Perioperative Evaluation of Nutritional Status to Predict Complications in Patients with Major Lower Extremity Amputation

Beyond their use in the management of traumatic injuries, MLEAs are highly morbid procedures performed for the treatment of dysvascular conditions such as end-stage critical limb ischemia and diabetic foot ulcer.^1,2 In individuals undergoing MLEA for these conditions, between 13% and 30% experience postoperative complications, and the associated 1-year mortality rate is approximately 30%.^3-5 Identification of factors
associated with such postoperative complications is imperative to minimizing morbidity.

Perioperative malnutrition is a well-known prognostic factor for surgical outcomes; Alb is a visceral protein that is often used as a surrogate for nutritional status and postoperative complications.^6-13 The European Society for Clinical Nutrition and Metabolism and the American Society for Parenteral and Enteral Nutrition define chronic disease-related malnutrition and acute disease or injury-related malnutrition as malnutrition caused by acute or chronic inflammation.^14,15 However, objectively measuring malnutrition in a chronically inflamed population has proven difficult.¹⁶ Low Alb is not specific for malnutrition because of its confounding role as a negative acute phase reactant.⁹ Further, Alb has a long half-life, between 2 to 3 weeks; thus, deficiencies may be more indicative of chronic nutrition than perioperative status.¹⁷ In contrast, pAlb has a short biological half-life of about 2 days, which suggests that it may be a more accurate biomarker for acute changes in nutritional status.^6-13,18-20

Investigation of nutrition within the chronic limb-threatening ischemia and diabetic foot ulcer populations reveals that preoperative nutritional index scores and Alb are independent predictors of morbidity and mortality.^21,22 However, the investigation of hypoalbuminemia in patients undergoing lower extremity amputation is much more limited.⁷ To the knowledge of the authors of the present study, there are no published studies on the predictive role of perioperative pAlb and Alb on postoperative complications in this patient population. Therefore, the principal aim of the present study was to assess Alb and pAlb levels and explore their relationship to postoperative complications in patients undergoing MLEA for chronic wounds.

Data Source and Patient Population

A retrospective review of data from patients who underwent MLEA at a tertiary wound care center from January 2017 to October 2021 was performed. The 3 documented criteria necessary for inclusion were MLEA, defined as a BKA or knee disarticulation; Alb level recorded within 30 days or pAlb level recorded within 5 days of MLEA; and follow-up within 30 days after surgery. This study was approved by the institutional review board (IRB No. 2018-173).

Perioperative Management

Patients were clinically evaluated perioperatively for medical comorbidities, nonhealing wounds, vascular disease, and the need for preoperative debridement and revascularization. Patients on outpatient anticoagulation regimens were transitioned to heparin when necessary for the maintenance of anticoagulation. Following their initial operation, patients were readmitted to the inpatient floor for additional medical and surgical management. This included multidisciplinary management of antibiotic administration, nutritional supplementation, and serial surgical debridement as necessary. Nutritional status was managed per the recommendations of the consulted dietician.

Pre- and post-debridement qualitative wound cultures were obtained at every operating room visit. Antibiotic administration was determined based on wound culture sensitivity. Per the judgment of each surgeon, the resolution of any preoperative infections was ensured via laboratory markers, qualitative wound cultures, and clinical assessments. If started, heparin was continued until the final closure of amputation, at which point outpatient medications were restarted, generally 24 to 48 hours postoperatively. After discharge, follow-up visits took place in the outpatient wound center clinic at weekly, biweekly, and monthly intervals as necessary per site healing and surgeon preference.

