Abstract
Hidradenitis suppurativa (HS) is a chronic inflammatory skin condition associated with disrupted skin integrity, while end-stage renal disease (ESRD) is marked by immune dysfunction and a heightened risk for systemic infections. The intersection of these conditions may increase susceptibility to bloodstream infections, yet this relationship remains underexplored. This study aimed to assess whether HS increases the risk of bacteremia, septicemia, and candidemia in patients with ESRD. A retrospective cohort study was conducted using the United States Renal Data System (USRDS). Adults aged 18–100 who initiated hemodialysis between 2005 and 2019 were included. A diagnosis of HS was determined via International Classification of Diseases (ICD)-9 and ICD-10 codes. Infectious outcomes, also assessed using ICD codes, were measured following HS diagnosis. Multivariable logistic regression models were used to estimate the relative risk (RR) of infection. Among 1,397,827 ESRD patients, 0.1% had HS, a prevalence that aligns with previously reported rates in the general U.S. population, suggesting that most HS cases were likely captured. After adjustment, HS was not significantly associated with diagnoses of bacteremia (RR = 0.98, 95% CI: 0.87–1.09) or candidemia (RR = 1.25, 95% CI: 0.86–1.83). However, HS was independently associated with a significantly increased risk of an ICD diagnosis of septicemia (RR = 3.24, 95% CI: 2.41–4.34). Although the prevalence of a bacteremia or candidemia diagnosis was not increased in ESRD patients with HS, these patients were more likely to be labeled as having “septicemia” per ICD codes. This highlights the importance of heightened clinical awareness and the potential need for tailored infection-prevention strategies in this subpopulation.
Introduction
Hidradenitis suppurativa (HS) is a chronic inflammatory skin condition characterized by painful nodules, abscesses, and sinus tracts, primarily affecting areas with apocrine glands such as the axillae, groin, and perianal regions. The global prevalence of HS is estimated to be between 0.00033% and 4.1%, with a more likely range of 0.7% to 1.2% in European and U.S. populations. 1 The pathogenesis of HS involves follicular occlusion, immune dysregulation, and bacterial infection, leading to significant morbidity and a reduced quality of life. 2 HS is associated with several comorbidities, including metabolic syndrome, inflammatory bowel disease, and psychiatric disorders, 3 and is more common in women and people of color, with a peak incidence in the second and third decades of life. 1
End-stage renal disease (ESRD) is an advanced form of chronic kidney disease in which kidney function declines to the point that dialysis or transplantation is required to sustain life. 4 Patients with ESRD are at an increased risk of infections due to chronic uremia, which impairs the immune system, and the need for frequent vascular access for dialysis. 5 The combination of these factors makes ESRD patients particularly vulnerable to systemic infections, which contribute significantly to morbidity and mortality. 5
The relationship between skin conditions and renal disease has been explored in various studies. For instance, psoriasis has been associated with an increased risk of certain infections in ESRD patients. 6 However, the specific impact of HS on infection risk in ESRD patients remains unexplored. Given the compromised skin integrity in HS and the heightened infection risk in ESRD, it is plausible that HS could exacerbate the risk of infections in this population. On the other hand, HS patients typically show elevated levels of several inflammatory mediators, including interleukin-17 (IL-17), a cytokine key for immune responses to microorganisms (reviewed in Curtis and Way, 7 Conti and Gaffen 8 ), and in particular antifungal immunity and prevention of candidiasis. 9 Therefore, we sought to determine whether HS might be associated with an altered risk of systemic infections, such as candidemia, bacteremia, or (non-specific) septicemia. We wished to determine whether a worsening of overall skin integrity due to HS might potentiate the existing risk in ESRD patients. To fill in the current gap in evidence, the purpose of this study was to determine the risk of systemic infection in patients with both ESRD and HS, in a study using data from the United States Renal Data System (USRDS). 10
Significance Statement
Hidradenitis suppurativa (HS) results in compromised skin integrity and is characterized by elevated levels of cytokines known to be important for antimicrobial responses.
