Diabetic hand infections managed at a regional general hospital: a retrospective cohort study of clinical features, surgical outcomes, and the impact of infection depth and comorbidities

Article information

Arch Hand Microsurg. 2025;30(4):204-211

Publication date (electronic) : 2025 November 28

doi : https://doi.org/10.12790/ahm.25.0029

Seong-Kee Shin, Seung-Yeon Cho, Kyu-Jin Kim

Department of Orthopaedic Surgery, Seoul Medical Center, Seoul, Korea

Corresponding author: Kyu-Jin Kim Department of Orthopaedic Surgery, Seoul Medical Center, 156 Sinnae-ro, Jungnang-gu, Seoul 02053, Korea Tel: +82-2-2276-8655 E-mail: fairgain7@gmail.com

Received 2025 August 20; Revised 2025 October 30; Accepted 2025 November 3.

Abstract

Introduction

Although hand infections in diabetic patients pose considerable clinical challenges, they remain underappreciated in both diagnosis and management. Increased infection risk in diabetic patients is attributed to compromised immune function, including leukocyte dysfunction and reduced bactericidal activity, along with impaired tissue healing, angiopathy, and peripheral neuropathy [1-3]. Infection involving any part of the limb, especially the hand, can rapidly advance, resulting in extensive soft tissue damage, necrosis, and, in severe cases, systemic sepsis [4]. Therefore, diabetic hand infections require prompt and appropriate management to prevent severe complications. Failure to initiate appropriate treatment in a timely manner may cause irreversible functional impairment, ranging from stiffness and contracture to limb loss, and in severe cases, fatal septic complications. The aim of this study is to evaluate the demographic and clinical characteristics of diabetic hand infections managed at our institution, and to investigate the relationship between the type of infection, insulin dependency, and the presence of renal failure with treatment outcomes, specifically, the need for multiple surgical interventions and the incidence of amputation.

Methods

All diabetic patients with hand infections who were treated in the Department of Orthopaedic Surgery at our regional general hospital between July 2015 and April 2025 were retrospectively reviewed. Medical records were reviewed to collect data on the type and duration of diabetes mellitus (DM), method of DM management, presence of concomitant renal failure, type and presumed etiology of the hand infection, treatment modalities applied, microbiological culture results, number of surgical interventions, and treatment outcomes at the time of discharge. Among the 23 included patients, 13 (56.5%) presented primarily to our hospital and 10 (43.5%) were referred from other institutions after prior treatment attempts. Seven patients (30.4%) were admitted via the emergency department. Before referral, prior interventions had been performed in a substantial proportion of patients, including systemic antibiotics, incision and drainage (I&D), or simple dressing. Patients with diabetic hand infections who completed follow-up care until discharge were included in this study. Patients who were nondiabetic, lost to follow-up care, or transferred to other institutions before completion of treatment were excluded from the study. Hand infections were categorized into two types: superficial hand infection (SHI), defined as involvement limited to the skin and subcutaneous tissue; and deep hand infection (DHI), defined as extension into the deep fascia, muscles, tendon sheaths, bones, or joints.

The primary outcome was all-cause mortality within 1 year of admission for diabetic hand infection. Medical records were reviewed to identify whether death was directly attributable to the upper-extremity infection. Infection-attributable mortality was defined as death directly resulting from septic complications of the hand infection.

In cases requiring surgical intervention, an aggressive surgical management protocol was implemented. Incisions were extended as necessary to include all regions demonstrating erythema or signs of inflammation. Wounds were debrided thoroughly until viable, healthy tissue was reached, and were left open for secondary healing or further assessment. All patients, regardless of whether they had superficial or DHIs, were managed as inpatients. They received daily wound care using diluted povidone-iodine and hydrogen peroxide solutions. Once the infection was clinically stabilized, patients were discharged and advised to continue wound care with dressing changes every other day. Repeat debridement procedures were performed as clinically indicated. Negative pressure wound therapy (NPWT) was applied when clinically indicated. As soon as the wound showed signs of granulation and healing, appropriate coverage techniques were applied based on the reconstructive ladder. Wound healing by secondary intention was permitted in cases where primary closure or reconstruction was not immediately required. If healing by secondary intention was not feasible, secondary closure or full-thickness skin grafting was attempted as the initial method of wound coverage. Flap coverage was undertaken in wounds where critical structures such as tendons, neurovascular bundles, bones, or joints were exposed. In cases where limb preservation was not feasible, primary amputation was undertaken as part of definitive management. A closed amputation technique with surgical drainage was employed in the majority of patients undergoing primary or secondary amputation. Given the small cohort and limited number of events, we primarily present descriptive summaries and univariable logistic regressions for predictor-outcome pairs. Body mass index (BMI, kg/m²), hemoglobin A1c (HbA1c, %), and duration of DM (years) were entered as continuous variables. When documentation provided only an interval, the midpoint of the reported range was used as duration, and this value was treated as a continuous variable in regression analyses. Insulin dependence, infection depth, renal failure (dialysis), and culture type were treated as categorical covariates. We report effect size with 95% confidence intervals (CIs). Multivariable models were explored but were prone to non-convergence and unstable estimates in this small dataset; therefore, they are presented only as exploratory and interpreted with caution. All statistical analyses were performed using IBM SPSS Statistics for Windows, ver. 30.0 (IBM Corp., Armonk, NY, USA). Demographic, clinical, microbiologic, and outcome data were collected for all patients. Detailed case-level information is presented in Supplementary Table 1.

