Relative risk of end-stage renal disease requiring dialysis in treated ankylosing spondylitis patients compared with individuals without ankylosing spondylitis: A nationwide, population-based, matched-cohort study.

OBJECTIVE: To examine the relative risk of end-stage renal disease (ESRD) requiring dialysis among treated ankylosing spondylitis (AS) patients compared with non-AS individuals.
METHODS: We used claims data from Taiwan's National Health Insurance Research Database obtained between 2003 and 2012, and enrolled 37,070 newly treated AS patients and randomly selected 370,700 non-AS individuals matched (1:10) for age, sex and year of index date. Those with a history of chronic renal failure or dialysis were excluded. After adjusting for age, sex, diabetes mellitus, hypertension, IgA nephropathy, frequency of serum creatinine examinations, use of methotrexate, sulfasalazine, ciclosporis, corticosteroid, aminoglycoside, amphotericin B, cisplatin, contrast agents and annual cumulative defined daily dose (cDDD) of traditional NSAIDs, selective cyclooxygenase-2 inhibitors (COX-2i) and preferential COX-2i, we calculated the adjusted hazard ratios (aHRs) with 95% confidence intervals using the Cox proportional hazard model to quantify the risk of ESRD in AS patients. We re-selected 6621 AS patients and 6621 non-AS subjects by further matching (1:1) for cDDDs of three groups of NSAIDs to re-estimate the aHRs for ESRD.
RESULTS: Fifty-one (0.14%) of the 37,070 AS patients and 1417 (0.38%) of the non-AS individuals developed ESRD after a follow-up of 158,846 and 1,707,757 person-years, respectively. The aHR for ESRD was 0.59 (0.42-0.81) in AS patients compared with non-AS individuals. However, after further matching for cDDD of NSAIDs, the aHR of ESRD was 1.02 (0.41-2.53). Significant risk factors included hypertension, IgA nephropathy and use of COX-2i.
CONCLUSIONS: The risk of ESRD was not significantly different between treated AS patients and non-AS individuals matched for age, sex, year of index date and dose of NSAID.

Background

Ankylosing spondylitis (AS) is a common immune-mediated inflammatory rheumatic disease affecting 0.11%–0.38% of the Taiwanese population [1, 2]. It is characterized by chronic spinal pain, stiffness, fatigue and progressive spinal ankylosis. AS patients usually experience an impaired health-related quality of life [3-7]. Other musculoskeletal manifestations include peripheral arthritis, enthesitis, and dactylitis; extra-articular manifestations may also develop.[8] The primary goal of AS management is to maximize health-related quality of life by controlling symptoms and inflammation [9]. For AS patients suffering from pain and stiffness, continuous use of non-steroidal anti-inflammatory drugs (NSAIDs) up to the maximum dose remains the first-choice medical therapy [9]. However, the potential renal toxicity of NSAIDs may limit long-term use of high-dose NSAIDs.

The incidence of end-stage renal disease (ESRD) requiring dialysis is increasing in Taiwan and in other parts of the world [10, 11]. Since 2000, Taiwan has had the highest incidence and prevalence of ESRD among the regions analyzed in the US Renal Data System [12]. Previous studies have suggested a possible association between immunoglobulin A (IgA) nephropathy and AS due to an increased prevalence of microscopic haematuria and a higher proportion of elevated serum IgA levels found in AS patients [13-16]. However, it remains unknown whether the risk of ESRD in treated AS patients is different from that in non-AS individuals.

Owing to the large number of National Health Insurance (NHI) beneficiaries in Taiwan, the NHI Research Database (NHIRD), which makes data available to researchers, is an invaluable resource for conducting longitudinal epidemiologic studies. Therefore, we analyzed nationwide claims data from NHIRD to examine the relative risk of ESRD requiring dialysis in newly treated AS patients compared with non-AS individuals.

Patients and methods

Ethics approval and consent to participate

The Institutional Review Board (IRB) of Taichung Veterans General Hospital (IRB number: CE17174B) approved this study. Informed consent could not be obtained as tracked personal information had been anonymized prior to data analysis.

Study design

We used a retrospective cohort design.

Data source

This study was conducted using claims data from NHIRD from 2003 to 2012. In 1995, Taiwan implemented a compulsory NHI program that currently covers over 99% of Taiwan’s population. The data in NHIRD includes comprehensive information on medication prescriptions, ambulatory care services, admission services and traditional medical services. Some personal and clinical data, including body weight, height, alcohol use, smoking, and data of laboratory tests, imaging, and pathology were not available in NHIRD. The Bureau of NHI (BNHI) has improved the accuracy of claims data in NHIRD by checking original medical records regularly [17]. The National Health Research Institutes managed NHIRD, and data were made available for research purposes after anonymization of personal information in accordance with privacy protocols.

Here, we utilized multiple NHIRD datasets, including 2003–2012 outpatient and inpatient claims files and enrolment files. We selected all newly treated AS patients during 2005–2012 to serve as the study cohort. The NHRI constructed a representative longitudinal health insurance database (LHID2000) of Taiwanese NHI enrollees by randomly selecting one million people who were enrolled in 2000. We selected a comparison cohort from the representative population in the LHID2000 and then extracted this cohort’s 2003–2012 claims data for analysis.

BNHI established a registry for catastrophic illness patients (RCIP) that serves as a database of patients with severe or major diseases. Patients with a certificate showing that they are on the RCIP are exempt from co-payment for all medical services related to their particular catastrophic illness. An RCIP certificate is only issued after a patient’s medical records have been carefully reviewed by at least two qualified specialists. We identified patients with ESRD requiring dialysis from the RCIP.