Outcome Variables and Study Groups

Electronic medical records were reviewed to collect preoperative and postoperative variables of interest. Preoperative data included patient demographics, comorbidities, and perioperative laboratory values. The primary outcomes of interest were surgical hematoma and infection following MLEA. Hematoma was defined as blood collection at the surgical site requiring intervention. Infections were determined based on purulent drainage, positive postoperative cultures, erythema around surgical incisions, or physician diagnosis. Secondary outcomes included dehiscence, reamputation, and the occurrence of postoperative complications. Complications that occurred within the first 30 days postoperatively were included for all patients. Patients were initially stratified according to preoperative nutritional markers, using standard clinically accepted thresholds of Alb less than 3.2 g/dL and pAlb less than 10 mg/dL for poor nutritional status.^16,23 Additional outcomes analysis at serially decreasing thresholds for preoperative Alb (in increments of 0.1 g/dL) and pAlb (in increments of 1 mg/dL) was also performed to identify thresholds that would yield significance for the major outcomes of interest. Preoperative demographics, perioperative laboratory values, and 30-day follow-up data were analyzed to determine independent predictors for the primary and secondary outcomes.

Statistical Analysis

Statistical analysis was performed using Stata/BE Software (Version 17, StataCorp LLC) and Prism (version 8.0.0 for Mac OS X; GraphPad Software). The Shapiro-Wilk test was performed to assess the normality of continuous variables. Descriptive statistics were calculated for all patient data, including basic and advanced parameters. All variables were compared between low and normal Alb and pAlb groups using bivariate analysis, using the Mann-Whitney test and unpaired 2-tailed t test as appropriate for continuous variables and the chi-square or Fisher exact test for categorical variables.

Univariate logistic regression analysis was used to calculate unadjusted ORs for hematoma and infection at the defined thresholds for pAlb (10 mg/dL) and Alb (3.2 g/dL). Variables that reached statistical significance in univariate analysis were included in multivariable regression analysis. Multivariable logistic regression was used to evaluate the independent effects of pAlb and Alb on postoperative complications while controlling for confounding factors in the model. Statistical significance was defined as P<.05.

Patient Demographics

A total of 337 patients who underwent lower leg amputation were identified. The study included 303 patients with Alb data, 76% of whom had low perioperative Alb levels (<3.2 g/dL). Of the 206 patients with pAlb data, 40% had low perioperative pAlb levels (<10 mg/dL).  The mean age of all 303 patients was 59.9 years ± 13.7 standard deviation. Demographic data for low and normal Alb and pAlb groups are presented in Table 1.

Patients with Alb less than 3.2 g/dL were significantly older than patients with normal Alb levels (61.3 years ± 12.8 and 55.4 years ± 15.4, respectively; P =.032). A higher proportion of patients in the low Alb group had diabetes (81.0% vs. 65.3%; P =.001), ESRD (37.2% vs. 21.8%; P =.01), and lower mean hemoglobin (8.6 g/dL vs. 9.9 g/dL; P <.001) compared with patients with normal Alb levels. Patients with pAlb levels less than 10 mg/dL were older than patients with normal pAlb levels (63.1 years ± 14.2 vs. 57.4 years ± 14.0; P =.042) and had higher CRP levels (75.4 mg/L ± 52.7 vs. 41.5 mg/L ± 35.7; P =.002), but they were less likely to have ESRD (21.7% vs. 39.0%; P =.01). No other significant differences were found between these groups. A separate analysis of low pAlb (<9 mg/dL) and normal pAlb (≥9 mg/dL) found that patients with pAlb less than 9 mg/dL were significantly older (64.8 years ± 13.6 vs. 57.2 years ± 14.1; P =.003) and had higher CRP levels (77.8 mg/L ± 57.4 vs. 44.2 mg/L ± 35.6; P =.012).