Results

Among 101 patients diagnosed with hand infections during the study period, 23 cases met the inclusion and exclusion criteria and were included in the final analysis. The average age of patients was 63.0 years. The demographic characteristics of the patients, along with the duration and treatment modalities of DM, are presented in Table 1. Of the 23 patients, 13 (56.5%) presented primarily to our hospital, 10 (43.5%) were referred from other institutions after prior treatment attempts (systemic antibiotics in 70%, I&D in 10%, and simple dressing in 20%), and seven (30.4%) were admitted via the emergency department. Among the patients, eight were diagnosed with SHI and 15 with DHI, and the majority of hand infections were either of unknown etiology or secondary to minor abrasions that were initially overlooked (Table 2). All patients were hospitalized for treatment, and nine of the 23 patients (39.1%) required multiple surgical interventions. All patients underwent operative procedures, including incision, drainage, and debridement. Primary amputation was performed in 11 of the 23 patients (47.8%), while secondary amputation was required in four patients (17.4%). Secondary amputation was performed in cases where the infection could not be controlled despite multiple I&D procedures, or when tissue necrosis developed following flap surgery. Three patients underwent flap surgery: local flaps were used in two cases and a regional flap in one. Flap necrosis developed in one case, leading to subsequent ray amputation (Fig. 1). NPWT was applied in four cases to manage soft tissue defects. Four cases were successfully managed with I&D alone. Among the 23 patients, major amputation was performed in four cases: wrist disarticulation in two patients, trans-radial amputation in one, and trans-humeral amputation in one. The trans-radial amputation was required due to persistent infection that failed to resolve following wrist joint I&D procedure (Fig. 2). The patients who underwent trans-humeral amputation initially presented to the emergency department with necrotizing fasciitis. Despite multiple I&D procedures, the infection was uncontrolled, ultimately necessitating major amputation (Fig. 3). In all cases, wound swabs were obtained for microbial culture and sensitivity testing, and antibiotic regimens were adjusted accordingly (Table 2). Staphylococcus aureus was found in four cases, and gram-negative infection (Enterobacter, Pseudomonas, Klebsiella, Serratia alone or in combination) was also found in eight cases. Fungal infections involving Candida species were identified in three cases (Candida albicans, Trichophyton). No anaerobic bacteria were identified among the 23 patients.

Table 1.

Patient demographics and DM characteristics

Table 2.

Hand infection characteristics and microbiology (n=23)

Fig. 1.

(A) A 79-year-old woman presented with a chronic ulcer on the dorsum of the hand that had persisted for several months. (B) Initial surgical debridement was performed to remove necrotic tissue from the ulcerated area. (C) Due to progressive necrosis of the second digit phalanx, digital amputation was performed, and a rotational flap was used to reconstruct the soft-tissue defect on the dorsum of the hand. (D) The rotational flap used for soft-tissue coverage became necrotic postoperatively. (E) A second ray amputation was performed following flap necrosis. (F) After exposure of the extensor tendon over the third metacarpal head, the defect was covered using acellular dermal matrix and negative-pressure wound therapy, followed by regular dressing changes.

Fig. 2.

(A) A 56-year-old male patient presented to the emergency department with left foot necrosis and generalized pain. (B) Initial plain radiograph of the wrist showed no remarkable findings at presentation. (C) After admission to the infectious disease department and below-knee amputation, a follow-up wrist radiograph obtained one month later demonstrated severe soft-tissue and osteomyelitic changes. (D) Although incision and drainage of the wrist joint were performed, a subsequent radiograph several days later showed worsening osteomyelitis. (E) Magnetic resonance imaging revealed severe bony destruction and osteolytic changes. (F) Due to progressive infection, transradial amputation was ultimately performed. (G) A clean amputation stump was achieved with ongoing dressing and wound care.

Fig. 3.