Patients with end-stage renal disease (ESRD), which is marked by immune dysfunction, show an increased risk for infection.
Skin diseases that disrupt barrier function, such as atopic dermatitis and psoriasis, have been linked to higher infection risk in patients with ESRD.
In simple bivariate models, HS was associated with an increased risk of bacteremia, candidemia, and septicemia in ESRD patients.
In the final model, after controlling for covariates, HS was not associated with a significantly increased risk of candidemia or bacteremia in ESRD patients.
HS was independently associated with a significant approximately 3-fold increased risk of septicemia in ESRD patients after controlling for covariates.
These findings suggest that clinicians may need to monitor ESRD patients with HS more closely for early signs of sepsis and consider tailored infection-prevention strategies.
Methods
Population
All ESRD patients in the USRDS on hemodialysis (HD), who had not had a kidney transplant, were between 18 and 100 years of age, and started dialysis between 2005 and 2019, were eligible for inclusion in the analysis. Subjects with missing or unknown race, sex, ethnicity, access type, or who had no follow-up were excluded from the analysis sample.
Main independent variable
The main independent variable was having a diagnosis of HS at least once following the start of dialysis. HS was determined using International Classification of Diseases (ICD)-9-CM and ICD-10-CM codes in hospital, detailed, or physician/supplier claims (Supplemental Table 1).
Outcome variables
Three different infectious outcomes were of interest: bacteremia, septicemia, and disseminated candidiasis/candidemia. The first occurrence of the infection diagnosis must have occurred following the HS diagnosis and was determined using ICD-9-CM and ICD-10-CM codes in hospital, detailed, or physician/supplier claims. The person-years at risk for each infectious outcome were determined using the number of years between the start of dialysis and the first occurrence of the infection diagnosis. For those without a diagnosis of the specific infection, the number of years between the start of dialysis and the date of death, for those who died, or December 31, 2019, for those who did not die, was used.
Demographic and clinical covariates or confounders
Demographic variables of interest included age at the start of dialysis, race, sex, ethnicity, and access type, and were determined from the Centers for Medicare and Medicaid (CMS) Medical Evidence Form 2728. Clinical variables of interest included tobacco use, alcohol use/dependence, diabetes, systemic lupus erythematosus (SLE), human immunodeficiency virus (HIV) infection, or an antibiotic use diagnosis. Clinical variables were determined using ICD-9-CM and ICD-10-CM codes in hospital, detailed, or physician/supplier claims data files or from the CMS Medical Evidence Form 2728 data file. All clinical risk factors must have occurred before the HS diagnosis for individuals with HS. For individuals without an HS diagnosis, the clinical risk factor must have occurred before the start of dialysis, plus the median time to HS (849 days), to ensure similar follow-up time between those with and without HS when determining whether a clinical diagnosis was made.
Statistical methods
All statistical analyses were performed using SAS 9.4 (SAS, Cary, NC), and statistical significance was determined using an overall alpha level of 0.05. Descriptive statistics were determined overall, by HS status, and by each infectious outcome. To examine potential demographic or clinical covariates or confounders of the association of HS with each outcome, Chi-square tests and two-sample t-tests, where appropriate, were used to examine differences between those with and without HS.
To assess the association of HS with each infectious outcome, logistic regression with an offset parameter of the natural log of the person-years at risk was used. For each outcome, HS, demographic, and clinical covariates were first examined in simple, bivariate logistic regression models. Then all variables were entered into a comprehensive multivariable model for each outcome, and a backward model-building strategy was used to arrive at the final model. First, the largest, non-statistically significant variable was removed from the model, and Akaike’s Information Criterion (AIC) and a −2-log-likelihood (−2LL) test were used to examine model fit. If the AIC was lower and the −2LL test was non-significant, the reduced model was considered a better fit to the data, the variable remained out of the model, and the next highest non-statistically significant variable was considered for removal from the model. If the AIC was higher or the −2LL test was statistically significant the variable was entered back into the model and the next highest non-statistically significant variable was considered for removal from the model. The final comprehensive multivariable logistic regression model for each infectious outcome consisted of HS and any demographic or clinical variable that was statistically significant or needed in the model to improve model fit to the data. The relative risk (RR) and 95% confidence interval (CI) are reported.