Prealbumin and Surgical Complications

Low pAlb at the threshold of less than 10 mg/dL was associated with a significantly higher rate of infection (12.1% vs. 3.3%; P =.014) and nonsignificant increases in hematoma rate (4.8% vs. 0.8%; P =.067) compared with normal pAlb (Table 2). For low pAlb at the threshold less than 9 mg/dL, both the hematoma rate (6.0% vs. 0.7%; P =.021) and infection rate (12.0% vs. 4.3%; P =.039) increased significantly
compared with pAlb greater than or equal to 9 mg/dL (Table 2). On bivariate regression analysis, low pAlb was found to be associated with increased odds of hematoma at the thresholds of less than 10 mg/dL (OR, 5.67; 95% CI, 1.02-105.90) and less than 9 mg/dL (OR, 7.07; 95% CI, 1.28-132.10) and increased odds of infection at those respective levels (OR, 3.72; 95% CI, 1.36-13.04 and OR, 2.73; 95% CI, 1.11-7.72, respectively). On multivariate analysis, pAlb was the only independent predictive factor of hematoma when controlling for age and CRP level (P =.021). However, neither age (P =.848), CRP level (P =.687), nor pAlb level (P =.168) was found to be an independent predictor of infection. No significant differences were found between groups with low or normal pAlb at either the 10 mg/dL threshold or the 9 mg/dL threshold (<10 mg/dL vs. ≥10 mg/dL; <9 mg/dL vs. ≥9 mg/dL) for dehiscence rate (7.2% vs. 10.6%; 9.0% vs. 9.3%), reamputation rate (3.6% vs. 0.8%; 3.0% vs. 1.4%), or overall complications rate (19.3% vs. 13.8%; 21.0% vs. 13.7%) (Table 2).

Albumin

No significant differences were found between the low Alb and normal Alb groups (<3.2 g/dL vs. ≥3.2 g/dL) in terms of hematoma rate (3.0% vs. 2.8%; P =.912) or infection rate (7.8% vs. 8.3%; P =.882) (Table 2). Bivariate regression analysis found no association between low Alb and hematoma (OR, 1.09; 95% CI, 0.26-7.45), infection (OR, 0.93; 95% CI, 0.37-2.65), or dehiscence (OR, 1.10; 95% CI, 0.45-3.10). No significant differences were found between the low Alb and normal Alb groups (<3.2 g/dL vs. ≥3.2 g/dL) for dehiscence (9.1% vs. 8.3%), reamputation (2.6% vs. 2.8%), and overall complications (16.9% vs. 18.1%) (Table 2).

This study demonstrates an increased propensity towards postoperative infection in patients with perioperative malnutrition when defined as a pAlb level less than 10 mg/dL. By contrast, low Alb (<3.2 g/dL) did not significantly predict postoperative complications (infection, hematoma, or dehiscence). Analysis at serially decreasing minimum thresholds revealed that pAlb less than 9 mg/dL was independently associated with increased rates of hematoma postoperatively. This suggests that acute malnutrition, indicated by perioperative pAlb less than 10 mg/dL or less than 9 mg/dL, can predict short-term complications in patients
who have undergone MLEA.^16,24 The results of the present study suggest that perioperative pAlb may have more utility than Alb in predicting MLEA-specific complications in the immediate postoperative period.^16,24

Preoperative malnutrition is associated with adverse clinical outcomes in patients with dysvascular disease.²⁵ Individuals who undergo MLEA for acute complications of chronic disease fulfill criteria for malnutrition according to the Nutritional Risk Screening and the Malnutrition Universal Screening Tool for adults, which may require intervention depending on the patient’s age, weight, decompensation, perioperative diet, and length of hospital stay.²⁶ Further, screening with the Mini Nutritional Assessment for persons age 65 and older would suggest that all such patients, regardless of disease or nutritional status, would possibly be at risk of malnutrition after MLEA owing to their mobility status and recent physical stress.²⁶