(A) A 66-year-old male patient presented to the emergency department with several days of generalized weakness. A plain radiograph of the left upper extremity revealed a radiolucent lesion. (B) Computed tomography scan demonstrated extensive gas-forming soft-tissue lesions extending from the distal humerus to the forearm on the left side. (C) Emergency fasciotomy and incision and drainage (I&D) were performed on the day of presentation. (D) The infection persisted despite multiple I&D procedures. (E) Due to uncontrolled infection, a transhumeral amputation was ultimately performed at the mid-humeral level.

The mean hospital stay among the patients was 44.7 days. Of the 23 patients, five patients (21.7%) died within 1 year of admission. On review of medical records, all five deaths were attributable to underlying medical comorbidities (e.g., end-stage renal disease [ESRD]-related complications, cardiovascular disease) rather than the hand infection itself; infection-attributable mortality was zero. On univariable logistic regression, dialysis was associated with higher odds of 1-year all-cause mortality (odds ratio [OR], 4.8; 95% CI, 0.66–35.2), although this did not reach statistical significance (p=0.12). Deep infection, insulin dependence, and polymicrobial infection were not significantly related to mortality. Among the five patients who died, four had deep infections and one had a superficial infection. No other significant complications were observed following treatment in this series.

On univariable logistic regression, deep infection was significantly associated with amputation (OR, 12.8; 95% CI, 1.7–97.2; p=0.013). The full multivariable model including all predictors did not converge due to the small sample size and high collinearity. Continuous predictors (BMI, HbA1c, and DM duration) did not demonstrate clear associations with the outcomes.

Discussion

In this study from a regional general hospital including a substantial proportion of referred and emergency cases, DHI was identified as a significant predictor of amputation in diabetic patients, consistent with prior research emphasizing the devastating impact of deep-seated infections in this population [5]. Deep infections often involve critical structures such as tendons, bones, and joints, which compromise both local immune response and the mechanical integrity necessary for hand function [6]. Furthermore, extensive tissue involvement may limit the efficacy of debridement alone, necessitating amputation to achieve infection control.

The overall amputation rate in our cohort was 65.2%, higher than previously reported rates ranging from 13.9% to 39.0% [7-11]. This discrepancy may reflect the high proportion of deep infections (65.2%) and comorbidities such as renal failure, and the referral pattern of our institution. Importantly, no patients died directly from the hand infection; all deaths were attributable to underlying medical comorbidities. However, dialysis was associated with 4.8-fold higher odds of 1-year all-cause mortality, although this did not reach statistical significance (p=0.12), likely reflecting the limited sample size. Taken together, these findings suggest that dialysis remains an important high-risk feature. This is consistent with evidence that ESRD is associated with poorer outcomes in diabetic infections [12,13].

Interestingly, other factors traditionally associated with poor diabetic wound outcomes, including BMI, HbA1c, DM duration, and insulin dependence, did not show significant associations with either amputation or the number of surgical interventions in our analysis. This finding may be due to the relatively small sample size or may suggest that once a severe hand infection develops, the anatomical depth of infection and systemic factors like renal failure play a more decisive role than glycemic control alone [6].

The microbiological profile in our study highlighted gram-negative pathogens as more common than traditionally expected, with Enterobacter, Pseudomonas, and Klebsiella frequently isolated. This pattern may be influenced by the hospital environment or prior antibiotic exposure, contributing to a shift from typical gram-positive organisms like S. aureus. Moreover, polymicrobial infections were present in nearly one-third of cases, though this was not statistically linked to increased mortality or amputation in this cohort.

Our protocol involving aggressive surgical debridement, NPWT, and staged reconstruction aligns with best practices in diabetic limb salvage [4]. However, despite these efforts, the mortality rate was 21.7%, emphasizing the systemic vulnerability of this patient group.

This study has several limitations. First, its retrospective design and small sample size limit generalizability; accordingly, this work is best interpreted as a descriptive, hypothesis-generating case series. We emphasize univariable findings and effect sizes with 95% CIs, and any multivariable results are exploratory. Second, because our hospital functions as a regional referral center, our sample likely over-represents severe or complicated cases transferred after prior treatment attempts at other institutions. This referral bias may overestimate the rate of adverse outcomes compared with community-based cohorts. However, our findings may be particularly relevant for clinicians managing severe diabetic hand infections in similar settings.

Conclusion

This single-center study from a regional general hospital, including referred and emergency cases, suggests that DHI appears to be associated with a higher risk of amputation in diabetic patients. Recognition of high-risk features and timely surgical management may help improve outcomes. Further prospective, multicenter studies are warranted to better define optimal treatment protocols.

Supplementary materials

Article information Continued