Clinical outcomes among patients with chronic kidney disease hospitalized with diabetic foot disorders: A nationwide retrospective study

Abstract Introduction Diabetic foot ulcerations or infections (DFUs/DFIs) are common complications of patients with diabetes. This study aimed to explore the impact of non‐dialysis and dialysis CKD on hospitalized patients with DFUs/DFIs. Methods A retrospective cohort study was conducted using the National Inpatient Sample database for the years 2017 and 2018. Patients hospitalized for DFUs/DFIs were included in the study. The primary outcome was lower limb amputations. The secondary outcomes were inpatient mortality, sepsis, length of stay (LOS), total hospitalization charges (THC) and disposition. Results A total of 121,815 hospitalizations were included (26.1% non‐dialysis CKD; 8.4% dialysis CKD). There was no significant difference in amputation rates between those on non‐dialysis CKD (adjusted odds ratio [aOR]: 0.96; 95% confidence interval [CI]: 0.87–1.06) and dialysis CKD (aOR: 1.04, [95% CI: 0.91–1.12]) when compared to non‐CKD group. Dialysis CKD group had increased odds of undergoing major amputation (aOR: 1.74, [95% CI: 1.32–2.29]), in‐hospital mortality (aOR: 3.77 [95% CI: 1.94–7.31]), sepsis (aOR: 1.83 [95% CI: 1.27–2.62]), longer LOS (adjusted mean difference [aMD]: 1.46 [95 CI: 1.12–1.80) and higher THC (adjusted mean difference [aMD]: $20,148 [95% CI: $15,968‐$24,327]). Non‐dialysis CKD group had increased odds of sepsis (aOR: 1.36 [95% CI: 1.02–1.82]), less likely to be discharged home (aOR: 0.87 [95% CI: 0.80–0.95]), longer LOS (aMD: 0.91 [95% CI 0.69–1.13]) and higher THC (aMD: $20,148 [95% CI: $15,968–$24,327]). Conclusion Patients with CKD on dialysis had higher odds of undergoing major amputation. CKD increased the odds of in‐hospital morbidity and resource utilization, with the most significant is for those on dialysis.


| INTRODUC TI ON
Foot ulceration is one of the most common complications of patients with diabetes. The lifetime incidence has been predicted to be more than 19% in patients with diabetes. 1 Approximately 58% of diabetic foot ulcers (DFU) would evolve into diabetic foot infection (DFI). 2 DFUs cause a significant burden for the government and debility to the patients themselves. The total estimated cost for management of diabetic foot in the United States ranges from $9 to 13 billion annually in addition to diabetes care itself. 3 Furthermore, up to one-fifth of patients with moderate or severe DFU will lead to amputation and mortality. 1,4 Amputation will affect their ability to perform daily tasks, which in the end will negatively influence their quality of life. 5 A study done by Wukich et al. showed that patients with diabetic foot perceived lower extremity amputation worse than death. 6 Therefore, understanding the risk factors is an essential issue in order to perform early detection of foot complications.
Diabetes is also frequently complicated by kidney disease.
Chronic kidney disease (CKD) is a known marker for the generalized vascular status of patients with diabetes. Overall, CKD is associated with an increased level of in-hospital mortality, postoperative complications, length of stay and hospital costs. 7,8 Several studies have demonstrated an increase in morbidity and mortality in patients on dialysis who develop foot ulcers. [9][10][11] However, studies on the influence of chronic kidney disease on diabetes-related foot ulcer hospitalization have been inconsistent. Lee et al., in a case-control study of 351 DFU subjects, demonstrated a significant relationship between low estimated glomerular filtration rate level and major amputation risk in patients diabetic foot. 12 On the contrary, other investigations have shown no significant relationship between non-dialysis CKD and DFUs/DFIs on hospital outcomes. 13,14 To our knowledge, the impact of non-dialysis and dialysis CKD on patients with DFUs/DFIs has not been explicitly addressed among hospitalized patients at a national level. Therefore, this study aims to explore the impact of CKD on clinical outcomes and resource utilization. In 2017 and 2018, the NIS provides up to 40 diagnoses and 25 procedures for each hospitalization record. The NIS contains a large sample size; therefore, it is ideal for developing national and regional estimates of multiple conditions. The International Classification of Disease, 10 th Revision, Clinical Modification/ Procedure Coding System (ICD-10-CM/PCS) coding system was used to report all medical diagnoses and inpatient procedures recorded in the NIS database. Detailed information is available at http://www.hcup-us.ahrq.gov. 15 The NIS database was queried for patients 18 years and older who had a principal diagnosis of foot ulcers or foot infections and any diagnosis of diabetes mellitus using ICD-10 codes. This cohort was further divided based on at least one secondary discharge diagnosis of non-dialysis CKD and dialysis CKD. Institutional Review Board approval was not needed because the NIS database has completely removed possible patient identifiers, state level and hospital identifiers. 15

