Association between atrial fibrillation and risk of end-stage renal disease among adults with diabetes mellitus.

Diabetes mellitus (DM) is an important risk factor in patients with end-stage renal disease (ESRD). DM is associated with the development of cardiovascular diseases, such as atrial fibrillation (AF), due to poor glycemic control. However, few studies have focused on the risk of developing ESRD among DM patients with and without AF. This study evaluated ESRD risk among DM patients with and without AF in Taiwan. Data were retrieved from one million patients randomly sampled from Taiwan's National Health Insurance Research Database, including 6,105 DM patients with AF propensity score-matched with 6,105 DM patients without AF. Both groups were followed until death, any dialysis treatment, or December 31, 2013, whichever occurred first. AF was diagnosed by a qualified physician according to the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM), using the diagnostic code 427.31. Patients aged <20 years or diagnosed with ESRD before the index date were excluded. A Cox proportional hazard regression model was used to calculate the relative ESRD risk. Among DM patients, those with AF have more comorbidities than those without AF. We also found a 1.18-fold (95% confidence interval [CI]: 1.01-1.46) increase in ESRD risk among patients with AF compared with those without AF. In addition, DM patients with hypertension, chronic kidney disease (CKD), or higher Charlson Comorbidity Index scores also have significantly increased ESRD risks than those without these complications. A 1.39-fold (95% CI: 1.04-1.86) increase in risk was observed for patients with AF among the non-CKD group. Our findings suggest that patients with DM should be closely monitored for irregular or rapid heart rates.

Introduction

Diabetes mellitus (DM) is a major public health priority worldwide, and the global prevalence of DM is projected to increase from 463 million (9.3%) in 2019 to 700 million (10.9%) in 2045 [1]. Patients with DM have high risks of cardiovascular morbidity and mortality [2], and DM is a leading cause of end-stage renal disease (ESRD), as approximately 40% of all patients with ESRD also have DM [3]. According to the United States Renal Data System annual report, Taiwan has the highest ESRD incidence and prevalence rates worldwide [4], leading to a heavy financial burden on Taiwan’s health care systems. Thus, preventive strategies that reduce the risks of developing ESRD should be implemented among patients with DM.

Among patients with ESRD, cardiovascular disease is considered a leading cause of mortality [5], accounting for 48% of total deaths [6]. The most common cardiac arrhythmia is atrial fibrillation (AF), characterized by rapid and irregular atrial activation and associated with increased risks of poor outcomes, such as stroke, heart failure, and mortality [7]. AF and ESRD share risk factors, such as age, hypertension, obesity, DM, vascular heart disease, and heart failure [8], and patients with ESRD on dialysis have a higher AF prevalence (13%–27%) than the general population (approximately 1%), suggesting that uremia may be associated with AF [9-11]. In addition, previous studies have shown that AF is associated with increased risks of stroke and mortality among patients with ESRD on dialysis [12, 13]. The risk of developing AF is even higher among patients with ESRD on dialysis who have one or more of the following risk factors: advanced age, hypertension, heart failure, coronary artery disease (CAD), peripheral vascular disease, and chronic obstructive pulmonary disease (COPD) [14, 15].

AF is associated with poor outcomes among patients with ESRD [16], and the incidence of AF was approximately 40% among patients with DM, which is higher than the incidence in patients without DM [17]. Patients with DM who also present with AF have increased risks of stroke, cardiovascular disease, mortality, and heart failure [18, 19]. However, the potential bidirectional association between AF and ESRD has not been evaluated, and the role of AF incidence on the progression from DM to ESRD has not been elucidated. Therefore, this study examined the association of AF incidence with the risk of developing ESRD among patients with DM by conducting a retrospective population-based cohort study using claims data from the National Health Insurance Research Database (NHIRD) of Taiwan.

Materials and methods

Data source

The National Health Insurance (NHI) program of Taiwan started in 1995 and currently covers nearly 99% of Taiwan’s population of approximately 23 million people [20]. Since 1999, the NHI program has made data available to researchers through the NHIRD, which contains registration files and claims data for NHI beneficiaries, including diagnoses, demographics, medication types, prescription dates, dosages, and prescription durations [21]. In the present study, we reviewed claims data for 1 million randomly sampled NHI beneficiaries. The datasets analyzed in the current study are available in the longitudinal health insurance database (LHID) 2000 repository. The National Health Research Institutes oversees all claims data and generates random identification numbers for insured patients to protect their privacy. Because the NHIRD contains deidentified and anonymously analyzed secondary data, the need for informed consent in this study was waived. This study was approved by the institutional review board of Tungs’ Taichung MetroHarbor Hospital in Taichung, Taiwan (106206N). All protocols used in this study were performed in accordance with relevant guidelines and regulations.

From among the 1 million NHI beneficiaries who were randomly sampled from the NHIRD, we first selected those diagnosed with DM according to the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM; code 250, n = 98,213) and prescribed at least one antidiabetic drug between 2000 and 2013. We divided these patients into those with (n = 8,886) and without AF (n = 89,327) according to ICD-9-CM code 427.31. After excluding those aged <20 years (n = 611) and those diagnosed with ESRD before the index date (n = 1,456), 6,819 patients with DM and AF remained. The AF group was matched on a 1:1 basis from among the 89,327 patients without AF. The date of AF diagnosis among the AF group was defined as the index for both the patients with AF and propensity score–matched non-AF counterparts. All patients were followed until death, any diagnosis of ESRD, or December 31, 2013, whichever came first. A flow chart showing the study patient selection process is presented in Fig 1.

Fig 1

A flow chart of the patient selection process for this study.

Definition of research variables

The main outcome of this study was ESRD occurrence, defined as having received dialysis treatment between 2000 and 2013. The main comorbidities that were controlled for in this study were hypertension (ICD-9-CM code 401), heart failure (ICD-9-CM code 428), hyperlipidemia (ICD-9-CM code 272), cardiovascular disease (ICD-9-CM codes 390–459), COPD (ICD-9-CM code 490–496), stroke (ICD-9-CM code 430–438), cancer (ICD-9-CM code 140–208), and chronic kidney disease (CKD; ICD-9-CM code 585). Study patients were considered to have one of these comorbidities if they underwent at least two ambulatory visits and one hospitalization associated with the respective diagnosis.

