Increased chronic kidney disease development and progression in diabetic patients after appendectomy: a population-based cohort study

Data source

Data for this study were retrieved from the Taiwan National Health Insurance Research Database (NHIRD), which contains data from all medical claims in the Taiwan National Health Insurance Program. This insurance program started in 1995 and covers >99% of the Taiwanese population of approximately 23 million people. Our research dataset was the Longitudinal Health Insurance Database 2005 (LHID2005), which includes the data from one million patients randomly selected from the NHIRD in 2005 and longitudinally linked with the NHIRD from 1996 to 2013. LHID2005 is managed and established by the National Health Insurance. Diagnostic codes used in the LHID2005 to identify diseases are taken from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). These diagnostic codes have been shown to have high accuracy and validity. (Cheng et al., 2011; Cheng et al., 2014; Hsu et al., 2014). This study was approved after a full ethical review by the Institutional Review Board of the Changhua Christian Hospital (approval number 151219), which waived the need for consent. Data were accessed anonymously.

Study population

We used a four-year look-back period (1996–1999) for newly identified appendectomy patients. From the hospitalization database of the LHID2005, we identified patients aged 18 to 100 years who underwent appendectomy from 2000 to 2013 (ICD-9-CM codes 47.0 and 47.1). The index date was defined as the date of appendectomy. Each identified appendectomy patient was randomly matched with four control patients according to propensity score. We excluded patients with a history of CKD before the index date, those aged <18 or >100 years, those with incomplete demographic data, and those with fewer than 30 days of follow-up.

Outcomes and relevant variables

Outcomes and comorbidities were diagnosed according to ICD-9 codes. The major outcome was a new diagnosis of CKD (ICD-9 codes 580–585), based on an outpatient diagnosis made at least three times or hospital discharge diagnosis made once during the follow-up period. ESRD was defined as undergoing dialysis therapy for at least 90 days, as in previous studies (Wu et al., 2016; Wu et al., 2017; Weng et al., 2017).

The following comorbidities potentially related to CKD were investigated: hypertension (ICD-9 401–405), diabetes mellitus (DM; ICD-9 250), hyperlipidemia (ICD-9 272), coronary artery disease (CAD; ICD-9 410–414), congestive heart failure (CHF; ICD-9 402–404, 425.4–425.9, 428), arrhythmia (ICD-9 427), stroke (ICD-9 430–438), peripheral artery occlusive disease (PAOD; ICD-9 443–444), gout (ICD-9 274), and chronic obstructive pulmonary disease (COPD; ICD-9 416.8, 416.9, 490–505, 506.4, 508, 508.1). Charlson Comorbidity Index (CCI) scores were used to measure the severity of comorbidities at baseline (Deyo, Cherkin & Ciol, 1992). Some long-term medications have been associated with renal outcomes, including antidiabetic agents, angiotensin-converting-enzyme inhibitors, angiotensin II receptor blockers, beta blockers, diuretics, nonsteroidal anti-inflammatory drugs (NSAIDs), analgesic drugs other than NSAIDs, and statins. All study subjects were followed from the index date until the first diagnosis of CKD or ESRD, withdrawal from the insurance system, or at the end of 2013.

Statistical analysis

Demographic and clinical characteristics in the appendectomy and non-appendectomy cohorts are presented as number (percentage) and mean ± standard deviation (SD). Differences in categorical and continuous variables were compared between the cohorts with chi-square tests and Student’s t-tests, respectively. To reduce the potential bias from uncontrolled confounding, we performed propensity-score-matching studies. To achieve a maximum power to illustrate a significant association between appendectomy and CKD, we used a 1:4 propensity-score-matching approach in our study. Propensity scores were calculated with multivariate logistic regression to predict the likelihood of undergoing appendectomy (Table S1). The incidence of CKD was calculated as the number of CKD events occurring during the follow-up year, divided by the total follow-up person-years for all subjects.

The Cox proportional hazards regression model with competing risks of death was applied to estimate the hazard ratio (HR) and 95% confidence interval (CI) for risk of incident CKD in patients after appendectomy compared with the non-appendectomy cohort. Multivariate Cox’s hazards analysis was used with the Fine–Grey competing risks model to estimate adjusted hazard ratios (aHRs), adjusting for the following confounders: appendectomy, demographic factors (age, sex, clinical visit frequency, and income), comorbidities (hypertension, DM, hyperlipidemia, gout, stroke, CAD, CHF, arrhythmia, PAOD, COPD, and CCI score), and long-term use of medications (statins, antidiabetic drugs, antihypertensive drugs, NSAIDs, and analgesic drugs other than NSAIDs). Subgroup analyses were used to distinguish CKD and ESRD risks in patients after appendectomy and in the control cohort according to various subpopulations. The cumulative incidence of CKD and ESRD in both study cohorts were examined with the Fine–Grey sub-distribution hazard approach and compared with Grey’s tests. All statistical analyses were performed with SAS 9.4 software (SAS Institute Inc., Cary, NC, USA). A two-tailed P value <0.05 was considered statistically significant.

Characteristics of the study population

Figure 1 shows a flowchart of the subject selection process and Table 1 shows the characteristics of the study population. This study comprised 51,915 patients, including 10,383 with new appendectomy and 41,532 propensity-score-matched controls not diagnosed with appendectomy. The follow-up times (mean ± SD) in these cohorts were 5.81 ± 3.98 years and 6.66 ± 4.02 years, respectively. There was no significant difference in sex or age between the appendectomy and non-appendectomy cohorts. The prevalence of comorbidities, including hypertension, DM, hyperlipidemia, CAD, CHF, arrhythmia, stroke, PAOD, gout, and COPD, were similar in both cohorts (Table 1). Compared with the control cohort, patients with appendectomy were slightly less likely to be taking long-term angiotensin-converting-enzyme inhibitors or angiotensin II receptor blockers (5.94% vs. 6.48%, P = 0.0433).