Results
Among the 1,397,827 individuals with ESRD included in the analysis, the average age was 64.9 years; 28% were Black, 6% were of another race, 15% were Hispanic, and 43% were female (Table 1). Nearly 82% were using a catheter, and 3.3% were using a graft for their dialysis access. Only 0.1% (n = 1455) had a diagnosis of HS (Table 1). Those with HS had a lower average age and had higher percentages of black, female, and graft or arteriovenous (AV) fistula access than those without HS. Those with HS also had higher percentages of tobacco usdependence, diabet/dependence, diabetes, systemic lupus erythematous (SLE), human immnodeficiency virus (HIV), and an antibiotic use diagnosis (Table 1).
Descriptive statistics overall and by HS with Chi-square and two-sample t-test results for differences between HS groups.
HS: hidradenitis suppurativa; SD: standard deviation; SLE: systemic lupus erythematosus; HIV: human immunodeficiency virus; RR: relative risk; CI: confidence interval.
Black-shaded cells indicate that a variable was not examined due to low or zero frequencies in simple models.
The outcome bacteremia was associated with an increased RR of HS of RR = 4.96 (95% CI: 4.46–5.53; Table 2) in simple models; however, after controlling for demographic and clinical covariates, the association decreased and became non-significant (RR = 0.98, 95% CI: 0.87–1.09). The association of HS was found to be confounded with tobacco use, diabetes, and age at incident dialysis, with tobacco use and diabetes resulting in a decreased RR, and age at incident dialysis resulting in nonsignificance. All three were confounders together, with none of them being a confounder in and of itself. Other variables that were associated with a decreased risk of bacteremia were increasing age, other race compared to the white race, and Hispanic ethnicity. Variables that were associated with an increased risk of bacteremia included black race compared to white race, catheter and graft access types compared to AV fistula, tobacco use, alcohol use/dependence, diabetes, SLE, HIV, and an antibiotic use diagnosis. Figure 1 shows the forest plot of the final model of HS-associated risk for bacteremia, controlling for the demographic and clinical covariates/confounders.
Descriptive statistics and logistic regression results for bacteremia.
SD, standard deviation; SLE, systemic lupus erythematosus; HIV, human immunodeficiency virus; RR, relative risk; CI, confidence interval.
Odds ratio is for a 1-year increase in age at incident dialysis.
Referent level.
Forest plot of the final logistic regression model of HS association with risk for bacteremia, controlling for demographic and clinical variables.
Risk of septicemia diagnosis was associated with HS with a RR = 40.62 (95% CI: 33.04–49.96) in simple models (Table 3). After controlling for demographic and clinical covariates, the association indicated that those with HS were at increased risk of septicemia; the magnitude of the association decreased to a RR = 3.24 (95% CI: 2.41–4.34). It was determined that the association of HS with septicemia was confounded by both tobacco use and diabetes. Other variables associated with a decreased risk of septicemia included black or other race compared to white race and Hispanic ethnicity. Variables associated with an increased risk of septicemia included age, female sex, tobacco use, alcohol use/dependency, diabetes, SLE, HIV, and an antibiotic use diagnosis. Figure 2 shows the forest plot of the final model of HS-associated risk for septicemia, controlling for the demographic and clinical covariates/confounders.
Descriptive statistics and logistic regression results for septicemia.
SD, standard deviation; SLE, systemic lupus erythematous; HIV, human immunodeficiency virus; RR, relative risk; CI, confidence interval.
Odds ratio is for a 1-year increase in age at incident dialysis.
Referent level.
Forest plot of the final logistic regression model of HS association with risk for septicemia, controlling for demographic and clinical variables.