Statistical analysis

The selection of patients without AF was performed using a 1:1 propensity score matching method [22], and matching was performed based on the nearest neighbor algorithm with a perfect proportion of 0.995 to 1.0 [23]. Propensity scores for all study patients were calculated using multivariable logistic regression adjusted for age, sex, geographic region, and Charlson Comorbidity Index (CCI) score [24].

The baseline data for patients with and without AF are presented as the frequency and percentage for categorical variables and as the mean and standard deviation for continuous variables. T-tests and χ2 tests were used to describe differences between groups with and without AF for categorical and continuous variables, respectively. The ESRD incidence rate was defined as the number of events divided by follow-up person-years, which were calculated as the time from the index date to the first received dialysis treatment, death, or December 31, 2013, whichever occurred first. We used Cox proportional hazards regression to determine adjusted hazard ratios (aHRs) and 95% confidence intervals (CIs) for ESRD comparing between groups with and without AF. By conducting the Schoenfeld residual test [25], we confirmed that the proportional hazards assumption was not violated. In multivariable analyses, we adjusted for all covariates shown in Table 2. In order to consider the competing risks of ESRD and death before ESRD, we used the cause‐specific hazard model to account for the distribution hazard for competing risks [26]. The Kaplan–Meier model was used to compare ESRD risk between groups with and without AF and between groups stratified according to CKD status (Fig 2). We further performed subgroup analyses to assess differences in AF risk among individual subgroups stratified by sex, age, hypertension, heart failure, dyslipidemia, CAD, COPD, stroke, cancer, CKD, and CCI score. Aside from the stratified variable being examined, all aHRs were adjusted for all other covariates (Fig 3). All p-values were two-sided, and any p-value <0.05 was considered significant. All analyses were computed using SAS version 9.4 (SAS Institute Inc, Cary, North Carolina).

Table 2

Association of DM progression to ESRD with AF, sociodemographic characteristics, and comorbidities.

	Crude HR (95% CI)	P-value	Adjusted HR (95% CI)	P-value
DM
Without AF	Ref.		Ref.
AF	1.36 (1.08–1.83)	0.011	1.18 (1.01–1.46)	0.043
Sex
Female	Ref.		Ref.
Male	0.67 (0.53–0.84)	0.001	0.68 (0.53–0.86)	0.015
Age, years
<60	Ref.		Ref.
60–69	0.81 (0.61–1.08)	0.809	0.67 (0.50–0.89)	0.013
70–79	0.46 (0.34–0.63)	0.462	0.40 (0.29–0.55)	<0.001
≥80	0.32 (0.20–0.51)	0.320	0.33 (0.20–0.54)	<0.001
DM duration	1.02 (0.98–1.05)	0.267	0.99 (0.94–1.03)	0.117
Hypertension
No	Ref.		Ref.
Yes	2.32 (1.82–2.95)	<0.001	1.89 (1.46–2.42)	<0.001
Heart failure
No	Ref.		Ref.
Yes	1.52 (1.11–2.09)	0.009	1.15 (0.83–1.60)	0.415
Dyslipidemia
No	Ref.		Ref.
Yes	2.07 (1.61–2.67)	<0.001	1.17 (0.89–1.52)	0.262
CAD
No	Ref.		Ref.
Yes	1.28 (0.84–1.97)	0.253	0.91 (0.59–1.41)	0.677
COPD
No	Ref.		Ref.
Yes	0.76 (0.53–1.10)	0.146	0.67 (0.44–1.02)	0.068
Stroke
No	Ref.		Ref.
Yes	0.89 (0.64–1.25)	0.892	0.62 (0.44–0.90)	0.029
Cancer
No	Ref.		Ref.
Yes	0.64 (0.34–1.20)	0.637	0.46 (0.25–0.88)	0.046
CKD
No	Ref.		Ref.
Yes	8.51 (6.54–11.06)	<0.001	6.35 (4.80–8.41)	<0.001
CCI score
<5	Ref.		Ref.
5–9	7.65 (3.13–18.69)	<0.001	7.53 (3.07–18.45)	<0.001
≥10	12.74 (5.22–31.07)	<0.001	13.25 (5.39–32.58)	<0.001

Abbreviations: AF, atrial fibrillation; CAD, cardiac artery disease; CCI, Charlson Comorbidity Index; CI, confidence interval; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disorder; DM, diabetes mellitus; ESRD, end-stage renal disease; HR, hazard ratio; Ref., reference.

Fig 2

Comparison of the end-stage renal disease curve between patients with diabetes with and without AF.

Fig 3

Risks of developing end-stage renal disease among patients with diabetes with and without atrial fibrillation, stratified according to demographic characteristics and comorbidities.

Results

Basic characteristics of study patients

The baseline characteristics of the patients and their matched controls are shown in Table 1. The average age was 68.7 years, and 51% of patients were men. Compared with DM patients without AF, those with AF were more likely to have hypertension, heart failure, dyslipidemia, CAD, COPD, stroke, and CKD; have lower percentages of cancer; and have CCI scores <5. As expected, DM patients with AF had higher prevalences of common comorbidities than those without AF. No significant difference was observed in the duration between the index date and new-onset ESRD between the case (7.6 years) and control groups (7.5 years).

Table 1

Distribution of demographic characteristics and comorbidities among DM patients with and without AF.