The Center for Wound Healing and Hyperbaric Medicine at MedStar Georgetown University Hospital uses a multidisciplinary approach aimed at optimizing aspects of care via collaboration between specialists, including plastic surgeons, registered dieticians, vascular surgeons, podiatrists, infectious diseases physicians, rheumatologists, orthotists/prosthetists, and peripheral nerve surgeons.²⁷ Postoperatively, these team members ensure enhanced blood flow, infection control, and augmented biomechanical function to provide the ideal setting for recovery after MLEA. Preoperatively, a standardized nutritional support algorithm is used to initiate and individualize the nutritional consult before amputation. Based on recommendations by a registered dietician, patients are routinely placed on a high-protein diet and high-protein oral nutritional supplements. Goal protein intake is guided by wound status and signs of protein malnutrition, such as muscle wasting, typically ranging from 1 to 1.5 g/kg daily.^28,29

The integration of nutritional assessment and intervention is particularly relevant for patients undergoing MLEA for nontraumatic indications owing to the increased nutritional demand associated with chronic and acute disease and surgical procedures and the projected implications for mobility.²⁵ Excessive inflammation resulting from injury and disease induces a state of increased resting energy expenditure (hypermetabolic state), anorexia, poor absorption, and negative nitrogen balance (hypercatabolic state).^24,30,31 However, preoperative evaluation and correction of these increased nutritional demands is limited in acute presentations.²⁴ Therefore, it is of paramount importance to use algorithmic approaches to the perioperative diagnosis and management of these pervasive nutritional deficits.²⁵ Incorporating the use of biomarkers into nutritional support algorithms may increase the precision of perioperative supplementation.^16,29,32-35 In the present study, using the clinically defined thresholds of 3.2 g/dL for Alb and 10 mg/dL for pAlb, 76.2% of the patients studied would be considered chronically malnourished and 40.29% would be considered acutely malnourished. Because nutritional optimization has a greater effect on pAlb than Alb, this discrepancy may reflect the effect of perioperative nutritional optimization.²⁹

The findings of the present study add to the literature surrounding the perioperative nutritional implications of biomarkers by emphasizing the potential role of pAlb in the nutritional evaluation of patients following MLEA for nontraumatic indications. In surgical patients, Alb is frequently used as a biomarker for nutritional status and predicts long-term complications, including postoperative morbidity and mortality, in a dose-response manner, although it inconsistently predicts short-term complications.^{3,7,10,13,24,32,36-41} The causal relationship between hypoalbuminemia and adverse surgical outcomes is thought to be largely a consequence of chronic inflammation.^16,18,24 Hypoalbuminemia further impedes wound healing by precipitating reactive oxygen species, interstitial edema, and tissue catabolism.²⁴ At low levels, Alb may predispose to complications by its failure to regulate inflammation, osmotic pressure, acid-base status, oxidation, and coagulation (Figure).^10,38,42,43  Therefore, it stands to reason that a person with chronic disease or chronic low nutritional intake would also have low Alb and subsequently have more adverse postoperative complications compared with a person with no comorbidities or with sufficient nutritional intake. The present study focused primarily on the propensity for short-term complications after MLEA. For such dependent variables, no statistically significant associations between Alb level and postoperative complications were observed. This may be because of the relative homogeneity of this population and the stability of Alb levels in the immediate setting of acute inflammation.

In the immediate perioperative setting of MLEA for severe wounds, ischemia, or infection, nutritional demands often abruptly increase as the result of extensive surgery, worsening disease status, or dietary restriction. At baseline, a majority of patients in the present study had diabetes or peripheral arterial disease, and many had ESRD, all of which would likely predict chronically lower Alb. The added nutritional demands of an acute illness, surgery, immobility, and prolonged hospital stay increase the overall nutritional demand. Still, fluctuations in energy requirements may not be captured by changes in Alb level during hospitalization. Further, it has not yet been determined if nutritional support can correct hypoalbuminemia.^16,18,28,44 Thus, Alb may not be the ideal biomarker for evaluating abrupt changes in nutritional demand in the perioperative setting.^{3,5,8,18,22,29,32,39,44-47}