| Outcome measures
The primary outcome was any lower limb amputation. Lower limb amputation was further divided into major and minor amputation.
Major amputation was defined as any amputation above the ankle joint. Minor amputation was defined as any amputation limited to the foot. The secondary outcomes included in-hospital mortality, sepsis, disposition, length of stay and total hospitalization charges.
Disposition was defined as either home discharge or all others (transfer to short-term hospital, skilled nursing facility, intermediate care, home health care, against medical advice and died). ICD-10-CM/PCS codes were used to obtain lower limb amputation, major amputation, minor amputation and sepsis from the cohort (Table S1). The other outcome variables were available in the NIS database.

| Statistical analysis
The data were analysed using STATA 16.0 (StataCorp). The NIS database represents a collection of a complex sampling design that includes weighting, clustering and stratification. All analyses were conducted using weighted samples based on guidelines outlined by HCUP NIS. 15 Baseline characteristics of patients and hospitals were compared among patients presented with DFUs/DFIs based on CKD status using Pearson chi-square for categorical variables and ANOVA for continuous variables. Potential confounders were identified based on previous literature, 10,16 and an initial univariate regression analysis was done with a cut-off p-value of .2. Adjusted

| DISCUSS ION
The current study analysed 121,815 admissions from the NIS da-  an increase in resource utilization, including length of stay and total hospitalization charges.
Several studies showed that CKD was associated with worse mortality and morbidity in patients with diabetic foot. 9-12 CKD promotes more severe peripheral vascular diseases by causing chronic inflammation, oxidative stress and inducing a prothrombotic state. The incidence of PAD was directly correlated with the stage of CKD. 17 A significant association between CKD and major amputation was observed in a retrospective cohort study of 669 individuals with foot ulcers. Compared to CKD stage 3, those with CKD stage 4-5 and CKD on dialysis had a higher risk for major amputation (hazard ratios 9.5 and 15.0, respectively; p < .005). 18 In addition, Sayiner et al. concluded that, in patients with DFUs, PAD tripled the odds of having major amputation. 19 On the contrary, some studies showed no significant relationship between diabetic foot amputation and CKD or initiation of dialysis. 13,14,20 Findings from the current study demonstrated that patients from non-dialysis CKD and CKD on dialysis groups had higher overall amputations rates compared to patients without CKD. After adjusting PAD and other confounders, dialysis treatment still significantly increased the odds ratio of undergoing major amputation by 74%. One possible explanation is patients receiving dialysis treatment had worse kidney function compared to other groups.
Furthermore, dialysis treatment itself might also decrease tissue oxygenation and blood flow of the foot. This effect was noticed to be more prominent in patients with diabetes than in patients without diabetes. 21 CKD has consistently been linked to adverse cardiovascular and renal outcomes. This association was not only seen in advanced CKD but also stage 1 or 2 CKD. 22   This study has some important limitations. Firstly, the NIS database uses ICD-10 codes to characterize diagnoses, procedures and hospitalization events. The database does not provide laboratory or imaging parameters and degrees or extent of DFUs/DFIs. Therefore, there is a possibility of misclassification of the diagnoses. Secondly, CKD might be under-reported in the NIS database. ICD-10 codes show a high accuracy for diabetic foot complications and high specificity and low sensitivity for CKD. 28,29 Consequently, some CKD patients were included in the non-CKD group. It means that the misclassification introduced would cause the statistic results more significant. Lastly, there is a risk of residual or unmeasured confounders in the retrospective analysis.

| CON CLUS ION
The present study demonstrated that the CKD population, particularly the dialysis CKD group, is posed to higher amputation rates, worse clinical outcomes and more enormous economic impacts on patients with DFUs/DFIs. The results from this study highlight the need for more research on ways to prevent diabetic foot complications in this high-risk population.

CO N FLI C T S O F I NTE R E S T S
The author reports no conflicts of interest in this work.

AUTH O R CO NTR I B UTI O N S
MS contributed in the conceptualization, data curation, formal analysis, methodology, software, validation, visualization, draft preparation, review and editing.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed during this study are publicly available from the data repositories HCUP database. The database is available online at http://www.hcup-us.ahrq.gov. 15