	AF	Without AF
	n (%)	n (%)	P-value
Sex	n = 6,105	n = 6,105
Male	3,126 (51.2)	3,138 (51.4)	0.828
Female	2,979 (48.8)	2,967 (48.6)
Age, years
<60	1,344 (22)	1,263 (20.7)	0.169
60–69	1,780 (29.2)	1,861 (30.5)
70–79	2,087 (34.2)	2,117 (34.7)
≥80	894 (14.6)	864 (14.2)
Hypertension
Yes	3,145 (51.5)	2,075 (34)	<0.001
No	2,960 (48.5)	4,030 (66)
Heart failure
Yes	1,197 (19.6)	279 (4.6)	<0.001
No	4,908 (80.4)	5,826 (95.4)
Dyslipidemia
Yes	1,499 (24.6)	1,229 (20.1)	<0.001
No	4,606 (75.5)	4,876 (79.9)
CAD
Yes	573 (9.4)	281 (4.6)	<0.001
No	5,532 (90.6)	5,824 (95.4)
PAOD
Yes	7 (0.1)	5 (0.1)	0.564
No	6,098 (99.9)	6,100 (99.9)
COPD
Yes	1,171 (19.2)	637 (10.4)	<0.001
No	4,934 (80.8)	5,468 (89.6)
Stroke
Yes	1,278 (20.9)	761 (12.5)	<0.001
No	4,827 (79.1)	5,344 (87.5)
Cancer
Yes	334 (5.5)	439 (7.2)	<0.001
No	5,771 (94.5)	5,666 (92.8)
CKD
Yes	433 (7.1)	321 (5.3)	<0.001
No	5,672 (92.9)	5,784 (94.7)
CCI score, mean ± SD	9.2 ± 3.5	9.2 ± 4.0	0.692
<5	800 (13.1)	1,114 (18.3)	<0.001
5–9	3,335 (54.6)	2,943 (48.2)
≥10	1,970 (32.3)	2,048 (33.6)
DM duration, mean ± SD	5.7 ± 4.7	5.6 ± 4.8	0.056
Duration period, mean ± SD	7.6 ± 4.6	7.5 ± 4.9	0.092

Abbreviations: AF, atrial fibrillation; CAD, cardiac artery disease; CCI, Charlson Comorbidity Index; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disorder; DM, diabetes mellitus; PAOD, peripheral arterial occlusive disease.

Hazard ratio estimation for new-onset ESRD

As shown in Table 2, male sex was associated with a lower risk of developing ESRD than female sex (aHR = 0.68, 95% CI = 0.53–0.86, p = 0.015). The risk of developing ESRD decreased (aHR = 0.67, 95% CI = 0.50–0.89, p = 0.007 for 60–69 years; aHR = 0.33, 95% CI = 0.20–0.54, p < 0.001 for ≥80 years) with increasing age compared with patients <60 years. AF, hypertension, CKD, and CCI score were strong predictors of ESRD in patients with DM. Compared with DM patients without AF, the those with AF had an increased risk of developing ESRD (aHR = 1.18, 95% CI = 1.01–1.46, p = 0.043). Patients with DM with hypertension had a higher risk of developing ESRD than those with DM without hypertension (aHR = 1.89, 95% CI = 1.46–2.42, p < 0.001). Furthermore, the risk of developing ESRD was higher for patients with DM with CKD (aHR = 6.35, 95% CI = 4.80–8.41, p < 0.001) than for those with DM without CKD. The risk of developing ESRD increased with increasing CCI scores (aHR = 7.53, 95% CI = 3.07–18.45, p < 0.001 for CCI 5–9; aHR = 13.25, 95% CI = 5.39–32.58, p < 0.001 for CCI ≥ 10) compared with CCI ≤ 5. Conversely, patients with DM with stroke and cancer were associated with decreased risks of developing ESRD than patients with DM without stroke and cancer. Furthermore, the cause‐specific hazard ratio of death before ESRD due to a competing was 1.57 (95% CI, 1.21–3.81).

Subgroup analyses

The influence of AF status on ESRD among patients with DM was examined by subgroup analyses stratified according to sex (male or female), age (<60, 60–69, 70–79, or ≥80 years), and potential comorbidities (Fig 3). The presence of certain comorbidities significantly increased the risk of developing ESRD among DM patients with AF. Among subgroups of patients with DM with comorbidities, heart failure (aHR = 3.23, 95% CI = 1.02–10.54, p = 0.048) and dyslipidemia (aHR = 1.67, 95% CI = 1.04–2.68, p = 0.035) significantly elevated the risks of developing ESRD compared with those without these comorbidities. However, a distinct difference was observed between DM patients with and without hypertension. Among patients with DM without CKD, those with AF had a higher risk of developing ESRD than those without AF (aHR = 1.31, 95% CI = 1.01–1.72, p = 0.037). When patients with DM were stratified into two groups based on CKD prevalence (Fig 2), the cumulative incidence curve analyzed by the log-rank test showed that among patients with DM without CKD, AF increased the risk of developing ESRD compared with those without AF (p = 0.023), but AF incidence did not appear to have a significant effect on ESRD risk in patients with DM with CKD.

Discussion

The present population-based study investigated the relationship between atrial fibrillation and ESRD among patients with DM using data from the Taiwan NHIRD and found that AF was associated with a higher risk of subsequent ESRD in patients with DM, and this association was significantly stronger among DM patients without CKD. To the best of the authors’ knowledge, this is the first study to investigate the relationship between AF and ESRD among patients with DM in Taiwan.

The primary new finding of this study is that AF is an important predictor of ESRD in patients with DM. Previous studies have reported that DM is an independent risk factor for AF [27-29], and metabolic syndrome is hypothesized to increase epicardial adipose tissue thickness, leading to the release of pro-inflammatory substances that contribute the endothelial dysfunction and fibrosis and influence structural and electrical atrial remodeling [30, 31]. However, no previously published studies have evaluated the impacts of incident AF on the progression from DM to adverse renal events. Our study of a large community-based cohort of patients with DM found that incident AF was associated with a higher rate of ESRD, suggesting that incident AF contributes to progression from DM to ESRD.