Definition of treated AS

Given that the NHIRD lacked data of laboratory tests and imaging to confirm the diagnosis of AS, the present study selected treated AS patients instead of individuals with AS diagnosis only as the study group to minimize misclassification bias. Treated AS patients were defined as having at least three ambulatory visits with an AS diagnosis [International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code 720.0] and concurrent prescription of NSAIDs, sulfasalazine (SSZ), methotrexate (MTX) or corticosteroid during 2003–2012. In Taiwan, AS diagnosis was based on the modified New York criteria for AS proposed in 1984 [18].

Study subjects

Newly treated AS patients identified from entire Taiwanese population

The study included all newly treated AS patients in Taiwan during 2005–2012. We excluded those with any ambulatory visits with an AS diagnosis and concurrent prescription of NSAIDs, SSZ or MTX before 1 January 2005. The index dates for treated AS patients were defined as the time of the first ambulatory visit with an AS diagnosis and concurrent prescription of NSAIDs, SSZ, MTX or corticosteroid. AS cases who had a chronic renal failure diagnosis (ICD-9-CM code 585, 586) or received dialysis before the index date were excluded.

Matched non-AS comparison group selected from a representative population of one million

Non-AS individuals were defined as having no ambulatory or inpatient AS diagnosis during 2003–2012. We randomly selected non-AS individuals from the LHID2000 matching treated AS patients (1:10) for sex, age and year of the index date (index year). In the sensitivity analysis, we re-selected the treated AS group and non-AS comparison group by additional matching (1:1) for average annual cumulative defined daily dose (cDDD) of NSAIDs. We used the time of the first ambulatory visit in the index year for any reason as the index date for the non-AS group. Those who had a chronic renal failure diagnosis (ICD-9-CM code 585, 586) or received dialysis before the index date were excluded.

Outcome

The study outcome was the time from the index date to the time of the first dialysis for ESRD. Patients who developed ESRD requiring dialysis were defined as being registered in the RCIP for chronic renal failure (ICD-9-CM code 585, 586) requiring long-term hemodialysis or peritoneal dialysis after the index date. We defined the censored date as 31 December 2012 (the last date of the data used) or the time of withdrawal from the NHI for any reason, such as leaving or death.

The risk of ESRD requiring chronic dialysis in AS patients

Incidence rate ratios (IRRs) with 95% confidence intervals (CIs) for ESRD requiring long-term dialysis were calculated in treated AS patients and compared with matched non-AS individuals. Cox proportional hazard regression was applied to calculate crude and adjusted hazard ratios (HRs) with 95% CI of ESRD requiring dialysis in AS patients compared with matched non-AS individuals.

Subgroup analysis

To test the interaction effect by age and sex on the relative risk of ESRD in treated AS patients compared with non-AS individuals, we conducted subgroup analyses of the IRRs with 95% CIs and adjusted HRs with 95% CIs for ESRD requiring dialysis were conducted based on age (≤40 years, >40 years) and sex.

Potential confounders

Potential confounders included baseline age, sex, comorbidities within one year before the index date, a history of IgA nephropathy, the frequency of testing serum creatinine (i.e. number/year = 0, 0 < number/year <1, number/year ≥1) and average annual cDDD of NSAIDs[19] during the follow-up period. Comorbidities included diabetes mellitus (DM) requiring anti-diabetic treatment, hypertension requiring anti-hypertensive treatment and IgA nephropathy. DM and hypertension were also included as confounders as they are both known risk factors for ESRD.[20-24] DM was defined as having at least one ambulatory visit or hospitalization that resulted in a DM diagnosis (ICD-9-CM code 250.x) with a concurrent prescription of any anti-diabetic drugs within one year before the index date. Hypertension was defined as having at least one outpatient visit or hospitalization with a hypertension diagnosis (ICD-9-CM codes 401–405) and a concurrent prescription of any anti-hypertensive agent within one year before the index date. A history of IgA nephropathy was defined as having at least three outpatient visits or one admission with an ICD-9-DM code 583.9 diagnosis before the index date. During the follow-up period, we also calculated the average annual number of cDDDs of NSAIDs [19] for adjustment and further matching. NSAIDs were categorized into three groups: traditional NSAIDs, selective cyclooxygenase-2 inhibitors (COX-2i) and preferential COX-2i (Table A in S1 Table). The average annual number of cDDDs of traditional NSAIDs were transformed to the categorical variable based on the quartiles in all subjects. The average annual numbers of cDDDs of selective COX-2i and preferential COX-2i were switched to categorical variables based on 50th percentile. NSAIDs were changed to categorical variables based on the quartiles in all subjects. We also adjusted the use of disease-modifying antirheumatic drugs (including MTX, SSZ and ciclosporin), corticosteroid, and other nephrotoxic agents (including aminoglycoside, amphotericin B, cisplatin and contrast agents).

Sensitivity analysis

Because NSAID is a major confounding factor, it is possible to have an inconsistent result if NSAID was matched rather than adjusted. We therefore conducted sensitivity analyses of the HRs with 95% CIs for risk of ESRD requiring dialysis by re-selecting AS cases and non-AS individuals after matching (1:1) for cDDDs of the three groups of NSAIDs.

Statistical analysis

Continuous variables are presented as a mean ± standard deviation and categorical variables as a percentage of patients. We examined the differences in continuous variables by Student’s t-test and categorical variables by Pearson’s χ2 test. We quantified the associations between covariates and the risk of ESRD requiring dialysis using Cox proportional regression analysis to estimate HRs with 95% CIs after adjusting for potential confounders. A log-rank test was used to examine the difference of cumulative incidence of ESRD requiring dialysis between AS patients and matched non-AS individuals. A two-tailed p-value < 0.05 was considered statistically significant. The significance of interaction effect by age group or gender on treated AS-associated risk of ESRD requiring dialysis was examined by calculating the p-value of the coefficient associated with the product of age group or gender and the indicator of treated AS using the Wald test. We performed all statistical analyses by SAS statistical software, version 9.3 (SAS Institute, Inc., Cary, NC, USA).

