Early-admission serum albumin and hospitalization burden in a selected serial-culture cohort of patients with diabetic foot ulcers and subsequent Pseudomonas aeruginosa detection: a retrospective cohort study.
Kan F, Wang A, Tan Y, Dai H • Frontiers in endocrinology • 2026
Within a selected serial-culture cohort of DFU patients with subsequent Pseudomonas aeruginosa detection, lower early-admission serum albumin was associated with greater hospitalization burden, though findings do not establish causality, transportability, a diagnosis of malnutrition, or a clinically actionable albumin threshold.
Key Findings
Results
Albumin 30 to <35 g/L was associated with longer length of stay and higher hospital charges compared with albumin ≥35 g/L.
Adjusted rate ratio (aRR) for LOS was 1.24 (95% CI, 1.10–1.40) for albumin 30 to <35 g/L vs. ≥35 g/L.
Adjusted charge ratio (aCR) for hospital charges was 1.28 (95% CI, 1.12–1.46) for the same albumin group.
Models used negative binomial regression for LOS and gamma log-link models for charges.
Cohort included 408 hospitalized DFU patients with a PA-positive wound culture after an earlier non-PA culture during the same admission.
Results
Albumin <30 g/L was associated with substantially longer LOS and higher charges compared with albumin ≥35 g/L.
aRR for LOS was 1.52 (95% CI, 1.29–1.78) for albumin <30 g/L vs. ≥35 g/L.
aCR for hospital charges was 1.67 (95% CI, 1.40–1.99) for albumin <30 g/L vs. ≥35 g/L.
These estimates represent a dose-response-like pattern across the three albumin categories.
Results
Each 5-g/L lower albumin was associated with approximately 19% longer LOS and 22% higher hospital charges.
Per 5-g/L lower albumin, aRR for LOS was 1.19 (95% CI, 1.14–1.25).
Per 5-g/L lower albumin, aCR for charges was 1.22 (95% CI, 1.16–1.28).
Baseline-only estimates changed by less than 3% on the log-ratio scale, suggesting covariate adjustment had minimal impact on albumin associations.
Results
The cohort had a median LOS of 14 days and median nominal hospital charges of CNY 18,708.
Median LOS was 14 days (IQR, 9–19 days).
Median nominal hospital charges were CNY 18,708 (IQR, 12,532–27,979).
The cohort comprised 408 patients, of whom 261 (64.0%) were men.
This was a single-center retrospective cohort study.
Results
Adding early-admission albumin to the core prognostic model increased the apparent AUC from 0.679 to 0.758 in the development sample.
Optimism-corrected AUC increased from 0.643 to 0.729 after adding albumin (ΔAUC, 0.079; bootstrap 95% CI, 0.036–0.120).
In the temporal assessment sample, AUC was 0.592 for the core model and 0.719 after albumin was added.
Corresponding Brier scores were 0.234 (core) and 0.204 (with albumin); calibration slopes were 0.625 (core) and 0.994 (with albumin).
The authors noted these small-sample performance estimates were treated as descriptive.
Results
The median bootstrap-optimal albumin cutoff was 35 g/L.
Median bootstrap-optimal albumin cutoff was 35 g/L (95% bootstrap percentile interval, 33–37 g/L).
The authors explicitly stated this does not establish a clinically actionable albumin threshold.
This cutoff corresponded to the categorical grouping used in the primary association analyses.
Methods
The study cohort was subject to immortal-time and collider-selection bias due to the serial-culture eligibility requirement.
Cohort eligibility required continued hospitalization until repeat sampling and later PA detection, a post-admission selection event.
The culture sequence was described as 'an operational ascertainment pattern, not evidence of nosocomial acquisition, microbiological conversion, or invasive PA infection.'
The authors stated that inference was conditional on a post-admission selection event susceptible to immortal-time and collider-selection bias.
The findings were explicitly stated to not establish causality, transportability, a diagnosis of malnutrition, or a clinically actionable albumin threshold.
Background
Early-admission serum albumin was characterized as reflecting multiple physiological processes rather than serving as a stand-alone measure of nutritional status.
The paper described albumin as reflecting 'inflammation, capillary permeability, hepatic synthesis, protein loss, hydration, and nutritional reserve.'
The authors explicitly stated it 'is not a stand-alone measure of nutritional status.'
This framing was used to contextualize the association findings without attributing effects solely to nutritional mechanisms.
What This Means
This research examined whether a blood protein level called serum albumin, measured shortly after hospital admission, was linked to how long patients stayed in the hospital and how much their care cost. The study focused on a specific group of 408 patients with diabetic foot ulcers who had wound cultures taken multiple times during the same hospital stay, with a bacteria called Pseudomonas aeruginosa detected on a later culture. The researchers found that patients with lower albumin levels at admission tended to have longer hospital stays and higher hospital charges, with each 5-unit drop in albumin associated with about 19% longer stays and 22% higher costs. Patients with the lowest albumin levels (below 30 g/L) had stays about 52% longer and charges about 67% higher than those with albumin of 35 g/L or above.
This research also explored whether albumin could help predict which patients would have worse outcomes. Adding albumin to a prediction model improved its accuracy in both the original group of patients studied and in a later group of patients used to test the model. However, the study authors were careful to note important limitations: the way patients were selected for this study (requiring them to remain hospitalized long enough for multiple cultures) could have introduced biases that make the findings hard to interpret. The finding that Pseudomonas appeared on a later culture does not necessarily mean the patient caught the bacteria in the hospital or developed a serious infection from it.
This research suggests that low albumin at hospital admission is associated with greater hospitalization burden in this specific patient population, but the authors emphasize that these results do not prove that low albumin causes worse outcomes, that albumin alone can diagnose malnutrition, or that there is a specific albumin level that should trigger a particular clinical action. The findings are specific to this single hospital and this unusual patient selection process, so they may not apply to other settings or patient groups.
Kan F, Wang A, Tan Y, Dai H. (2026). Early-admission serum albumin and hospitalization burden in a selected serial-culture cohort of patients with diabetic foot ulcers and subsequent Pseudomonas aeruginosa detection: a retrospective cohort study.. Frontiers in endocrinology. https://doi.org/10.3389/fendo.2026.1918069