Multiple biological mechanisms may underlie the accelerated progression to ESRD from DM in patients with AF. First, AF triggers systemic inflammation [32]. A cohort study showed that ESRD risk in patients with DM was strongly associated with elevated concentrations of circulating pro-inflammatory cytokines, which might be responsible for kidney dysfunction [33]. Second, increasing evidence supports an association between inflammation and AF [34-36]. Additionally, various inflammatory biomarkers have been associated with AF, suggesting an important role for inflammation in AF [34], resulting in a prothrombotic state that induces renal micro-infarcts, decreased renal function, and rapid progression from DM to ESRD. Third, AF reduces left ventricular systolic and diastolic function, which may accelerate ESRD progression and impair kidney function through altered hemodynamics, reduced renal perfusion, and renin-angiotensin-aldosterone system activation [37, 38]. Fourth, some medications used to treat AF are nephrotoxic and can result in renal impairment.

We found that the association of incident AF with ESRD risk was stronger among DM patients without CKD. This finding is particularly interesting because CKD is generally accepted to increase the risk of developing AF [39-41]. The reasons for this finding are unclear; however, AF may result in stronger impacts among patients with DM without CKD because these patients are less likely to have other concomitant comorbidities that contribute to ESRD risk. Further studies are warranted to clarify this finding.

Patients with DM with AF were significantly more likely to have comorbid conditions compared with the patients without AF, including hypertension, heart failure, dyslipidemia, CAD, COPD, stroke, and CKD (Table 1). Remarkably, elevated blood pressure has been known to increase the risks of ESRD occurrence [42, 43]. As shown in univariate analyses (Table 2), hypertension plays a significant role in ESRD occurrence. However, in subgroup analyses according to the presence of hypertension, incident AF was not significantly associated with ESRD among DM patients with hypertension. The reasons for this are complex. For example, the role of hypertension as a risk factor may be masked by successful strategies designed to control blood pressure. However, data regarding the use of blood pressure medications were not available in the Taiwan NHIRD.

The present study has some limitations. First, individual behaviors associated with ESRD development were not able to be evaluated, and not all clinical data were fully available in the Taiwan NHIRD. Thus, we were not able to adjust for other potential confounding factors, such as relevant medications, blood pressure, residual renal function, and hemoglobin level. Second, all diagnoses were based on ICD-9-CM diagnosis codes, which do not differentiate among AF types and pose a risk for coding bias. Lastly, the results generated from our study may not be applicable to other countries as the study was conducted in only one country.

Conclusions

In conclusion, incident AF is associated with a relative increase in the risk of ESRD development among patients with DM. Further research remains necessary to delineate contributing factors that lead to AF development in the setting of DM and elucidate potential modifiable pathways through which AF contributes to the progression to ESRD.

STROBE statement.

(DOC)

Click here for additional data file.

20 Jan 2022

Association between Atrial Fibrillation and Risk of End-Stage Renal Disease among Adults with Diabetes Mellitus

PLOS ONE

Dear Dr. Lim,

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.

Please submit your revised manuscript by Mar 05 2022 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're 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.

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). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Abduzhappar Gaipov

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

The methodological issues should be improved

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

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Thank you for stating the following financial disclosure: "No"

At this time, please address the following queries:

a) Please clarify the sources of funding (financial or material support) for your study. List the grants or organizations that supported your study, including funding received from your institution.

b) State what role the funders took in the study. If the funders had no role in your study, please state: “The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.”

c) If any authors received a salary from any of your funders, please state which authors and which funders.

d) If you did not receive any funding for this study, please state: “The authors received no specific funding for this work.”

Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

3. Thank you for stating the following in the Acknowledgments Section of your manuscript:

"This work was funded by grants from Tungs’ Taichung Metroharbor Hospital (TTMHH-106R)."

We note that you have provided funding information that is not currently declared in your Funding Statement. However, funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form.

Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows:

"This work was funded by grants from Tungs’ Taichung Metroharbor Hospital (TTMHH-106R)."

Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[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: Yes

Reviewer #2: No

**********

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

Reviewer #1: Yes

Reviewer #2: 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: No

Reviewer #2: No

**********

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: Yes

Reviewer #2: Yes

**********

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: 1) As authors evaluate the effect of AF in progression of ESRD among DM patients, it would be useful to have a discussion

about the prevalence of AF among DM patients. Authors had discussed the joint effect of AF and ESRD on mortality and morbidity,

but the focus on DM patients is missing in introduction. Are the discussed excessive risk is similar for DM patients or different?

2) Authors declare taking age, gender, region and CCI score for calculation of PS score, but justification on variable selection

for Propensity score analysis is missing. Was it based on statistical methods or clinical importance? Did author have chance to adjust

on more number of variables at this stage? Adjusting for duration of DM or severity at index date might have introduced better

similarity among groups.

3) As authors conducted the retrospective cohort study, thorough explanation of dates is important. At what point the age was taken? Index date, the onset of DM or end of follow-up?

The explanation of defining index date is unclear in "Methods" part. Was it taken after PS matching and individual date of AF incidence

was taken for specific controls? If it is true, this have some drawbacks. Despite the age being controlled, the onset of DM might be different, unless you control it,

but there was no indication on adjustment on DM onset date or its duration. Thus, taking the date of AF onset as an index date for controls might introduce bias

as those with AF might tend to have longer DM duration compared to those without and those with AF will have higher risk of development of other complications at

baseline. Suggestion is again to control for DM duration and if possible the age of DM onset. Also, changing the index date to the date of DM onset might reduce bias.

Additionally, the way of counting the mentioned comorbidities in the study is vague. Was it counted throughout the 2000-2013 period or only after the index date?

If the information on comorbidities' incidence before index date is available, these information might have been used to PS score calculation to make comparison

group homogeneous regarding baseline comorbidities.

4) As there patients might die before developing ESRD, it might worth to perform competing risk analysis.

5) It would be better to include the initial distribution of demographic characteristics and comorbidities before PS matching to be able to evaluate on the goodness of

balancing and how distribution of comorbidities changes.

