Baseline atrial fibrillation is associated with contrast‐induced nephropathy after cardiac catheterization in coronary artery disease: Systemic review and meta‐analysis

Abstract

Background

Atrial fibrillation (AF) is the most common arrhythmia, independently associated with significant mortality and morbidity. Recent studies suggest that AF is potentially associated with contrast‐induced nephropathy (CIN) in patients with coronary artery disease (CAD) undergoing catheterization. However, the association was not conclusive. Thus, we assessed the association between AF in patients with CAD and CIN by a systematic review of the literature and a meta‐analysis.

Hypothesis

AF is a predictor of CIN in patients with CAD.

Methods

We comprehensively searched the databases of MEDLINE and EMBASE from inception to April 2018. Included studies were published observational studies that compared the risk of CIN among CAD patients with AF vs those without AF. Data from each study were combined using the random‐effects, generic inverse variance method of DerSimonian and Laird to calculate risk ratios and 95% confidence intervals (CIs).

Results

Eight cohort studies from June 2007 to November 2017 were included in this meta‐analysis involving 16,691 subjects with CAD (1,030 with AF and 15,661 without its presence). The presence of AF was associated with CIN (pooled risk ratio = 2.17, 95% CI: 1.50‐3.14, P < 0.001, I ² = 54.1%). In our subgroup analysis by urgency and multivariable adjustment, both groups still showed substantial association between AF and CIN (P < 0.05).

Conclusions

AF increased the risk of CIN up to two fold among patients with CAD compared to the absence of it. Our study suggests that the presence of AF in CAD is prognostic for the development of CIN.

Abbreviations

ACS

acute coronary syndrome

AF

atrial fibrillation

CAD

coronary artery disease

CHF

congestive heart failure

CI

confidence interval

CIN

contrast‐induced nephropathy

CKD

chronic kidney disease

DM

diabetes mellitus

LVEF

left ventricular ejection fraction

MI

myocardial infarction

NSAIDs

nonsteroidal anti‐inflammatory drugs

OR

odd ratio

PCI

percutaneous coronary intervention

1. INTRODUCTION

Contrast‐induced nephropathy (CIN), also known as contrast‐induced acute kidney injury, defined as an impairment of renal function occurring within 48 hours after contrast media administration, is the third leading cause of hospital‐acquired acute kidney injury.1 Proposed mechanisms include acute tubular necrosis from renal vasoconstriction causing medullary hypoxia and direct cytotoxic effect from the contrast media.2, 3, 4 CIN was found associated with poor clinical outcomes, including long‐term mortality. Since it is preventable, identifying patients with elevated risk is necessary. Several conditions associated with CIN were investigated. Recognized risk factors include, but not limited to, preexisting chronic kidney disease (CKD), diabetes mellitus (DM), older age, congestive heart failure (CHF), hypertension, myocardial infarction (MI), use of intra‐aortic balloon pump and increased volume of contrast medium.4 Several tools to predict CIN were developed. One of the most known tools, which was proposed in 2004 by Mehran et al., included eight variables; hypotension, intra‐aortic balloon pump, CHF, CKD, DM, age more than 75 years, anemia and volume of contrast.5

Atrial fibrillation (AF) is one of the most common supraventricular arrhythmias in adults and is even more prevalent in patients undergoing percutaneous coronary intervention (PCI). AF is associated with significant morbidity and mortality6 and was proven to be associated with particularly poor outcomes in patients with acute coronary syndrome (ACS).7 Recent studies suggest AF may be associated with increased risk of CIN in patients undergoing PCI.8, 9, 10, 11, 12, 13, 14, 15 Nonetheless, the association was not conclusive and no meta‐analysis has been performed to address this issue. Hence, we conducted this systematic review and meta‐analysis to determine the impact of AF on the development of CIN.

2. METHODS

The meta‐analyses were conducted in adherence with the recommendation of the MOOSE guidelines (Meta‐analysis of Observational Studies in Epidemiology) including items such as background, search strategy, methods, results, discussion and conclusions (Supporting Information Table S1).16

2.1. Search strategy

Two investigators (Chanavuth Kanitsoraphan and Veraprapas Kittipibul) independently searched for published observational studies indexed in MEDLINE and EMBASE databases from inception to March 2018 using a search strategy (Figure 1) that included the terms “atrial fibrillation,” “Contrast‐induced nephropathy,” and “Contrast‐induced kidney injury”. Only English language publications were included. A manual search for additional pertinent studies and review articles using references from retrieved articles was also completed.

Figure 1.

Search methodology and selection process. AF, atrial fibrillation; CIN, contrast‐induced nephropathy

2.2. Study eligibility criteria

Two main criteria were assessed for inclusion of studies. First was the reporting of incidence of CIN in patients undergoing cardiac catheterization who presented with and without AF. Second was reporting of relative risk, hazard ratio, odds ratio, incidence ratio, or standardized incidence ratio with 95% confidence intervals (CIs) (or sufficient raw data for the calculation), all of which were only related to CIN risk. Patients underwent cardiac catheterization who did not have AF were used as controls. Study eligibility was independently determined by two investigators (Chanavuth Kanitsoraphan and Veraprapas Kittipibul). Differences were resolved by mutual consensus. The Newcastle‐Ottawa quality assessment scale was used to evaluate each study in three domains: recruitment and selection of the participants, similarity and comparability between the groups, and ascertainment of the outcome of interest among cohort studies.17

2.3. Quality assessment of included studies

The Newcastle‐Ottawa scale (0‐9) was used to evaluate included studies on three domains: selection, comparability, and outcomes. Higher scores represent higher study quality. All studies received a score of 7 to 9, which reflected high quality of included studies. Detailed evaluation of each study is presented in a supplementary table (Supporting Information Table S2).

