Examining Albumin as a Bioindicator of Healing Capability in Patients With Diabetic Foot Ulcers: A Retrospective Review

DFUs are a significant source of morbidity in patients with diabetes and are a major cause of leg amputations in the developed world.^1-5 Among patients who undergo lower extremity amputation owing to complications of DFU, the 5-year mortality rate is 46.2% for minor amputation and 56.6% for major amputation.⁶ It is estimated that only half of DFUs heal within 1 year in the United States,⁷ and amputation often is necessary to prevent life-threatening infection. Prediction of DFU outcomes helps optimize management strategies. Several biomarkers, including albumin, Hb, HbA_1C, C-reactive protein level, erythrocyte sedimentation rate, and white blood cell count, have been shown to correlate with long-term outcomes in patients with DFU, but there is a paucity of literature examining the short-term predictive value of trends in these markers.^2,8-14 

The current study aims to add increased temporal acuity to the connection between albumin level and ulcer status to help clinicians better estimate treatment success over the course of care. The authors hypothesized that lower albumin levels at presentation would correlate with worse outcomes at 12 weeks of treatment. It was also hypothesized that negative albumin level trends would correlate with worse healing (defined by increasing ulcer size) compared with stable or improving albumin level trends. 

A retrospective review was conducted using as inclusion criteria ICD-10 codes of patients with plantar DFU who presented to the Atrium Health Wake Forest Baptist Wound Care and Hyperbaric Center between 2010 and 2022. Patients with end-stage heart, kidney, or lung disease; osteomyelitis; chronic steroid use; uncontrolled diabetes (HbA_1C >12%); tobacco use; peripheral artery disease; or venous insufficiency were excluded. Traumatic or surgical wounds were also excluded, as were charts without at least 2 albumin level measurements and wound care visits over 12 weeks.

Wound size, interventions, and laboratory data over 2 to 4 monthly visits were collected. Average initial albumin levels among healed and unhealed patients were compared using 2-tailed t tests assuming unequal variance. The chi-square test was used to compare proportions, and linear regression was used to examine the relationship between change in albumin level and wound size. P values less than .05 were considered statistically significant. 

A total of 793 charts were reviewed, and 62 patients met all inclusion criteria. The median time from wound onset to presentation to the wound care center was 43 days, with an average initial wound area of 5.36 cm² (range, 0.03 cm²–79.7 cm²). Of 138 albumin measurements, 9 were in the hypoalbuminemia range (<3.0 g/dL) but were not less than 2.2 g/dL. Of the 62 patients included in the study, 7 had hypoalbuminemia at 1 or more visits. The average albumin level at presentation was 3.7 g/dL. 

After 12 weeks of treatment, 24 patients’ ulcers were healed. Initial ulcer size was significantly smaller in the healed group than in the nonhealed group (1.08 cm² and 7.83 cm², respectively; P < .05), but there was no significant difference in initial albumin level between healed and nonhealed patients (3.75 g/dL and 3.65 g/dL; P = .49). Time to initial presentation was also similar between groups (P = .35).

Results of the Pearson correlation coefficient indicated that there is a nonsignificant, very small negative relationship between monthly changes in albumin level and monthly changes in wound size (r = 0.144, P = .240) (Figure 1). This was also true when percentage change in wound size was used as the dependent variable. More generally, patients with positive changes in albumin value were not more likely to have decreased wound sizes compared with patients with negative or unchanged albumin values (P = .11) (Figure 2). Similarly, patients with large negative trends in albumin level (<0.5 g/dL per month) were no more likely than patients with stable or positive trends in albumin level to have increased wound sizes (4 of 14 and 7 of 36, respectively; P = .48). In other words, a positive or negative change in albumin level within the normal range did not predict a change in wound size. 

In patients with at least 1 hypoalbuminemia value (<3.0 g/dL), healed proportion and average final ulcer size were not significantly different compared with that of patients with normal albumin levels (healed proportion: 2 of 7 and 22 of 55, respectively [P = .56]; final ulcer size: 3.8 cm² and 6.0 cm², respectively [P = .47]). Wound size changes in patients who trended from hypoalbuminemia to normoalbuminemia were not statistically different from those in patients who trended towards hypoalbuminemia or from those in patients who maintained normoalbuminemia. Patients who trended from normoalbuminemia to hypoalbuminemia had significantly increased wound sizes compared with stable, normoalbuminemic patients (0.04 cm² and −1.17 cm², respectively; P < .05) (Figure 3). A mild inverse correlation (18%) was present between albumin level change and ulcer area change among patients who trended into or out of hypoalbuminemia. 