Results

We identified 37,070 newly treated AS cases and randomly selected 370,700 non-AS individuals matched with AS cases, in a 1:10 ratio, for age, sex and the year of the initial AS-related treatment date (index date). The mean age ± SD was 42.3 ± 16.7 years, and 63.1% of study subjects were men (Table 1). AS patients had significantly higher proportions of DM, hypertension and IgA nephropathy than non-AS individuals.

Table 1

Demographic data and clinical characteristics of 37,070 AS cases and 370,700 non-AS individuals matched for age, sex and year of index date.

	Non- AS	AS	p-value
	(n = 370,700)	(n = 37,070)
Age, years, mean ± SD	42.3 ± 16.7	42.3 ± 16.7	1.000
Sex
Female	136,740 (36.9)	13,674 (36.9)
Male	233,960 (63.1)	23,396 (63.1)	1.000
Comorbidities
Diabetes mellitus	18,337 (5.0)	1,995 (5.4)	<0.001
Hypertension	45,710 (12.3)	6,156 (16.6)	<0.001
IgA nephropathy	484 (0.1)	89 (0.2)	<0.001
Frequency of serum creatinine examinations during the follow-up period	0.5 ± 2.7	1.7 ± 2.4	<0.001
Frequency group			<0.001
Number/year = 0	212,540 (57.3)	5799 (15.6)
0 < number/year < 1	102,148 (27.6)	13,098 (35.3)
Number/year ≥ 1	56,012 (15.1)	1,8173 (49.0)
Medications
NSAIDs			<0.001
Never used	44,278 (11.9)	127 (0.3)
Ever used	326,422 (88.1)	36,943 (99.7)
Traditional NSAIDs			<0.001
cDDD/year ≤2	100,562 (27.1)	3,890 (10.5)
2 <cDDD/year≤6	99,791 (26.9)	4,170 (11.2)
6 <cDDD/year≤14	93,241 (25.2)	7,218 (19.5)
cDDD/year >14	77,106 (20.8)	21,792 (58.8)
Selective COX-2i			<0.001
cDDD/year ≤8	358,990 (96.8)	18,721 (50.5)
cDDD/year >8	11,710 (3.2)	18,349 (49.5)
Preferential COX-2i			<0.001
cDDD/year ≤2	343,151 (92.6)	18,427 (49.7)
cDDD/year >2	27,549 (7.4)	18,643 (50.3)
Methotrexate use	1,462 (0.4)	3,791 (10.2)	<0.001
Sulfasalazine use	1,777 (0.3)	21,901 (59.1)	<0.001
Ciclosporin	299 (0.1)	475 (1.3)	<0.001
Corticosteroid use	155,502 (42.0)	22,645 (61.1)	<0.001
Aminoglycoside	3,261 (0.9)	351 (1.0)	0.188
Amphotericin B	137 (0.04)	22 (0.1)	0.037
Cisplatin	1,456 (0.4)	152 (0.4)	0.613
Contrast agents	12,792 (3.5)	2,928 (7.9)	<0.001

Results are shown as number (%) unless specified otherwise.

Abbreviations: AS, ankylosing spondylitis; HR, hazard ratio; NSAIDs, non-steroidal anti-inflammatory drugs; cDDD, cumulative defined daily dose; COX-2i, cyclooxygenase-2 inhibitors.

Table 2 shows a comparison of the incidence rates of ESRD requiring dialysis between AS cases and non-AS individuals. Fifty-one (0.14%) of the 37,070 AS patients and 1417 (0.38%) of the non-AS individuals developed ESRD requiring dialysis. The incidence rate of ESRD was significantly lower in treated AS cases compared with non-AS individuals, and this finding was consistent across all age and sex subgroups, except the >40 years age group (Table 2). Fig 1 shows the cumulative incidence of ESRD requiring dialysis among AS patients and non-AS individuals matched by age, sex and index date (log rank test p < 0.001). As shown in Table B in S1 Table, the risk of ESRD requiring dialysis was significantly lower in treated AS patients compared with non-AS individuals after adjustment for potential confounders (HR, 0.59; 95% CI, 0.42–0.81). As shown in Table C in S1 Table, age and sex did not have interaction effects.

Table 2

Comparison of the incidence rates of end-stage renal disease requiring dialysis between treated AS patients and matched non-AS individuals.

Group	Total	Event (%)	Total person-years	IR (/105 years)	IRR (95% CI)
All subjects
Non-AS	370,700	1,417 (0.38)	1,707,757	83	1.00
Treated AS	37,070	51 (0.14)	158,846	32	0.39 (0.29–0.51)
Age ≤ 40 years
Non-AS	185,000	128 (0.07)	846,739	15	1.00
Treated AS	18,500	9 (0.05)	80,015	11	0.74 (0.38–1.46)
Age > 40 years
Non-AS	185,700	1,289 (0.69)	861,018	150	1.00
Treated AS	18,570	42 (0.23)	78,830	53	0.36 (0.26–0.48)
Female
Non-AS	136,740	572 (0.42)	628,187	91	1.00
Treated AS	13,674	19 (0.14)	57,977	33	0.36 (0.23–0.57)
Male
Non-AS	233,960	845 (0.36)	1,079,571	78	1.00
Treated AS	23,396	32 (0.14)	100,869	32	0.41 (0.28–0.58)

Matched variables include age, sex and year of the index date.