In simple models, HS was associated with an increased risk for candidemia with a RR = 4.31 (95% CI: 2.96–6.28). In the final model, although HS remained associated with an increased risk for candidemia (Table 4) with a RR = 1.25 (95% CI: 0.86–1.83), the association was no longer significant. This association of HS was found to be confounded with tobacco use, diabetes, and age at incident dialysis, resulting in non-significance. Again, all three were confounders together, with none being a confounder in and of itself. Variables associated with a decreased risk of candidemia included age at dialysis onset, other race compared to the white race, and Hispanic ethnicity. Variables associated with an increased risk of candidemia included black compared to white race, female sex, catheter and graft access types compared to AV fistula, tobacco use, alcohol use/dependence, diabetes, SLE, HIV, and an antibiotic use diagnosis. Figure 3 shows the forest plot of the final model of HS on candidemia, controlling for the demographic and clinical covariates or confounders.
Descriptive statistics and logistic regression results for disseminated candidiasis/candidemia.
SD, standard deviation; SLE, systemic lupus erythematosus; HIV, human immunodeficiency virus; RR, relative risk; CI, confidence interval.
Odds ratio is for a 1-year increase in age at incident dialysis.
Referent level.
Forest plot of the final logistic regression model of HSD association with risk for disseminated candidiasis/candidemia, controlling for demographic and clinical variables.
Discussion
The aim of this study was to assess whether the risk of systemic infections, including bacteremia, septicemia, and candidemia, was increased in patients with ESRD who also have HS compared to ESRD patients without HS. From the USRDS total cohort of 1,397,827 ESRD patients, 0.1% were identified with a diagnosis of HS. Among ESRD patients, those with HS were generally younger, predominantly female, and more likely to be Black. They also had higher rates of tobacco use, alcohol use/dependence, diabetes, SLE, HIV, and an antibiotic use diagnosis compared to those without HS.
Although ESRD patients are highly susceptible to bacteremia due to frequent vascular access and immune dysfunction, 5 the risk of bacteremia in patients with both HS and ESRD was not significantly elevated in our adjusted models, despite initial indications of increased risk. The lack of significant association of HS with bacteremia in our study may be because of the confounding effects of age, diabetes, and tobacco use, which were controlled for in our analysis. Septicemia, a severe bloodstream infection characterized by a systemic inflammatory response, is a significant concern for patients with ESRD due to their compromised immune systems and other comorbidities. In our study, we found that ESRD patients with a diagnosis of HS had an increased risk of having an ICD diagnosis of septicemia, even after adjusting for various demographic and clinical covariates. This finding aligns with the understanding that ESRD patients are particularly vulnerable to systemic infections, which are likely exacerbated by the presence of skin conditions, such as HS, that compromise skin integrity. The increased risk of septicemia in this population detected in this study suggests the need for vigilant monitoring and management of infection risks, especially in those with additional risk factors such as tobacco use and diabetes.
Our study did not find a statistically significant associated increased risk for candidemia in patients with HS after adjusting for covariates and confounders. The risk associated with HS was non-significant because of the confounding effects of diabetes and other comorbidities, which are prevalent in the ESRD population. 11 Previous research has shown that candidemia is a common complication in immunocompromised individuals, including those with ESRD. 12 It is possible, too, that the diagnosis of septicemia may include cases of candidemia; indeed, the Infectious Diseases Society of America recommends considering the use of empiric antifungal therapy in patients with a fever of unclear origin and known risk factors for invasive candidiasis, such as dialysis. 13 The role of HS in the context of susceptibility to this infection in patients with ESRD remains unclear, warranting further investigation.