Reviewer #2: Thank you for sharing your manuscript with me for review. In general the study has an interesting research question and has sufficiently large study sample to explore the research question. However, there are several major limitations in statistical analysis and description of the methods that question methodological robustness of the study. Please my comments below:

1. I appreciate that the authors are attempting to examine the association between AF and ESRD among DM patients where the exposure variable is AF and ESRD is the outcome. However, the following statement in the introduction section is a bit confusing: “…Furthermore, the impact of AF incidence on the progression of DM to ESRD has not been elucidated.”

2. Not familiar with the term “insurance cohort study”. I would recommend to omitting “insurance” word from the study design description as it could cause confusion.

3. In this sentence, “All protocols used in this study were performed in accordance with relevant guidelines and regulations”, could the authors clarify what does it mean relevant guidelines and regulations?

4. In this sentence, “… we first selected those who had been diagnosed with DM according to the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM; code 250, n = 98,213) and at least one antidiabetic drug”, what did author mean by “at least one antidiabetic drug”?

5. I am not familiar with the NHIRD, but I am guessing it could contain diagnoses about comorbidities. Have the authors attempted to look for AF and DM in all data including among comorbidities or variables that could point to these diseases? For example, elevated blood glucose level could be indication for DM or ECG records for AF?

6. It would be nice to see the definitions of the ICD codes that were used in the manuscript. At least in the Supplementary section.

7. This sentence is not clear “After excluding those aged <20 years (n = 611) and diagnosed with TMD before the index date (n = 1,456) among the AF group, patients were then matched on a 1:1 basis with 89,327 patients without AF, with the initial date of AF diagnosis for a given AF subject being defined as the index date for the without AF group with which he or she was matched.”

8. What does “TMD” stand for?

9. Clearly define what index date means in this context.

10. The description of the propensity score matching approach should be placed in the statistical analysis subsection.

11. Did the authors adjusted for “age, gender, geographic region, and Charlson Comorbidity Index (CCI)” or did they use these variables to calculate propensity scores? Why were these variables selected for the propensity score calculation?

12. Why did the authors choose to include the covariates in Table 2 when building the multivariable model? Were other covariates available from the NHIRD database that could potentially confound the relationship of AF with ESRD?

13. Since death is a competing risk for ESRD occurrence, the authors could perform additionally competing risk regression analysis. I would strongly encourage to perform such analysis.

14. It is not clear why the researchers included some of the variables in the multivariable model in Table 2 that were used for propensity score calculations such as age, sex and CCI? Have the authors determined whether patients with AF and without AF have similar distributions in terms of the variables that were used to calculate the propensity scores?

15. In the subgroup analyses, it is not clear how adjusted HRs were calculated. Did the authors include all interactions together in one model or did they calculate estimates by including one interaction at a time?

16. Why were the confidence intervals for heart failure, CAD and cancer too wide? Have the authors tried to minimize numerical issues when calculating HRs with such wide confidence intervals?

17. It would be also interesting to see if unmatched groups had similar results when performing the multivariable regression analyses? The authors could provide as additional table in the Supplementary section.

18. In Table 1, CCI score has two p-values. Is this a typo or the authors performed two tests?

19. The first two sentences in the Results section are redundant and can be omitted.

20. The following statement “that AF in an important predictor of ESRD in patients with DM” is too bold and not supported given limitations of the study as the authors pointed out some of the important covariates are missing from the multivariable analysis. The authors should avoid such questionable claims and frame the discussion around the association, not causality.

21. What did the authors mean by “pose a risk for coding bias”? Measurement errors?

**********

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

Reviewer #2: 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.]

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 PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

21 Apr 2022

Response to Reviewers’ comments

Ref.: PONE-D-21-17026

Reviewers’ comments:

Reviewer #1:

Reviewer #1: 1) As authors evaluate the effect of AF in progression of ESRD among DM patients, it would be useful to have a discussion about the prevalence of AF among DM patients. Authors had discussed the joint effect of AF and ESRD on mortality and morbidity, but the focus on DM patients is missing in introduction. Are the discussed excessive risk is similar for DM pati ents or different?

Reply: Thank you for your comment. Please see Main document_Clean file. We have added the following sentences and associated references to the introduction (P3L25–26 and P4L1–3): “AF is associated with poor outcomes among patients with ESRD [16], and the incidence of AF was approximately 40% among patients with DM, which is higher than the incidence in patients without DM [17]. Patients with DM who also present with AF have increased risks of stroke, cardiovascular disease, mortality, and heart failure [18, 19].”

2) Authors declare taking age, gender, region and CCI score for calculation of PS score, but justification on variable selection for Propensity score analysis is missing. Was it based on statistical methods or clinical importance? Did author have chance to adjust on more number of variables at this stage? Adjusting for duration of DM or severity at index date might have introduced better similarity among groups.

Reply: Thank you for your comment. We selected age, sex, geographic region, and CCI score as the variables included in the propensity score because the inclusion of additional risk factors would result in the loss of many samples and representation. We opted to use the CCI score before the index date to control for the presence of comorbidities during matching.

3) As authors conducted the retrospective cohort study, thorough explanation of dates is important. At what point the age was taken? Index date, the onset of DM or end of follow-up? The explanation of defining index date is unclear in “Methods” part. Was it taken after PS matching and individual date of AF incidence was taken for specific controls? If it is true, this have some drawbacks. Despite the age being controlled, the onset of DM might be different, unless you control it, but there was no indication on adjustment on DM onset date or its duration. Thus, taking the date of AF onset as an index date for controls might introduce bias as those with AF might tend to have longer DM duration compared to those without and those with AF will have higher risk of development of other complications at baseline. Suggestion is again to control for DM duration and if possible the age of DM onset. Also, changing the index date to the date of DM onset might reduce bias. Additionally, the way of counting the mentioned comorbidities in the study is vague. Was it counted throughout the 2000-2013 period or only after the index date? If the information on comorbidities’ incidence before index date is available, these information might have been used to PS score calculation to make comparison group homogeneous regarding baseline comorbidities.