Abbreviations: AS, ankylosing spondylitis; IR, incidence rate; IRR, incidence rate ratio; CI, confidence interval.

Fig 1

The cumulative incidences of ESRD requiring dialysis among 37,070 AS patients and 370,700 non-AS individuals matched for age, sex and year of index date.

Given that use of NSAIDs is one of the most critical risk factors for the development of ESRD [25], we re-selected 6,621 AS patients and 6,621 non-AS subjects by further matching (1:1) for cDDDs of three groups of NSAIDs to estimate the risk of developing ESRD requiring dialysis associated with AS. Table 3 shows a comparison of the demographic and clinical data from both groups. The frequency of serum creatinine examination was higher in AS patients than in non-AS individuals. The proportions of comorbid hypertension and use of MTX, SSZ and ciclosporin were significantly higher in the treated AS group than in the non-AS group. However, the proportion of DM patients was not different between the treated AS group and the non-AS group. The proportions of aminoglycoside use and cisplatin use were lower in the treated AS group than in the non-AS comparison group. Table 4 shows that the IRR of ESRD requiring dialysis was not significantly different between AS patients and non-AS individuals, and this finding was consistent across all subgroups stratified by age and sex. As shown in Table 5, after adjusting for potential confounders, the risk of ESRD requiring dialysis was still not significantly different between the treated AS group and the non-AS group (HR, 1.02; 95% CI, 0.41–2.53). Significant risk factors included hypertension, IgA nephropathy, the frequency of serum creatinine follow-up ≥1, and use of selective COX-2i. Fig 2 shows the cumulative incidence of ESRD requiring dialysis among treated AS patients and non-AS individuals matched for age, sex, index date and NSAIDs dose (log rank test p < 0.808). As shown in Table D in S1 Table, age and sex did not have interaction effects.

Table 3

Demographic data and clinical characteristics of 6,621 treated AS patients and 6,621 non-AS individuals matched for age, sex, year of the index date and average annual numbers of cDDD of NSAIDs.

	Non-AS	AS	p-value
	(n = 6,621)	(n = 6,621)
Age, mean ± SD years	40 ± 14	40 ± 14	1.000
Sex			1.000
Female	2,129 (32.2)	2,129 (32.2)
Male	4,492 (67.8)	4,492 (67.8)
Comorbidities
Diabetes mellitus	309 (4.7)	287 (4.3)	0.356
Hypertension	722 (10.9)	823 (12.4)	0.006
IgA nephropathy	10 (0.15)	17 (0.26)	0.178
Frequency of serum creatinine examination during the follow-up period, mean ± SD number per year	0.5 ± 1.0	0.7 ± 1.4	<0.001
Frequency group
Number/year = 0	2,929 (44.2)	2,001 (30.2)
0 < number/year < 1	2,584 (39.0)	3,196 (48.3)
Number/year ≥ 1	1,108 (16.7)	1,424 (21.5)
Medications
NSAIDs			0.436
Never used	115 (1.7)	127 (1.9)
Ever used	6,506 (98.3)	6,494 (98.1)
Traditional NSAIDs, mean ± SD cDDD/year group	22.8 ± 20.8	23.0 ± 20.8	0.696
cDDD/year ≤ 10	1,720 (26.0)	1,664 (25.1)	0.645
10 < cDDD /year ≤ 18	1,585 (23.9)	1,598 (24.1)
18<cDDD /year ≤ 30	1,630 (24.6)	1,676 (25.3)
cDDD > 30	1,686 (25.5)	1,683 (25.4)
Selective COX-2i, mean ± SD cDDD/year	0.3 ± 3.4	0.4 ± 3.4	0.775
cDDD = 0	6279 (94.8)	6220 (93.9)	0.026
cDDD > 0	342 (5.2)	401 (6.1)
Preferential COX-2i, mean ± SD cDDD/year	0.7 ± 2.9	0.8 ± 2.9	0.194
cDDD = 0	5,394 (81.5)	5,280 (79.8)	0.012
cDDD > 0	1,227 (18.5)	1,341 (20.3)
Methotrexate use	35 (0.5)	266 (4.0)	<0.001
Sulfasalazine use	25 (0.4)	2873 (43.4)	<0.001
Ciclosporin use	5 (0.1)	37 (0.6)	0.005
Corticosteroid use	3,902 (58.9)	3,838 (58.0)	0.259
Aminoglycoside	65 (1.0)	37 (0.6)	0.005
Amphotericin B	1 (0.02)	4 (0.06)	0.189
Cisplatin	49 (0.7)	25 (0.4)	0.005
Contrast agents	326 (4.9)	364 (5.5)	0.137

Results are shown as number (%) unless specified otherwise.

Abbreviations: AS, ankylosing spondylitis; HR, hazard ratio; NSAIDs, non-steroidal anti-inflammatory drug; cDDD, cumulative defined daily dose; COX-2i, cyclooxygenase-2 inhibitors.

Table 4

Comparison of the incidence rates of end-stage renal disease requiring dialysis between 6,621 AS patients and 6,621 matched non-AS individuals.

