Association Between Chronic Osteomyelitis and Risk of End-Stage Renal Disease A Nationwide Population-Based Cohort Study

Association Between Chronic Osteomyelitis

and Risk of

End-Stage Renal Disease

A Nationwide Population-Based Cohort Study

Shih-Yi Lin, MD, Cheng-Li Lin, MSc, Chun-Hung Tseng, MD, Yen-Jung Chang,

PhD,

I-Kuan Wang, MD, Hung-Chieh Yeh, MD, and Chia-Hung Kao, MD

INTRODUCTION

E

impairment and interfere with work productivity. In addition, the high cost of treatment for patients with ESRD causes a

financial burden for health care systems.2 Hence, early recognition

and prevention of risk factors for ESRD could diminish its social and economic cost.

Over the past decade, attention has been focused on

identifying the risk factors for ESRD.3 Age, male sex, diabetes,

hypertension, coronary artery disease (CAD), and congestive heart failure (CHF) are currently considered risk factors for ESRD.4,5 Inflammation has also been reported to be associated

with ESRD risk.6,7 The pathophysiological explanation for the

ESRD risk associated with inflammation is still unresolved but likely involves a systemic response and vascular atherosclerosis.

8,9 Exploring and evaluating the ESRD risks associated with

chronic inflammation-related diseases is necessary to elucidate the association between ESRD and inflammation.

Chronic osteomyelitis (COM), a lasting infection of the bones, typically evokes intense inflammation within bony

structures and nearby soft tissues.10 COM is difficult to eradicate,

and its therapy typically requires weeks, months, or years to complete.11 COM has been reported to increase the risk of

Because CAD and stroke have many risk factors in common with ESRD,17 investigating the possible relationship between

COM and ESRD is warranted. No study has connected COM with the risk of ESRD. We used a nationwide population

database to assess the association of COM and risks of developing ESRD in a cohort study over a follow-up period of

14 years.

MATERIALS AND METHODS

Data Source

Data were extracted from the National Health Insurance Research Database (NHIRD) of the Taiwan National Health Insurance (NHI) program. This insurance program has provided health care for>99%of the >23 million residents of Taiwan and contracted with 97% of Taiwan’s hospitals and clinics. Taiwan

launched a NHI in 1995, operated by a single buyer, the government. Medical reimbursement specialists and peer review should

scrutinize all insurance claims. The diagnoses were based on the International Classification of Diseases, Ninth Revision (ICD-9) codes that were judged and determined by related specialists and physicians according to the standard clinical criteria. If these hospitals or doctors made the wrong codes or diagnoses, they would be punished to pay a lot of penalty. Therefore, the diagnoses and codes used in this study should be correct and reliable.18 For this study,we usedNHIRadministrative data19 that

contains health care data including records of inpatient claims, a registry of catastrophic illness patients, and a registry of beneficiaries. Records were linked using a scrambled, anonymous

identification number for each patient to obtain a longitudinal medical history. Diagnoses are coded according to the ICD-9. This study was approved to exempt from requiring informed consent by the Institutional Review Board of China Medical University (CMU-REC-101–012).

Study Patients

A retrospective cohort study was conducted to examine the association betweenCOM 9 code 730.1) and ESRD (ICD-9 code 585) development. Patients aged _20 years with newly diagnosed COM between 1997 and 2010 were included in the study cohort. The index date was the date of COM diagnosis. Those with a history of chronic kidney disease or ESRD before

the index date were excluded. The comparison cohort comprised patients who had no history of COM, chronic kidney

disease, or ESRD and were frequency matched with the study cohort at a ratio of 1:4 according to age (every 5 years), sex, and index year of diagnosis. The IDC-9 codes about COM (ICD-9 code 730.1) and ESRD (ICD-9 code 585) used in this study from NHIRD in Taiwan are highly reliable, because many related studies about ESRD and COM were already published.20–23

Outcome Measures

Both cohorts were followed until a diagnosis of ESRD or until loss to follow-up, death, termination of insurance, or the end of 2010. ESRD was identified from the Registry for

Catastrophic Illness Patient Database. Registration for catastrophic illness requires a diagnosis made by a physician and

pathological confirmation or other supporting medical information; these documents are formally reviewed by the Bureau

of NHI. Conditions diagnosed before the index date, namely, diabetes (ICD-9 code 250), hypertension (ICD-9 codes 401– 405), hyperlipidemia (ICD-9 code 272), coronary heart disease (CHD) (ICD-9 codes 411.1, 411.81, 411.89, 413, 414.0, 414.8, and 414.9), CHF (ICD-9 code 428), hyperuricemia (ICD-9 code 790.61), gout (ICD-9 code 274.9), and proteinuria (ICD-9 code 791) were identified as comorbidities.