Reply: Thank you for your comment. The index date was the date of AF incidence among patients with DM with AF, and this date was used as the index date for all matched patients with DM without AF. The index date was defined in the Methods as follows: “The date of AF diagnosis among the AF group was defined as the index for both the patients with AF and propensity score–matched non-AF counterparts.” (Please see P5L9–10). The study was focused on patients with and without AF among patients with DM, and all patients were followed from the index date until death, any ESRD diagnosis, or December 31, 2013, whichever came first. Although we selected the date of AF diagnosis as the index day, we included DM duration and the age of DM diagnoses as adjusted variables (Please see revised Tables 1 and 2 and Figure 3, P7L11–14 and P7L17–24 and P8L7–8 and P8L12–13 in the Main document_Clean file).

4) As there patients might die before developing ESRD, it might worth to perform competing risk analysis.

Reply: Thank you for your comment. The adjusted hazard ratio of death before ESRD due to a competing was 1.57 (95% CI, 1.21–3.81), which was similar to the original hazard ratio.

5) It would be better to include the initial distribution of demographic characteristics and comorbidities before PS matching to be able to evaluate on the goodness of balancing and how distribution of comorbidities changes.

Reply: Thank you for your comment. We performed propensity score matching based on sex, age, and CCI to control the distribution of demographic characteristics among patients with DM with and without AF.

Reviewer #2: Thank you for sharing your manuscript with me for review. In general the study has an interesting research question and has sufficiently large study sample to explore the research question. However, there are several major limitations in statistical analysis and description of the methods that question methodological robustness of the study. Please my comments below:

1. I appreciate that the authors are attempting to examine the association between AF and ESRD among DM patients where the exposure variable is AF and ESRD is the outcome. However, the following statement in the introduction section is a bit confusing: “…Furthermore, the impact of AF incidence on the progression of DM to ESRD has not been elucidated.”

Reply: Thank you for your comment. We have revised the sentence as follows. “However, the potential bidirectional association between AF and ESRD has not been evaluated, and the role of AF incidence on the progression from DM to ESRD has not been elucidated.” (Please see P4L3–5 in the Main document_Clean file)

2. Not familiar with the term “insurance cohort study”. I would recommend to omitting “insurance” word from the study design description as it could cause confusion.

Reply: Thank you for your comment. We had have removed the word “insurance” and appreciate your recommendation. (Please see P4L7 in the Main document_Clean file)

3. In this sentence, “All protocols used in this study were performed in accordance with relevant guidelines and regulations”, could the authors clarify what does it mean relevant guidelines and regulations?

Reply: This sentence means this study was performed in accordance with the institutional review board guidelines and regulations.

4. In this sentence, “… we first selected those who had been diagnosed with DM according to the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM; code 250, n = 98,213) and at least one antidiabetic drug”, what did author mean by “at least one antidiabetic drug”?

Reply: Because all individuals were categorized as having DM based only on data available in the NHIRD, we defined the criteria for DM as having both a DM diagnosis code and the use of at least one antidiabetic drug to ensure our cohort was appropriately diagnosed.

5. I am not familiar with the NHIRD, but I am guessing it could contain diagnoses about comorbidities. Have the authors attempted to look for AF and DM in all data including among comorbidities or variables that could point to these diseases? For example, elevated blood glucose level could be indication for DM or ECG records for AF?

Reply: The NHIRD contains diagnosis codes for AF, DM, and comorbidities, and we attempted to evaluate the relationship between AF and ESRD among DM patients after adjusting for comorbidities.

6. It would be nice to see the definitions of the ICD codes that were used in the manuscript. At least in the Supplementary section.

Reply: Thank you for your comment. We have provided definitions of the ICD codes in the Methods section. (Please see P5L3, P5L6 and P5L17–21 in the Main document_Clean file)

7. This sentence is not clear “After excluding those aged <20 years (n = 611) and diagnosed with TMD before the index date (n = 1,456) among the AF group, patients were then matched on a 1:1 basis with 89,327 patients without AF, with the initial date of AF diagnosis for a given AF subject being defined as the index date for the without AF group with which he or she was matched.”

Reply: Thank you for your comment. We have revised this sentence as follows: “After excluding those aged <20 years (n = 611) and those diagnosed with ESRD before the index date (n = 1,456), 6,819 patients with DM and AF remained. The AF group was matched on a 1:1 basis from among the 89,327 patients without AF. The date of AF diagnosis among the AF group was defined as the index for both the patients with AF and propensity score–matched non-AF counterparts.” (Please see P5L6–10 in the Main document_Clean file)

8. What does “TMD” stand for?

Reply: Thank you for your comment. The “TMD” was an error and has been revised to “ESRD.” (Please see P5L7 in the Main document_Clean file, and Figure 1)

9. Clearly define what index date means in this context.

Reply: The index date was the date of AF incidence among patients with DM with AF, and this date was used as the index date for all matched patients with DM without AF. (Please see P5L9–10 in the Main document_Clean file)

10. The description of the propensity score matching approach should be placed in the statistical analysis subsection.

Reply: Thank you for your recommendation. The description of the propensity score matching approach has been placed in the first section of the statistical analysis section. (Please see P5L25–26 and P6L1-3 in the Main document_Clean file)

11. Did the authors adjusted for “age, gender, geographic region, and Charlson Comorbidity Index (CCI)” or did they use these variables to calculate propensity scores? Why were these variables selected for the propensity score calculation?

Reply: Thank you for your comment. We used the propensity score matching based on the sex, age, and Charlson Comorbidity Index (CCI) score because the inclusion of too many risk factors would result in the loss of numerous samples and representation. We opted to use the CCI score before the index date to control for comorbidities during the matching process.

12. Why did the authors choose to include the covariates in Table 2 when building the multivariable model? Were other covariates available from the NHIRD database that could potentially confound the relationship of AF with ESRD?

Reply: Thank you for your comment. We selected covariates that could be potential confounding factors in the progression to ESRD. Unfortunately, some potentially confounding factors that might contribute to ESRD are not available in the NHIRD, including individual behaviors, some clinical data, relevant medications, blood pressure, residual renal function, and hemoglobin levels. We have added this as a limitation of our study in the last section of the Discussion (Please see P10L16–20 in the Main document_Clean file).