Our findings align with existing literature highlighting increased infection risks in ESRD patients. Previous studies have demonstrated that skin conditions such as psoriasis, atopic dermatitis, and pemphigus vulgaris are linked to heightened infection risk in ESRD populations, although the specific infections and risk factors may vary. 14 For example, research on atopic dermatitis in ESRD patients revealed an elevated risk of cellulitis and herpes zoster, but not bacteremia or septicemia. 15 Similarly, psoriasis is associated with an elevated risk of multiple infections, including cellulitis, conjunctivitis, systemic inflammatory response syndrome, pericarditis, fungemia, onychomycosis, and herpes zoster in ESRD patients. 6 On the other hand, studies on pemphigus vulgaris suggest an increased risk of various infections in simple models, although associations became non-significant after adjusting for confounders such as diabetes. 14 These findings highlight the complex interplay of factors influencing infection risk in ESRD patients with skin conditions, warranting continued investigation and targeted clinical interventions. In contrast, studies have also found HS to be associated with renal dysfunction, 16 showing a potential interaction between HS and the development of ESRD.
Infection risk may also be shaped by comorbidities frequently associated with HS—most notably, diabetes mellitus. Several studies have established that HS is often comorbid with metabolic syndrome and type 2 diabetes. In a German cohort, 20.2% of HS patients had diabetes, with those at Hurley stage III showing a 5.3-fold higher risk of this comorbidity compared to those at earlier stages. 17 The mechanism of this association between the two may relate to shared inflammatory mechanisms involving IL-1β and TNF-α, which are elevated in both conditions and known to impair insulin signaling. 2 Furthermore, among ESRD patients with diabetes, poor glycemic control has been linked to a higher risk of certain infections. One study found that patients on hemodialysis with Hemoglobin A1c (HbA1c) ≥ 8.5% had 22%–23% higher infection risks compared to those with an HbA1c level between 5.5 and 6.5%. 18 Finally, anecdotal evidence points to a potential bidirectional relationship between HS and diabetes. In one case report, a patient with severe HS achieved diabetes remission following aggressive HS treatment, with HbA1c levels dropping from 10.1% to 5.5% over 36 weeks and resultant discontinuation of antidiabetic medications. 19 These findings underscore the need for integrated dermatologic and metabolic care strategies in this population.
Tobacco smoking is another established risk factor for HS. Longitudinal data from a national cohort of over 6 million adults showed that individuals who quit smoking had a significantly lower risk of developing HS compared to sustained smokers (adjusted hazard ratio: 0.68, 95% CI: 0.56–0.83), with the risk approaching that of never-smokers after 3 years. 20 In contrast, relapsed and new smokers retained the elevated risk profile. 20 Smoking may exacerbate HS through proinflammatory cytokine activation (e.g., TNF-α, IL-1β), immune dysregulation, and disruption of hair follicle homeostasis—mechanisms that could also contribute to higher infection susceptibility in ESRD patients with HS who also smoke. 20
Beyond these comorbidities, ESRD itself is an immunocompromised state. Chronic uremia disrupts both innate and adaptive immunity, leading to neutrophil dysfunction, T and B cell depletion, impaired antigen presentation, and accelerated immunosenescence. 21 Dialysis can further impair immune function, which is typically only restored after renal transplantation. 21 This altered immune profile may leave ESRD patients with HS particularly vulnerable to systemic infections, such as septicemia, even after accounting for traditional clinical risk factors. Taken together, these findings emphasize the need to consider metabolic, behavioral, and immunologic risk factors when managing infection risk in ESRD patients with HS.
We hypothesized that patients with both ESRD and HS might be at increased risk for systemic infections. We also considered the possibility that, since HS is an inflammatory disease characterized by elevated levels of several cytokines, these molecules might stimulate the immune system and actually reduce the risk. Our results suggest that, in this ESRD patient population, the increased inflammation largely balances out the skin disruption such that there is neither an increased nor decreased risk of bacteremia or candidemia. However, in the case of a code-based diagnosis of septicemia, there is a roughly 3-fold increased risk of this diagnosis in ESRD patients with an HS diagnosis. HS-related small abscesses and breaks in the skin likely allow local bacterial translocation on an ongoing basis. Perhaps the activation of the immune system, either by the HS or by the continuous microbial translocation, is somewhat protective against small-inoculum hematogenous dissemination resulting in candidemia and/or bacteremia. However, in terms of septicemia/sepsis, the HS-related dysregulation of the immune system (with chronic elevation of IL-17 and other cytokines) may alter the interaction between host and pathogen, increasing the risk of an overreaction of the immune system with resultant sepsis syndrome in response to larger or multiple inocula.