Group	Total	Event (%)	Total person-years	IR (/105 years)	IRR (95% CI)
All subjects
Non-AS	6621	13 (0.20)	35,396	37	1.00
Treated AS	6621	11 (0.17)	33,272	33	0.90 (0.40–2.01)
Age ≤ 40 years
Non-AS	3489	3 (0.09)	18,725	16	1.00
Treated AS	3489	3 (0.09)	17,751	17	1.05 (0.21–5.23)
Age > 40 years
Non-AS	3132	10 (0.32)	16,671	60	1.00
Treated AS	3132	8 (0.26)	15,522	52	0.86 (0.34–2.18)
Female
Non-AS	2129	4 (0.19)	11,520	35	1.00
Treated AS	2129	4 (0.19)	10,785	37	1.07 (0.27–4.27)
Male
Non-AS	4492	9 (0.20)	23,875	38	1.00
Treated AS	4492	7 (0.16)	22,487	31	0.83 (0.31–2.22)

Matched variables include age, sex, year of the index date and average annual cumulative defined daily doses of three groups of non-steroidal anti-inflammatory drugs during the follow-up period.

Abbreviations: AS, ankylosing spondylitis; IR, incidence rate; IRR, incidence rate ratio; CI, confidence interval.

Table 5

Crude and multivariable-adjusted analyses of the risk of end-stage renal disease requiring dialysis associated with variables among 6,621 AS patients and 6,621 matched non-AS individuals, as shown by HRs with 95% CIs.

Variable	Crude	Adjusted
	HR (95%CI)	HR (95%CI)
AS
Non-AS	Reference	Reference
Treated AS	0.91 (0.41–2.02)	1.02 (0.41–2.53)
Comorbidities
Diabetes mellitus
No	Reference	Reference
Yes	9.11 (3.78–21.98)	1.19 (0.46–3.09)
Hypertension
No	Reference	Reference
Yes	18.55 (7.69–44.75)	6.86 (2.39–19.70)
IgA nephropathy
No	Reference	Reference
Yes	57.75 (13.56–246.04)	14.05 (2.91–67.95)
Frequency of serum creatinine examination during the follow-up period
Number/year = 0	Reference	Reference
0 < number/year < 1	1.37 (0.12–15.12)	1.20 (0.11–13.56)
Number/year ≥ 1	42.54 (5.72–316.26)	9.99 (2.38–167.99)
Medication use vs. not use
Methotrexate	4.02 (0.95–17.09)	4.13 (0.86–19.92)
Sulfasalazine	0.55 (0.17–1.86)	0.44 (0.11–1.75)
Ciclosporin*	-	-
Corticosteroid	1.39 (0.57–3.35)	1.41 (0.55–3.62)
Aminoglycoside*	4.56 (3.57–5.83)	1.30 (1.01–1.66)
Amphotericin B*	-	-
Cisplatin*	-	-
Contrast agents	1.44 (0.34–6.12)	0.61 (0.14–2.75)

Matched variables included age, sex, year of the index date and average annual number of cumulative defined daily dose of non-steroidal anti-inflammatory drugs. Adjusted variables included diabetes, hypertension, IgA nephropathy, frequency of serum creatinine examinations during the follow-up period, use of methotrexate, sulfasalazine, ciclosporin aminoglycoside, amphotericin B, cisplatin and contrast agents.

Abbreviations: AS, ankylosing spondylitis; HR, hazard ratio; CI, confidence interval; NSAIDs, non-steroidal anti-inflammatory drugs; cDDD, cumulative defined daily dose; COX-2i, cyclooxygenase-2 inhibitors.

*None of users developed end-stage renal disease requiring dialysis.

Fig 2

The cumulative incidences of ESRD requiring dialysis among 6,621 AS patients and 6,621 non-AS individuals matched for age, sex, index date and NSAID dose.

Discussion

To the best of our knowledge, this is the first nationwide, population-based cohor study to estimate the relative risk of ESRD requiring dialysis in AS patients receiving medical therapy compared with matched non-AS individuals. We found that treated AS patients had a lower risk of ESRD requiring dialysis than non-AS individuals matched for age, sex and year of the index date. However, this finding was not robust after additionally matching (1:1) for the doses of three groups of NSAIDs. NSAIDs may increase the risk of renal function impairment; therefore, patients with the impaired renal function may have been excluded from this study. NSAIDs are the first-line medication recommended for symptomatic AS patients [9]. Given that 99.7% of AS patients and 88% of matched non-AS individuals received NSAIDs therapy (p < 0.001) and the doses of NSAIDs were markedly higher in treated AS patients than in non-AS individuals, adjustment for NSAID dose may not have eliminated its confounding effect. Samia et al. reported a high prevalence of renal disease (15.1%) and ESRD (3.3%) in 212 AS cases, with a mean follow-up of 12 years in hospital [16]. However, hospital-based data are subject to selection bias. In the aforementioned study, the lack of a comparison cohort meant it was not possible to conclude that AS patients had an increased risk of ESRD [16]. Kang et al. found that the prevalence of renal failure in AS patients was not different from that in matched non-AS controls in Taiwan (0.7% vs. 0.6%, p = 0.421) [26]. However, there were some differences between Kang et al.’s study and ours. First, their study lacked information on medication; thus, non-treated AS cases may have been enrolled [26]. The validity of AS diagnosis may be of concern to those who have never received NSAID therapy, given that NSAIDs are the first-line treatment for symptomatic AS patients. Second, based on their results, it is not possible to determine whether AS per se influences the risk of renal failure because the use of medication, especially NSAIDs, may be unequally distributed between AS cases and non-AS controls [26]. Third, incident AS cases were excluded and the prevalence and incidence of dialysis due to ESRD in AS patients were not estimated.