In contrast, pAlb may be more helpful in evaluating the nutritional needs of patients undergoing MLEA as a result of complications from chronic diseases. pAlb has a much shorter half-life than albumin (~2 days vs. 2-3 weeks, respectively) in the range of CRP (t1⁄2 ~19 hours), which may more precisely reflect fluctuations in a patient’s perioperative nutritional demands.⁴⁸ Unlike CRP, pAlb is synthesized in greater amounts, is inversely related to inflammation, and has been demonstrated to predict recovery from malnutrition and overall disease.^{16,24,30,43,44} Further, a low pAlb level has been found to be predictive of short-term surgical complications such as surgical site infection, poor wound healing, and free flap compromise.^{16,38,46,49-51} The results of the present study corroborate the findings of these previous studies and suggest that a low pAlb level may be associated with postoperative infection and hematoma following nontraumatic MLEA.

Malnutrition is thought to impair wound healing in part owing to immune system compromise.^{2,10,18,24,31,42,52-54} Mechanistically, this may be related in part to a low pAlb level. The primary physiologic function of pAlb is to aid in transporting both thyroxine and vitamin A.^{4,19,43,55-57} Thus, a sudden drop in pAlb levels can cause hyperthyroidism (catabolism and hypermetabolism) and disruption of retinol transport (impaired lymphocyte maturation and weakened immune responses).^41,43 It seems as though a low pAlb level may influence the prolonged infections, inflammation, and poor wound healing observed in patients with acute and chronic inflammation.^41,43 Although the topic is beyond the scope of the present study, pAlb may also be useful in perioperative monitoring because pAlb has been observed to rise with enteral nutritional intervention and recovery from infection.^24,30,58-63

Although biomarkers cannot reliably diagnose malnutrition, Alb and pAlb provide additional information that can complement the clinical nutritional assessment of individuals experiencing chronic or acute inflammation. Patients who require nontraumatic MLEA often have compromised nutritional reserves and may require increased nutrition following amputation owing to diet, resting energy expenditure, and decreased mobility. The results of the present study suggest that clinical application of pAlb may deepen the perioperative nutritional assessment of this high-risk population and encourage further investigation of its utility in experimental evaluations of nutritional supplementation.^6,50,52,64

The conclusions of the present study are inherently limited by the retrospective design and the consistency of medical documentation. In addition, the independent variables assessed in this study and corroborated for nutritional evaluation are greatly affected by individual and dynamic patient characteristics related to surgical procedure, perioperative medication use, disease profiles, and dietary composition.²⁵ Thus, assessing the need for intervention requires an understanding of the competing external influences on these biomarkers. While the authors of the present study controlled for the covariates with significance on univariate analysis, this single-institution study has limited generalizability, and several factors require further investigation. In terms of procedural influences, AKAs are calorically more demanding than BKAs and may require more rigorous evaluation and treatment. Regarding medical management, medications such as warfarin are often manipulated in the perioperative setting, and the reciprocal interactions with Alb and pAlb may predispose patients to complications.^65,66 These areas of future work are expected to refine the relationship between Alb and pAlb and outcomes according to multidisciplinary care with integrated nutritional assessment, perioperative pharmacologic regimen, CRP/pAlb ratio, amputation level (AKA vs. BKA), and perioperative nutrition (ie, diet or nutritional supplementation).^11,67,68

The findings of the present study suggest that pAlb may be a useful biomarker in the armamentarium for the perioperative nutritional evaluation of patients undergoing MLEA for the management of chronic wounds. Specifically, it was observed that a low perioperative pAlb level was associated with postoperative hematoma (pAlb <10 mg/dL) and infection (pAlb <9 mg/dL). By contrast, a low perioperative Alb level (<3.2 g/dL) was not predictive of postoperative complications. This remained true for serially decreasing levels of Alb. These findings expand the literature and indicate areas for further research on the interventional responses to increased perioperative nutritional demands. While there are no current guidelines for the specific nutritional management of similar patients, surgeons should continue working closely with dieticians to optimize these patients’ nutrition during and after the hospital stay.