This study utilized data from the USRDS database and has several limitations. All diagnoses were based on medical billing codes from administrative claims data or documented on form CMS-2728, as laboratory and diagnostic testing results were unavailable; therefore, the data rely on the accuracy of the administrative claims data. Indeed, the lack of information on confirmation of the systemic infections through positive cultures and reliance on ICD codes only may introduce bias as providers may have based coding on general symptoms or signs that raised suspicion for systemic infection, particularly in terms of septicemia. The study population included patients with HS and systemic infection diagnoses recorded during inpatient encounters, which may have excluded patients diagnosed in outpatient settings or those missed due to coding errors. Consequently, not all cases of HS and systemic infections may have been captured if they were not documented during a hospital admission. Nevertheless, the study identified an HS prevalence of 0.1%, which aligns with previously reported rates in the general U.S. population, suggesting that most HS diagnoses were likely included. While associations were observed in the data, causation cannot be established. In addition, the lack of direct testing to confirm HS and systemic infection diagnoses, as well as the inability to assess HS severity from the available data, may introduce inaccuracies and limit the granularity of our findings. Studies validating ICD codes for dermatologic conditions, including HS, suggest that infection risk estimates improve with multiple diagnostic codes or specialist-confirmed diagnoses. 22 Furthermore, research has shown that the prevalence of HS in ESRD patients may be underestimated due to limited clinical awareness and under-reporting in medical records. 23 Nevertheless, a previous study examining the accuracy of HS diagnosis in a retrospective claims database found that 89.6% of patients with two HS diagnostic codes were confirmed to have HS. 22 Thus, in spite of these challenges, the large dataset and substantial number of individuals in the HS group provide strength to the observed associations.
Future research should address the above limitations by incorporating prospective study designs and more detailed clinical data to improve diagnostic accuracy and understanding. Furthermore, additional prospective or retrospective investigations could examine potential mechanisms by which HS increases the risk of septicemia, such as whether systemic infiltration occurs through abscesses of skin or other soft tissue. Alternatively, currrent procedural termonology (CPT) codes for surgical debridement (more specifically, “unroofing”) or wide excision could be used to estimate the severity of the disease and to determine how well-controlled it is. Despite these limitations, our study provides valuable insight into the relationship between HS and systemic infections in ESRD patients. Compared to those without HS, ESRD patients with HS were generally younger, more likely to be Black and female, and had higher rates of tobacco use, alcohol use/dependence, diabetes, SLE, HIV, and an antibiotic use diagnosis—factors that may contribute to elevated infection risk. These findings highlight the need for increased clinical attention and more tailored infection-prevention efforts in this high-risk subgroup.
Supplemental Material
sj-docx-1-imj-10.1177_10815589261462559 – Supplemental material for The association of hidradenitis suppurativa with increased risk of systemic infections in end-stage renal disease
Supplemental material, sj-docx-1-imj-10.1177_10815589261462559 for The association of hidradenitis suppurativa with increased risk of systemic infections in end-stage renal disease by Adaora Ntukogu, Fabiha Anwar, Rushan Momin, Jennifer L. Waller, Stephanie L. Baer and Wendy B. Bollag in Journal of Investigative Medicine
Footnotes
Ethical considerations
The Augusta University Institutional Review Board deemed this research not to constitute human subject research.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported in part by the Translational Research Program of the Augusta University Department of Medicine. WBB is supported by a VA Research Career Scientist Award (#IK6BX005691). SLB is provided research time by the VA Augusta Healthcare System.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data availability statement
The data underlying this article were provided by the United States Renal Data System (USRDS) under a data use agreement. Data will be available from the USRDS upon execution of a data use agreement.
Disclaimer
The data reported here have been supplied by the USRDS. The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as official policy or interpretation of the USRDS or the United States Government. The contents of this article do not represent the views of the Department of Veterans Affairs or the United States Government.
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
