levels in preoperative workups. Many patients with hip and knee arthritis undergo months of conservative management (eg, physical therapy and corticosteroid injections) before considering surgery, and it would be wise to optimize their nutritional status in this period to minimize the risk of perioperative complications. The physician should use these data to provide informed consent of the increased risk to patients planning to undergo TJAs with elevated malnutrition markers. Because this research is retrospective in nature, albumin should be studied prospectively in hypoalbuminemic and normoalbuminemic patients and their postoperative outcomes should be measured. Regarding transferrin and TLC, future research should help elucidate their predictive value and determine the value of preoperatively optimizing them and their effect in mitigating postoperative complications.  Graph showing the forest plot of transferrin articles.