Consistent with previous studies [20-24], we found that hypertension increased the risk of ESRD. A higher frequency of serum creatinine examination was associated with an increased risk of ESRD. This finding might be explained by reverse causality (i.e., impaired baseline renal function leading to more frequent follow-up). Consistent with a study by Chang et al. [25], we found that use of selective COX-2 inhibitors was associated with an increased risk of ESRD. However, use of preferential COX-2i tended to be associated with a lower risk of ESRD requiring dialysis (HR, 0.37; 95% CI, 0.12–1.09; p-value = 0.071). Previously reported studies might explain this finding. In 1996, an open study reported that the preferential COX-2i, meloxicam, did not influence renal function or lead to meloxicam accumulation in 25 patients with rheumatic diseases with mild renal function impairment [27]. In 1997, another study demonstrated that individuals with moderate renal function impairment had lower meloxicam concentrations in plasma with corresponding higher plasma clearance compared with those with normal renal function, suggesting that there is no need to adjust meloxicam dosage in patients with mild-to-moderate renal impairment [28]. An animal study showed that meloxicam had a renal protective effect in diabetic rats by reducing COX-2 expression in the kidney [29]. An earlier animal study showed that preferential COX-2i might compromise renal perfusion [30]. However, future prospective randomized controlled studies are warranted to confirm the possible protective effect of preferential COX-2i against ESRD.

The main strength of this study was the use of a nationwide population-based cohort, which provided a large sample size and avoided selection bias. However, this study has some limitations. First, the analysis was based on claims data; therefore, it was not possible to be absolutely certain that the AS diagnosis was accurate. However, BNHI has increased the accuracy of diagnosis and requires a routine check of the original medical record [17]. The exclusion of patients who did not receive AS-related medical therapy may also have improved the accuracy of AS diagnosis. Second, as there may have been a long period between symptom onset and AS diagnosis (8–11 years), some early AS patients may have been misclassified as non-AS individuals. Mild AS patients may not have visited a physician for a long period [31], and thus may have been misclassified as non-AS individuals. Third, although we adjusted for the frequency of serum creatinine examination, we could not completely avoid detection bias. However, such bias should have led to an overestimation of ESRD risk in the AS cases. Based on our findings, it is likely that AS did not increase the risk of ESRD. Fourth, data regarding some potential confounding factors, such as the use of tobacco, alcohol, over-the-counter medications and traditional herbal medicines, were not collected in NHIRD. Fifth, the lack of some clinical data, such as serum creatinine level, urine routine, human leukocyte antigen-B27, imaging findings and pathologic data limited further adjustment or matching to confirm our findings. Finally, we cannot generalize these results to AS patients who did not receive pharmacological therapy or to non-Taiwanese populations.

Conclusions

This nationwide, population-based cohort study revealed that the risk of ESRD requiring long-term-dialysis in treated AS patients was not significanly different from that in non-AS individuals. Further large, population-based studies using renal function data are warranted to confirm the lack of association between treated AS and risk of ESRD.

Data of 37,070 AS patients and 370,700 non-AS individuals matched for age, sex and year of index date.

(SAV)

Click here for additional data file.

6,621 AS patients and 6,621 non-AS individuals matched for age, sex, index date and NSAID dose.

(SAV)

Click here for additional data file.

Table A and Table B.

Additional data.

(DOCX)

Click here for additional data file.

11 Feb 2020

PONE-D-19-23326

Risk of End-Stage Renal Disease Requiring Dialysis in Treated Ankylosing Spondylitis: a Nationwide, Population-based, Matched-cohort Study

PLOS ONE

Dear Dr. Chen,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

It will be important to make it clear, as per reviewer 1's comments, whether this was an analysis to examine the association of Ankylosing Spondylitis with ESKD, or examine the association between NSAID use and ESKD, and describe to what extent these can be seperated.

We would appreciate receiving your revised manuscript by Mar 22 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Nicholas M Selby, BMedSci BMBS MRCP DM

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements:

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.plosone.org/attachments/PLOSOne_formatting_sample_main_body.pdf and http://www.plosone.org/attachments/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Thank you for stating in your Funding Statement:

"This study was supported in part by grant TCVGH-1067313C from Taichung Veterans General Hospital, Taiwan. HHC received partial funding from this grant. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript."

Please provide an amended statement that declares *all* the funding or sources of support (whether external or internal to your organization) received during this study, as detailed online in our guide for authors at http://journals.plos.org/plosone/s/submit-now.  Please also include the statement “There was no additional external funding received for this study.” in your updated Funding Statement.

Please include your amended Funding Statement within your cover letter. We will change the online submission form on your behalf.

3. Your ethics statement must appear in the Methods section of your manuscript. If your ethics statement is written in any section besides the Methods, please move it to the Methods section and delete it from any other section. Please also ensure that your ethics statement is included in your manuscript, as the ethics section of your online submission will not be published alongside your manuscript.

4. We note that there appears to be an issue with your uploaded figures 1 and 2 and they are instead opening with an error message only. Can you please upload replacement Figures 1 and 2 and ensure the figures follow the guidelines (https://journals.plos.org/plosone/s/figures)

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Dear authors,

Thank you for your efforts on this manuscript. I have some comments for your consideration

Major comments:

1- I would like you to clearly state the aim of this study. Do you want to test the association between AS and development of ESRD or do you want to test the association between the intake of treatment among AS patients and development of ESRD. As you did in this work, we included AS treated patients and non-As population and then assessed the development of ESRD, which means that your assumption is drug intake among AS patients predispose to development of ESRD. Now in this case, you also need to consider AS patients without intake of treatment as a control group.