Statistical Analysis

The proportionate distributions of sociodemographic

characteristics and comorbidities between the cohorts with and without COM were compared using the x2 test. The sex-,

comorbidity-, and age-specific incidence rates of ESRD per 10,000 person-years of follow-up for each cohort were calculated. Poisson regression was used to estimate the COM-to-comparison incidence rate ratio (IRR) with a 95%confidence interval (CI).To investigate the risk of developing ESRD associated with COM, Cox proportional hazard regression modelswere used to estimate the hazard ratios (HRs) of developing ESRD in patients with COM compared with those without COM. The Kaplan–Meier

methodwas used with the log-rank test to compare the probability of ESRD-free events between the 2 cohorts. All statistical

(Version 9.2 for Windows; SAS Institute, Inc., Cary, NC). A Kaplan–Meier survival curve was plotted using R software (Version 2.14.1; R Development Core Team, Vienna, Austria). Statistical significance was set at a¼0.05.

RESULTS

We identified 24,267 patients newly diagnosed with COM between 1997 and 2010 and 97,068 patients in the non-COM comparison cohort (Table 1). The mean follow-up years in COM cohort is 5.29_3.96 years and in non-COM cohort is 6.21_3.88 years. In both the cohorts, more than half of the patients were _55 years, and 66.5% were male. Compared with the comparison cohort, patients with COM were more likely to have diabetes (28.0% vs 6.05%; P<0.001), hypertension (30.1% vs 11.2%; P<0.001), hyperlipidemia (6.97% vs

2.66%; P<0.001), CHF (14.1% vs 5.09%; P<0.001), hyperuriemia and gout (2.53% vs 0.50%; P<0.001), and proteinuria

(0.37% vs 0.11%; P<0.001).

Overall, the incidence rate of ESRD was 4.55-fold higher in the COM cohort than in the non-COM cohort (59.6 vs 13.1 per 10,000 person-years), with the adjusted HR being 2.01 (95% CI: 1.81–2.25). Sex-specific analysis showed that patients with COM had a higher risk of developing ESRD compared with patientswithoutCOMfor bothwomen (adjustedHR¼1.81, 95% CI: 1.52–2.17) and men (adjusted HR¼2.03, 95% CI: 1.76– 2.34). Age-specific analysis showed that the IRR was the highest in younger adults aged 20 to 34 years (IRR¼34.0, 95%CI: 28.5– 40.5), with the adjusted HR being 17.8 (95%CI: 5.18–61.4). The corresponding adjusted HR decreased to 1.30 (95% CI: 1.10– 1.54) for the oldest group. The adjusted HR of ESRD of the comorbidity-specific COM cohort relative to the non-COM

cohort was significant for both the subgroups without comorbidity (adjusted HR¼1.57, 95% CI: 1.23–2.00) and with comorbidity (adjusted HR¼2.25, 95% CI: 1.97–2.57) (Table 2).

Table 3 shows the incidence rate and adjusted HR of ESRD according to the presence of individual comorbidity. A higher incidence rate of ESRD was observed in patients having any comorbidity in both the cohorts. COM patients with

with the non-COM patients with no comorbidity (adjusted HR of diabetes¼1.53, 95% CI: 1.23–1.85; adjusted HR of hypertension ¼2.06, 95% CI: 1.77–2.40; adjusted HR of

hyperlipidemia¼2.02, 95% CI: 1.80–2.28; adjusted HR ofCHD¼1.99, 95%CI: 1.76–2.24; adjustedHR ofCHF¼2.06, 2.06, 95% CI: 1.82–2.34; adjusted HR of hyperuricemia and gout¼2.04, 95% CI: 1.82–2.28; adjusted HR of

proteinuria¼2.02, 95% CI: 1.81–2.25). Table 4 shows the incidence rate and adjusted HR of ESRD stratified by age categorization and the presence of comorbidity. Younger COM adults aged 20 to 34 years without comorbidities have a higher ESRD risk than non-COM adults aged 20 to 34 years

without comorbidities (adjustedHR¼8.14, 95%CI: 2.04–32.6). The Kaplan–Meier survival analysis showed that patients with

COM had a significantly higher rate (5.8%) of ESRD development than that of the non-COM cohort (Figure 1).