13. Since death is a competing risk for ESRD occurrence, the authors could perform additionally competing risk regression analysis. I would strongly encourage to perform such analysis.

Reply: Thank you for your comment. The adjusted hazard ratio of death before ESRD due to a competing was 1.57 (95% CI, 1.21–3.81), which was similar to the original hazard ratio.

14. It is not clear why the researchers included some of the variables in the multivariable model in Table 2 that were used for propensity score calculations such as age, sex and CCI? Have the authors determined whether patients with AF and without AF have similar distributions in terms of the variables that were used to calculate the propensity scores?

Reply: We performed propensity score matching to identify patients with AF and without AF with similar demographics (sex and age) and comorbidities (CCI) to provide a better comparison between these two groups.

15. In the subgroup analyses, it is not clear how adjusted HRs were calculated. Did the authors include all interactions together in one model or did they calculate estimates by including one interaction at a time?

Reply: Thank you for your comment. We compared individual subgroups stratified by a single variable at a time, including sex, age, hypertension, heart failure, dyslipidemia, CAD, COPD, stroke, cancer, CKD, and CCI score. Except for the examined variable, all adjusted HRs were adjusted for all covariates. We have revised this explanation in the manuscript. (Please see P6L17-21 in the Main document_Clean file)

16. Why were the confidence intervals for heart failure, CAD and cancer too wide? Have the authors tried to minimize numerical issues when calculating HRs with such wide confidence intervals?

Reply: Thank you for your comment. Because of propensity score matching, the numbers of patients with heart failure, CAD, or cancer were fewer than patients with other comorbidities, and these subgroups showed a reduction in ESRD incidence but with wide confidence intervals.

17. It would be also interesting to see if unmatched groups had similar results when performing the multivariable regression analyses? The authors could provide as additional table in the Supplementary section.

Reply: Thank you for your comment. In our study design, the index date for non-AF patients was taken at the time of matched AF-patient onset. The observation period began from the index date to the final follow-up date, which is why non-AF (patients) were not subjected to Cox proportional regression analysis and lack matching index dates. As a result, we are unable to provide the results of unmatched groups.

18. In Table 1, CCI score has two p-values. Is this a typo or the authors performed two tests?

Reply: Thank you for your comment. The first p-value is the comparison of continuous variables by t-test. The second p-value is the comparison of categorical variables by the chi-square test.

19. The first two sentences in the Results section are redundant and can be omitted.

Reply: Thank you for your comment. We have deleted these first two sentences (Please see the Main document_Clean file).

20. The following statement “that AF in an important predictor of ESRD in patients with DM” is too bold and not supported given limitations of the study as the authors pointed out some of the important covariates are missing from the multivariable analysis. The authors should avoid such questionable claims and frame the discussion around the association, not causality.

Reply: Thank you for your comment. We had revied the last sentence of the Discussion to describe the finding of an association between incident AF and ESRD (Please see P11L1-2 in the Main document_Clean file).

21. What did the authors mean by “pose a risk for coding bias”? Measurement errors?

Reply: Thank you for your comment. Because NHIRD data represents health insurance declaration data, a very small chance of coding bias exists.

Submitted filename: Response to reviewers.docx

Click here for additional data file.

6 Jun 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Abduzhappar Gaipov

Academic Editor

PLOS ONE

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

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

Reviewer #2: (No Response)

**********

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

Reviewer #2: Partly

**********

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

Reviewer #1: Yes

Reviewer #2: No

**********

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: No

Reviewer #2: No

**********

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

Reviewer #2: 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: Thank you for letting me review given article. I have no further comments. Comments were addressed as fully as possible.

Reviewer #2: I appreciate that the authors' responses to my comments. However, I still have a few concerns that might be helpful in strengthening the manuscript:

The authors indicated that some risk factors have missing values which result in smaller sample size and a loss of statistical power. Given presence of missing values of important covariates for analysis, it would be interesting to conduct multiple imputations (MI). MI will allow a better statistical power and adjustment for more appropriate set of covariates to draw conclusions.

I see that the authors performed competing risk analysis. I would suggest including the description of methods used to perform the competing risk analysis and provide these results in the manuscript.

While the authors indicated that groups were matched using PS, Table 1 shows that there are differences in some of the important covariates. I do understand that the authors included these covariates later in the multivariable regression. However, why would the authors perform PS matching (while missing some of important covariates) and then including them in the final analysis. Would it be easier to just conduct multivariable analysis without matching? And then conduct analysis on the matched groups? I am just struggling to understand with concepts why the authors done the analysis this 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

Reviewer #2: 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.]

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 PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

4 Jul 2022

Reviewer #1: Thank you for letting me review given article. I have no further comments. Comments were addressed as fully as possible.

Reply: Thank you.

Reviewer #2: I appreciate that the authors' responses to my comments. However, I still have a few concerns that might be helpful in strengthening the manuscript:

The authors indicated that some risk factors have missing values which result in smaller sample size and a loss of statistical power. Given presence of missing values of important covariates for analysis, it would be interesting to conduct multiple imputations (MI). MI will allow a better statistical power and adjustment for more appropriate set of covariates to draw conclusions.

Reply: Thank you for your comment. MI is a tool to deal the missing research data essentially that happens when people fail to respond to a survey. The NHIRD is a registered claims dataset that has fewer missing data problem. In the previous study was showed that a rule of thumb blow of 5% missingness was not necessary to multiple imputation.1 So that MI might not be suitable to use in the study that is not somewhat true to the original in order to add the statistic power. On the other hand, we chose the PS matching in the statistical analysis of observational data to construct an artificial control group by matching each treated unit with a non-treated unit of similar characteristics that attempts to reduce bias arises.