2- We usually consider a ration of 1:4 as a maximum between cases: controls or exposed versus non exposed

3- You mentioned in the abstract that you matched for age, sex, index date, and NSAIDs use while in the text NSAIDs was not mentioned

4- I don't get your justification for doing propensity score analysis in this manuscript. Usually we do PS to overcome the problem of selection bias or confounding by indication in observational studies so as to be similar to RCTs. Your manuscript considered AS who were receiving treatment from the beginning versus non-AS population, so what is the benefit of PS analysis

5- IgA nephropathy is considered as renal disease so why did not you exclude patients with previous history of IgA nephropathy

6- You considered age and sex as potential confounders although you matched for them from the beginning

7- You should consider PS in your cox analysis later

8- What is your rational for the moderate-to severe renal disease classification. I would prefer to report if they have history of CKD or not

9- Why did you adopt the sensitivity analysis

10- How did you deal with loss to follow up

11- what about the history of other drugs which may predisposing to renal impairments

12- What is the importance of presenting the data by age < 40 and > 40 years and also by gender although you already matched for it

13- why did you adjust for age and sex again in cox analysis? this leads to over-adjustment

14- There is no need to report matched factors in the tables

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

13 Mar 2020

Review’s Comments to the Author:

Major comments:

1. I would like you to clearly state the aim of this study. Do you want to test the association between AS and development of ESRD or do you want to test the association between the intake of treatment among AS patients and development of ESRD. As you did in this work, we included AS treated patients and non-As population and then assessed the development of ESRD, which means that your assumption is drug intake among AS patients predispose to development of ESRD. Now in this case, you also need to consider AS patients without intake of treatment as a control group.

Author Response: Thank you for your great comment.

(1) We wanted to test the association between AS and the development of ESRD. Because NSAIDs are the most important confounding factor in the relationship between AS and ESRD, we first adjusted the dose of NSAIDs in the Cox regression analysis and then conducted a sensitivity analysis by matching the dose of NSAIDs. We selected treated AS patients instead of patients with AS diagnosis only because the NHIRD lacked data on laboratory and imaging examinations to confirm the diagnosis of AS: some patients might be diagnosed with AS temporarily before the relevant data were available. Therefore, we selected treated AS patients instead of individuals with AS diagnosis only as the study group to minimize misclassification bias. We did not include patients with AS diagnosis but not receiving AS-related treatment because these patients may be either real AS patients or non-AS individuals. Using ‘treated AS’ instead of ‘AS’ is to clarify that the study population only included AS patients receiving treatment, but not untreated AS patients.

(2) To avoid confusion, we revised the title as follows: ‘Relative Risk of End-Stage Renal Disease Requiring Dialysis in Treated Ankylosing Spondylitis Patients Compared with Individuals without Ankylosing Spondylitis: a Nationwide, Population-based, Matched-cohort Study’. We also added a statement in the subsection of ‘Definition of treated AS’ in the Methods section: ‘Given that the NHIRD lacked data on laboratory tests and imaging to confirm the diagnosis of AS, the present study selected treated AS patients instead of individuals with AS diagnosis only as the study group to minimize misclassification bias.’

2. We usually consider a ration of 1:4 as a maximum between cases: controls or exposed versus non exposed

Author Response:

(1) Miettinen (1969) and others have shown that there is little to gain by using a matching ratio >4 (i.e., the ‘diminishing returns’ phenomenon once the matching ratio exceeds 4). (David GK, Lawrence LK, Hal M. Epidemiologic research. In: John Wiley & Sons, ed. New York, 1982:396). However, a matching ratio >4 is still valid. The reason for choosing 10 as the matching ratio to select the comparison group in the initial analysis was to provide more cases in the re-selected comparison group by further matching for the cDDD of NSAIDs in the sensitivity analysis.

2- You mentioned in the abstract that you matched for age, sex, index date, and NSAIDs use while in the text NSAIDs was not mentioned

Author Response:

(1) We mentioned ‘NSAID use’ as another matching variable in the sensitivity analysis in the subsection ‘Matched non-AS comparison group selected from a representative population of one million’ in the Methods section: ‘We randomly selected non-AS individuals from the LHID2000 matching treated AS patients (1:10) for sex, age, and year of the index date (index year). In the sensitivity analysis, we re-selected the treated AS group and non-AS comparison group by additional matching (1:1) for the average annual cumulative defined daily dose (cDDD) of NSAIDs.’

4- I don't get your justification for doing propensity score analysis in this manuscript. Usually we do PS to overcome the problem of selection bias or confounding by indication in observational studies so as to be similar to RCTs. Your manuscript considered AS who were receiving treatment from the beginning versus non-AS population, so what is the benefit of PS analysis

Author response:

(1) We apologize for the error in the statement regarding the matching method. In fact, we matched the treated AS group and the non-AS comparison group directly for sex, age, and year of the index date, with additional matching for the cDDD of NSAIDs in the sensitivity analysis.

(2) We revised the statement as follows: ‘We randomly selected non-AS individuals from the LHID2000 matching-treated AS patients (1:10) for sex, age, and year of the index date (index year). In the sensitivity analysis, we re-selected the treated AS group and non-AS comparison group by additional matching (1:1) for the average annual cumulative defined daily dose (cDDD) of NSAIDs.’

5- IgA nephropathy is considered as renal disease so why did not you exclude patients with previous history of IgA nephropathy

Author response:

(1) We did not exclude patients with a previous history of IgA nephropathy, nor did we mentioned it in the original manuscript. However, we did not add IgA nephropathy in the covariates initially. Therefore, we revised our analysis by adding IgA nephropathy in the adjusted variable and revised the data as shown in the Table and in the text.

6- You considered age and sex as potential confounders although you matched for them from the beginning

Author response:

(1) Sjölander and Greenland investigated the trade-off between bias and variance in deciding whether adjustment for matching variables is advisable (Stat Med 2013; 32(27):4696-708.) and reviewed the validity of matching variables. On page 4701 of this paper, they concluded that it is usually not valid to ignore the matching variables when adjusting for additional confounders. Of note, it is valid to ignore the matching variables when adjusting for additional covariates if any of the following criteria hold true in the target population: i. The unmatched covariates are conditionally independent of the exposure, given the matching variables. ii. The unmatched covariates are conditionally independent of the matching variables, given the exposure. iii. The outcome is conditionally independent of the matching variables, given the exposure and unmatched covariates.