DISCUSSION

Previous studies have shown a link between chronic inflammatory diseases, such as hepatitis C infection,24,25 hepatitis B

infection,26 stroke,27 gout,28 periodontal disease,29 and herpes

zoster, and ESRD.30 Systemic inflammation is a suspected

mechanism in the relationship between these diseases and an increased risk of ESRD. Using the nationally representative NHIRD to compare patients with COM and controls between

1997 and 2010, this study showed that COM, a chronic inflammatory disease, is associated with an increased risk of ESRD.

Numerous studies have shown a causal link between ESRD risk and old age,31,32 male sex,33 diabetes,33,34 hypertension,33– 35 hyperlipidemia,34,36 CAD,37 and CHF.38 Our data revealed

that COM patients with at least 1 of these comorbidities had an increase in ESRD risk compared with non-COM patients without comorbidities. Further analysis of the interaction between COM and individual comorbidity as well as the risk of ESRD in COM patients and matched controls with or without these comorbidities differed. Our results demonstrate that COM is potentially an independent risk factor for ESRD.

There are several possible physiopatholoical mechanisms accounting for COM cohort that has higher ESRD risk than

non-COM cohort. Chronic infection may cause infection-associated glomerulonephritis that would predispose to nephrons damage, glomerulosclerosis, and thus decline of renal function reserve. Antibiotics for treating COM may also have direct nephrotoxicity or interacting drug–drug toxic effects on renal function of COM patients. A prospective long-term follow-up of COM patients with kidney biopsy data would be necessary to help clarify the causality of COM and ESRD.

Old age has been recognized as a crucial risk factor for ESRD.39 However, in the current study, the COM subgroup of

patients aged 20 to 34 years had an up to 17.8-fold increased ESRD risk compared with the non-COM subgroup of patients

aged 20 to 34 years. This result is attributable to at least 2 factors. First, the competing risk between ESRD and death is higher in elderly people compared with younger patients.40

Therefore, elderly patients might have an increased risk of death from other causes before they are required to initiate dialysis. Second, elderly people have higher possibility to refuse longterm dialysis in consideration of lifespan and underlying complex comorbidities.41 Finally, we collected data in a retrospective

manner and applied strict criteria to enroll patients with and those without COM. The relatively low number of ESRD events in patients without COM aged 20 to 34 years might have caused bias.

Our study has several strengths. First, this retrospective study had a follow-up length of 14 years, and a strict criteria was used to the catastrophic illness registration criteria used to identify ESRD. The long-term follow-up and strict definition of

ESRD diagnosis criteria strengthened the time- and severitydependent effects ofCOMonESRDdevelopment. Second,COM

and age- and sex-matched controls were selected from a dataset exceeding 22 million enrollees and encompassing >99% of the population of Taiwan. This near-total-population sample coupled with a strict case-to-control ratio of 1:4 increased the generalizability, precision, and reliability of its results. Third, an NHI

monitoring and auditing system is implemented to supervise insurance claims to prevent overdiagnosis and medical resource waste. This NHI surveillance program ensures the validity of

diagnosis. Finally, all recognized comorbidities and risk factors of ESRD(ie, hypertension, diabetes, hyperlipidemia,CHF,CAD, hyperuricemia and gout, and proteinuria) were considered and adjusted in this study, and the results suggest that COM is an independent risk factor for ESRD.

Several limitations of this study should be noted. First, we had no definite information on the levels of blood pressure, serum glucose, and serum lipids of the patients. Our study may thus have a confounding variability bias.

The second limitation is that the database used for our research did not provide information on lifestyle and personal health behaviors, including smoking, drinking, and obesity; these variables are known to be related to ESRD. Finally, the results of this study were obtained from insurance claims to calculate the risk of ESRD among the COM patients. Hence, patients who refused long-term dialysis or COM management may have caused us to underestimate or overestimate the effects of COM on ESRD development. This possible bias was minimized because the NHI covers >99% of Taiwan’s population.

Our investigation showed that COM is an independent risk factor for ESRD. Patients with COM have a higher prevalence of conventional risk factors for ESRD. The ESRD risk of patients with COM increases if they have comorbidities (ie, hypertension, diabetes, CAD, CHF, hyperlipidemia, hyperuricemia and gout, and proteinuria). Younger patients with COM

have a higher risk of ESRD. Our findings could be used to prompt clinical alerts and develop renal function screening programs for patients with COM, particularly younger patients