Reference:

1.Jakobsen JC, Gluud C, Wetterslev J, Winkel P. When and how should multiple imputation be used for handling missing data in randomised clinical trials - a practical guide with flowcharts. BMC Med Res Methodol. 2017;17(1):162. Epub 2017/12/07. doi: 10.1186/s12874-017-0442-1. PubMed PMID: 29207961; PubMed Central PMCID: PMCPMC5717805.

I see that the authors performed competing risk analysis. I would suggest including the description of methods used to perform the competing risk analysis and provide these results in the manuscript.

Reply: Thank you for your comment. Please see Main document_Clean file. We have add the description of competing risk analysis in the Methods. (Please see P6L15-17)

While the authors indicated that groups were matched using PS, Table 1 shows that there are differences in some of the important covariates. I do understand that the authors included these covariates later in the multivariable regression. However, why would the authors perform PS matching (while missing some of important covariates) and then including them in the final analysis. Would it be easier to just conduct multivariable analysis without matching? And then conduct analysis on the matched groups? I am just struggling to understand with concepts why the authors done the analysis this way.

Reply: Thank you for your kindly comment. In the statistical analysis of observational data, PS matching is a quasi-experimental method such as the investigator uses statistical techniques to construct an artificial control group by matching each treated unit with a non-treated unit of similar characteristics that attempts to reduce bias arises. Because a difference in the treatment outcome between treated and untreated groups may be caused by a factor that predicts treatment rather than the treatment itself. We had used this statistical analysis published some journal articles based on NHIRD1-8. In our earliest study8, we used PS matching among patients with peritoneal dialysis and hemodialysis, because of patients treated with PD were different from patients treated with HD in terms of health status. The PD and HD cohorts were homogenous in sociodemographic characteristics and comorbid conditions at the baseline after using the propensity score-matched design, which reduces the selection bias of samples and potential confounding effect. The general rule of practice in the previous study was showed that the covariates can be added into a regression adjustment that could be dramatically remove residual confounding bias if remaining imbalance after PS matching.9

References:

1.Chen YY, Fan HC, Tung MC, Chang YK. The association between Parkinson's disease and temporomandibular disorder. PLoS One. 2019; 14(6):e0217763.

2.Lu CW, Chang YK, Lee YH, Kuo CS, Chang HH, Huang CT, Hsu CC, Huang KC. Increased Risk for Major Depressive Disorder in Severely Obese Patients after Bariatric Surgery - a 12-year Nationwide Cohort Study. Ann Med. 2018; 13:1-30. doi: 10.1080/07853890.2018.1511917.

3.Chang HH, Chang YK, Lu CW, Huang CT, Chien CT, Hung KY, Huang KC, Hsu CC. Statins Improve Long Term Patency of Arteriovenous Fistula for Hemodialysis. Sci Rep. 2016; 6:22197. doi: 10.1038/srep22197.

4.Lu CW, Chang YK, Chang HH, Kuo CS, Huang CT, Hsu CC, Huang KC. Fracture Risk After Bariatric Surgery: A 12-Year Nationwide Cohort Study. Medicine (Baltimore). 2015; 94(48): e2087.

5.Hung SC, Chang YK, Liu JS, Hsu CC, Tarng DC. Mortality and metformin use in patients with advanced chronic kidney disease. Lancet Diabetes Endocrinol. 2015 Aug;3(8):605-14. doi: 10.1016/S2213-8587(15)00123-0.

6.Hsu CC, Wang H, Hsu YH, Chuang SY, Huang YW, Chang YK, Liu JS, Hsiung CA, Tsai HJ. Use of Nonsteroidal Anti-Inflammatory Drugs and Risk of Chronic Kidney Disease in Subjects With Hypertension: Nationwide Longitudinal Cohort Study. Hypertension. 2015; 66(3): 524-33. doi: 10.1161/HYPERTENSIONAHA. 114.05105.

7.Kuo KL, Hung SC, Liu JS, Chang YK, Hsu CC, Tarng C. Iron supplementation associates with low mortality in pre-dialyzed advanced chronic kidney disease patients receiving erythropoiesis-stimulating agents: a nationwide database analysis. Nephrol Dial Transplant. 2015; 30(9): 1518-25. doi: 10.1093/ndt/gfv085.

8.Chang YK, Hsu CC, Hwang SJ, Chen PC, Huang CC, Li TC, Sung FC. A Comparative Assessment on Survival between Propensity Score Matched Patients with Peritoneal Dialysis and Hemodialysis in Taiwan. Medicine. 2012; 91(3): 144-151. doi: 10.1097/MD.0b013e318256538e.

9.Nguyen TL, Collins GS, Spence J, Daures JP, Devereaux PJ, Landais P, et al. Double-adjustment in propensity score matching analysis: choosing a threshold for considering residual imbalance. BMC Med Res Methodol. 2017;17(1):78. Epub 2017/04/30. doi: 10.1186/s12874-017-0338-0. PubMed PMID: 28454568; PubMed Central PMCID: PMCPMC5408373.

Submitted filename: Response to reviewers.docx

Click here for additional data file.

15 Aug 2022

Association Between Atrial Fibrillation and Risk of End-Stage Renal Disease Among Adults With Diabetes Mellitus

PONE-D-21-17026R2

Dear Dr. Lim,

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

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. 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 help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- 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.

Kind regards,

Abduzhappar Gaipov

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

Reviewer #2: (No Response)

**********

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

Reviewer #2: Partly

**********

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

Reviewer #1: Yes

Reviewer #2: No

**********

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: (No Response)

Reviewer #2: No

**********

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

Reviewer #2: 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: There is no further comments. All comment were addressed by the authors.

Reviewer #2: I do not fully agree with the authors' responses and selected analytical plan, but theoretically the results should be quite similar (to other alternative proposed approach)

**********

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

Reviewer #2: No

**********

17 Aug 2022

PONE-D-21-17026R2

Association Between Atrial Fibrillation and Risk of End-Stage Renal Disease Among Adults With Diabetes Mellitus

Dear Dr. Lim:

I'm 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 let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, 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.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Abduzhappar Gaipov

Academic Editor

PLOS ONE