(2) Given that, in our study, we adjusted for additional covariates and none of the aforementioned criteria were met, it is reasonable to adjust matching variables including sex, age, and cDDD of NSAIDs.

(3) However, for fear of confusion, we did not show crude and adjusted HR (95% CI) for the matching variables in our revised manuscript.

7- You should consider PS in your cox analysis later

Author response:

(1) We did not use PS for matching or adjustment in the study.

8- What is your rational for the moderate-to severe renal disease classification. I would prefer to report if they have history of CKD or not.

Author response:

(1) We excluded those who had a history of CKD (ICD-9-CM code 585) or renal failure (ICD-9-CM code 586), given that they are at high risk of ESRD requiring dialysis during a short period of follow-up if they are treated with NSAIDs, the main treatment for AS. Because the proportions of a history of moderate-to-severe renal disease were not different between treated AS patients and non-AS individuals, for fear of confusion, we have removed the moderate-to-severe renal disease from the list of covariates.

(2) Also, given that IgA nephropathy was considered to be a risk factor of ESRD and treated AS patients had a higher proportion of having a history of IgA nephropathy than non-AS individuals, we considered IgA nephropathy as a potential confounder. We added ‘a history of IgA nephropathy’ in the list of potential confounders (2nd line in the subsection ‘Potential confounders’ in the ‘Methods’ section) and also added a statement in Line 11-13 in this subsection: ‘A history of IgA nephropathy was defined as having at least three outpatient visits or one admission with an ICD-9-DM code 583.9 diagnosis before the index date.’

9- Why did you adopt the sensitivity analysis

Author response:

(1) Because NSAID use is a major confounding factor, it is possible to obtain an inconsistent result if NSAID use was matched rather than adjusted. Also, the initial analysis performed by adjusting the dose of NSAIDs showed that treated AS patients seemed to have a lower risk of ESRD requiring dialysis than non-AS individuals.

(2) However, given that the proportion of receiving NSAID treatment is significantly higher than that of non-AS individuals, we cannot exclude the possibility that the result might be biased due to confounding by indication if the dose of NSAIDs was adjusted rather than matched.

10- How did you deal with loss to follow up

Author response:

(1) Because the present study used claim data, we defined the censored date as 31 December 2012 (the last date of the data used) or the time of withdrawal from the NHI for any reason, such as leaving or death (see the subsection ‘Outcome’ in the ‘Methods’ section).

11- what about the history of other drugs which may predisposing to renal impairments

Author response:

(1) We also considered other drugs with potential nephrotoxicity, including the use of ciclosporin, aminoglycosides, amphotericin B, cisplatin, and contrast agents as potential confounders in the revised manuscript.

12- What is the importance of presenting the data by age < 40 and > 40 years and also by gender although you already matched for it

Author response:

(1) Regarding the rationale for performing a subgroup (stratified) analysis (David GK, Lawrence LK, Hal M. Epidemiologic research. In: John Wiley & Sons, ed. New York, 1982:387-8), the matching process itself ensured that the crude risk ratio for the matched data will provide the correct point estimate of the population relative risk (RR), given no effect modification assumption. However, if the matched variables have a modification effect, the point estimate of RR will be different between the subgroups. For example, the point estimate of RR may be 1 in the whole study population (combined males and females), but <1 in males and >1 in females. Also, a stratified analysis is required for both precision and validity reasons.

(2) We therefore revised the subsection of ‘Subgroup analysis’ in the Methods section: ‘To test the interaction effect by age and sex on the relative risk of ESRD in treated AS patients compared with non-AS individuals, we conducted subgroup analyses of the IRRs with 95% CIs and adjusted HRs with 95% CIs for ESRD requiring dialysis were conducted based on age (≤40 years, >40 years) and sex.’

(3) We described the method of statistical analysis for testing the modification effect in the ‘Statistical analysis’ subsection: ‘The significance of the interaction effect by age group or gender on the treated AS-associated risk of ESRD requiring dialysis was examined by calculating the P value of the coefficient associated with the product of age group or gender and the indicator of treated AS using the Wald test.’

(4) The cut-off for the age group (40 years) was based on the mean age of the population in the sensitivity analysis.

13- why did you adjust for age and sex again in cox analysis? this leads to over-adjustment

Author response: Please refer to the response to question 6.

14- There is no need to report matched factors in the tables

Author response: We have deleted the data on matched factors in the tables.

Submitted filename: Response-Letter-AS_and_ESRD_risk-20200312.docx

Click here for additional data file.

25 Mar 2020

Relative Risk of End-StageRenal Disease Requiring Dialysis in Treated Ankylosing Spondylitis Patients Compared with Individuals without Ankylosing Spondylitis: a Nationwide, Population-based, Matched-cohort Study

PONE-D-19-23326R1

Dear Dr. Chen,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Nicholas M Selby, BMedSci BMBS MRCP DM

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Thanks for your efforts in modifying the manuscript and justifying all comments in an intelligible way

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

6 Apr 2020

PONE-D-19-23326R1

Relative Risk of End-Stage Renal Disease Requiring Dialysis in Treated Ankylosing Spondylitis Patients Compared with Individuals without Ankylosing Spondylitis: a Nationwide, Population-based, Matched-cohort Study

Dear Dr. Chen:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Nicholas M Selby

Academic Editor

PLOS ONE