Long-term risk of incident CKD and ESRD

During the follow-up period, the proportion of patients who developed CKD or ESRD was significantly higher among appendectomy patients than in the control cohort (CKD: 4.34% vs. 3.44%, P < 0.0001; ESRD: 0.34% vs. 0.18%, P = 0.0019; Table 1). The cumulative incidence curve of CKD and ESRD according to the Fine–Grey method was also significantly higher for subjects with appendectomy than for the control cohort (Grey’s test: P = 0.012 for CKD, P = 0.014 for ESRD; Figs. 2A and 2B). The incidence rates of CKD and ESRD were 6.52 and 0.49, respectively, per 1,000 person-years for appendectomy patients and 5.93 and 0.31, respectively, per 1000 person-years for the control cohort (Table 2). To evaluate the reliability of our results, we used three models to adjust the risk for incident CKD and ESRD among appendectomy patients compared with the control cohort (Table 2). According to propensity-score-matched data, the risks of CKD and ESRD were significantly higher among appendectomy patients than among controls (CKD: crude HR [cHR] 1.15, 95% CI [1.03–1.27], P = 0.012; ESRD: cHR 1.65, 95% CI [1.10–2.47], P = 0.014). Even after adjustment for the propensity score, the risk of CKD and ESRD remained higher in the appendectomy group than in the control cohort (CKD: adjusted HR (aHR) 1.13, 95% CI [1.01–1.25], P = 0.029; ESRD: aHR 1.62, 95% CI [1.09–2.42], P = 0.018). Finally, we performed multivariate analysis adjusting for all confounding variables listed in Table 1. The results of this analysis showed that the risks of CKD and ESRD were similarly higher in the appendectomy group than in the control cohort (CKD: aHR 1.13, 95% CI [1.01–1.26], P = 0.037; ESRD: aHR 1.59, 95% CI [1.06–2.37], P = 0.024).

Subgroup analysis of CKD risk according to age, sex, comorbid conditions, and DM in appendectomy patients versus control cohort

Table 3 shows that the HR for incident CKD was significantly higher in the appendectomy group than in the control cohort only for males (aHR 1.16, 95% CI [1.00–1.34], P = 0.045) and for those without comorbid conditions (aHR 1.29, 95% CI [1.08–1.55], P = 0.006). The association between appendectomy and CKD risk was absent in female patients, in individual age groups, and in those with comorbid conditions (1–2 versus ≥ 3 comorbid diseases), including DM. The interactions between all stratified subgroups and appendectomy were not significant (all P_interaction > 0.05).

Subgroup analysis of ESRD risk according to age, sex, comorbid conditions, and DM in appendectomy patients versus control cohort

Table 4 shows that the HR for incident ESRD was significantly higher for appendectomy patients than for the control cohort only among patients who were middle-aged (30 −64 years) (aHR 2.05, 95% CI [1.17–3.57], P = 0.012) and that there was a marginally higher risk among females (aHR 1.68, 95% CI [0.97–2.92], P = 0.067). The association between appendectomy and ESRD risk was absent for male patients, for other age groups (<30 years and ≥ 65 years), and for those with comorbid conditions (0, 1–2, and ≥ 3 comorbid diseases). However, the interactions between sex, age group, comorbid conditions, and appendectomy were not significant (all P_interaction > 0.05). In particular, appendectomy patients with concomitant DM were found to have a significantly increased risk of ESRD compared with the control cohort patients with DM (aHR 2.08, 95% CI [1.27–3.43], P = 0.004; Table 4). The interactions between appendectomy and DM for risk of ESRD were significant (P = 0.022).

Interaction effects between appendectomy and comorbidities or medication use on CKD and ESRD risk

Interaction analyses for CKD risk are shown in Fig. 3A. Patients with appendectomy had a higher risk of CKD compared with the control cohort in most subgroups, except patients with stroke, gout, COPD, use of analgesic drugs other than NSAIDs, and use of statins. Interaction analyses for ESRD risk are shown in Fig. 3B. An increased risk of ESRD was also observed consistently in most subgroups, except those with stroke, use of NSAIDs, and use of statins. However, the interactions between appendectomy and concomitant chronic diseases or medication use were not significant for any of these subgroups (all P_interaction > 0.05; Figs. 3A and 3B).

Limitations

This study had limitations. First, the NHIRD does not include information about smoking history, body mass index, family history of renal disease, blood pressure, lipid profile, glucose or uric acid concentrations, proteinuria, dietary habits, use of over-the-counter drugs or herbal remedies, or data on chronic glomerulonephritis or chronic interstitial nephritis. Although we performed propensity-score matching and adjusted for various confounders, these unmeasured confounders might have affected our results. Second, renal outcomes were mainly identified with ICD-9-CM codes. Because CKD severity and estimated glomerular filtration rate were unknown, the number of patients with CKD and ESRD could be underestimated. However, ICD-9-CM codes are recognized as reliable indicators of CKD, ESRD, and comorbidities. Third, results derived from a retrospective cohort study are generally of lower statistical quality than those from prospective studies because of potential biases. Finally, the majority of Taiwan’s population is of Chinese ethnicity; the results of this study may not be applicable to other populations.