The association between hypoalbuminemia and poor DFU outcome is well established.^2,8–14 Malnourishment impairs healing by increasing infection risk and interrupting processes such as coagulation and extracellular remodeling.¹⁵ However, the utility of albumin level as a nutritional status metric is controversial. A systematic review of 63 articles concluded that albumin and prealbumin levels remain stable until the extremes of starvation (ie, body mass index <12 or >6 weeks of starvation).¹⁶ Rather, the presence of hypoalbuminemia in malnourished and ill patients is more likely the result of increased inflammation and vascular permeability.¹⁷ The association between hypoalbuminemia and poor DFU outcome may therefore be explained by the common-cause confounder of chronic inflammation, which negatively affects wound healing.¹⁸

While several bioindicators are associated with adverse DFU outcomes in the long-term, few studies have examined how well short-term changes in these variables predict wound healing. In 2011, Christman et al¹⁹ published the results of such a study, showing that baseline HbA_1C significantly correlates with wound healing rate. That study explored several other variables, such as race and ethnicity, tobacco use, body mass index, white blood cell count, and neuropathy, none of which was significantly associated with wound healing rate. Albumin was not examined.

The authors of the present study aimed to determine whether albumin level is similarly predictive of DFU wound healing, hypothesizing that lower albumin levels at presentation would correlate with worse outcomes at 12 weeks of treatment and that negative albumin trends would correlate with impaired healing. No connection was established between DFU outcomes and albumin status. The most likely explanation is that, while chronic inflammation and subsequent hypoalbuminemia predict broad outcomes such as amputation, the inflammatory changes measured with monthly albumin values are too minor to associate outcome differences at the authors’ timescale. It is also probable that the maximum treatment course of 12 weeks was too brief to accurately gauge DFU outcomes; less than 12 weeks of data were available for many patients in the present study. However, a significant correlation was established between normoalbuminemic to hypoalbuminemic trends and worse healing (Figure 3); thus, the second half of the hypothesis was accepted. This finding aligns with what is currently known about albumin, but it contributes to the literature by showing that the predictive value of hypoalbuminemia exists on a much shorter timescale than is currently established. This observation should be interpreted in the context of a small patient sample (3 patients trended from normoalbuminemia to hypoalbuminemia) and the absence of inverse significance in hypoalbuminemic to normoalbuminemic patients.

Several studies have found significant differences within the normoalbuminemia range when comparing positive and negative long-term outcomes.^5,6,9 In the present study, short-term changes within the normoalbuminemia range had no correlation with ulcer healing (Figures 1, 2). Albumin’s prognostic value thus remains limited to abnormally low values, which, as mentioned previously, aligns with how the laboratory value is currently used clinically. To the knowledge of the authors of the present study, this is the first study to examine whether short-term albumin trends are predictive of DFU wound healing.

Accounting for potential confounders is difficult in retrospective studies and is especially challenging when exploring complex processes such as wound healing. The present study excluded major confounding comorbidities such as vascular disease, tobacco use, end-stage organ disease, and infection; however, untracked factors such as at-home resources, out-of-office wound care, stress, other drug use, and metabolic status all influence the results. Medical variables such as anemia, kidney function, and HbA_1C likely also influence the results. A probable cause for exclusion was the ICD-10 codes used. Instead of setting strict inclusion criteria, future studies may instead use multivariate analysis to determine which of many potential variables best predict short-term ulcer healing. This approach would also help power the study by excluding fewer patients and would have improved generalizability, since most patients with DFUs have comorbidities. 

Because the correlation between monthly albumin change and monthly wound area change was small, a post hoc power calculation was performed to determine whether the present study was adequately powered to detect a significant correlation. This study had 80% power to detect a correlation of 35% or greater at an alpha level of .05, which indicates that the correlation of 0.144 could not have been significant at the sample size. The primary conclusion that monthly albumin changes correlate minimally with monthly wound area changes is unaffected by this finding.

Albumin is commonly measured in patients with DFU to gauge metabolic status and wound healing capacity. The data from the present study suggest that monthly changes from normoalbuminemia to hypoalbuminemia are predictive of worse wound healing; however, this conclusion is limited by a small sample size. Short-term changes within normal albumin ranges have little predictive value concerning wound healing. Thus, the utility of albumin as a bioindicator of healing capability remains limited to below-normal values. The ideal laboratory marker for malnutrition remains elusive and requires further investigation.