2.4. Definition

CIN was defined differently based on each recruited study (Table 1).

Table 1.

The clinical characteristics and summary of the included studies

First author	Abe	Balli	Duzel	Inohara	Moo	Roghi	Roubin	Sedhai
Year	2017	2016	2017	2015	2013	2008	2015	2017
Country	Japan	Turkey	Turkey	Japan	Korea	Italy	Spain	United States
Study type	Retrospective cohort	Prospective cohort	Prospective cohort	Prospective cohort	Retrospective cohort	Prospective cohort	Retrospective cohort	Retrospective cohort
Participant description	Patients underwent first elective PCI	STEMI patients underwent primary PCI	NSTEMI patients underwent PCI	Patients underwent PCI	Patients underwent PTCA	Patients underwent PCI	ACS patients underwent CAG	Patients hospitalized after CAG
Exclusion criteria	Emergent PCI, hemodialysis	Known obstructive CAD, severe VHD, LVEF < 30%, SCr ≥1.5 mg/dL, recent MI, IABP	Severe VHD, Severe HF, IABP, eGFR <30 mL/min/1.73 m2, unsuccessful PCI, emergent CABG after PCI	Hemodialysis, cardiopulmonary support, cardiac arrest, baseline SCr > 8 mg/dL	Exposure to other CM, ESRD, malignancy, acute infection	Acute STEMI, hemodialysis	Long‐term hemodialysis	ESRD
Participants (N)
AF	462	26	28	157	23	141	139	54
Non‐AF	5054	624	1017	3800	712	2719	1381	364
AF prevalence	9.08	4.17	2.75	4.13	3.23	5.19	10.07	14.84
Mean age (years)	N/A	59.9 ± 11.9	63.6 ± 7.3	N/A	64.8 ± 10.6	63 ± 10	67.1 ± 12.7	69.1 ± 13.8
CIN definition	Elevation in peak SCr ≥0.5 mg/dL from baseline within 5 days after PCI	Increase in SCr by >25% or 0.5 mg/dL from baseline within 72 hours after PCI	Increase in SCr by ≥25% or 0.5 mg/dL from baseline within 72 hours after PCI	Increase in SCr of 50% or 0.3 mg/dL from baseline after PCI within 48 hours	Increase of ≥0.5 mg/dL or ≥25% in SCr during 72 hours following PTCA	Increase in SCr of ≥0.5 mg/dL from baseline at 24 hours after PCI	Increase in SCr of 0.5 mg/dL within 72 hours of CM exposure	Increase in SCr by ≥25% or 0.5 mg/dL from baseline within 72 hours after CM exposure
CIN prevalence
AF	33 (7.1%)	9 (34.6%)	6 (21.4%)	18 (11.5%)	2 (8.7%)	10 (7.1%)	N/A	8 (13.8%)
Non‐AF	185 (3.6%)	50 (8.0%)	80 (7.9%)	340 (8.9%)	62 (8.7%)	96 (3.5%)		8 (2.3%)
All	218 (3.9%)	59 (9.1%)	86 (8.2%)	358 (9.0%)	64 (8.7%)	106 (3.7%)		16 (3.8%)
Odd ratio (95% CI)	N/A	6.80 (2.77‐16.70)	3.83 (1.24‐11.84)	N/A	N/A	2.08 (1.06‐4.10)	1.69 (0.89‐3.22)	4.11 (1.40‐12.07)
Confounder adjustment	N/A	Demographics, DM, LVEF, WBC, platelets, eGFR, baseline SCr, medications	Age, DM, OAC Use, Mehran risk score, baseline SCr, eGFR, prior CABG, peak troponin, LVEF	N/A	N/A	Demographics, comorbidity, LVEF, fluoroscopy time, eGFR, medications	Age, HTN, STEMI, LVEF, SCr on admission	HF, CKD
Conclusion by authors	Persistent CIN was independently associated with increased long‐term mortality	AF can contribute to CIN development in patients with STEMI	AF can be related to CIN development in patients with NSTEMI	Predictors of CIN included older age, HF, DM, previous PCI, HTN, higher baseline SCr and ACS.	Diastolic dysfunction may be a useful parameter in CIN risk stratification	ARF following PCI occurs almost exclusively in patients with CKD or LV dysfunction	Development of CIN is an independent predictor of new‐onset AF in ACS patients	Preexisting AF may increase risk for CIN.
Newcastle‐Ottawa quality assessment	9	9	9	8	8	8	9	9

Abbreviations: ACS, acute coronary syndrome; AF, atrial fibrillation; ARF, acute renal failure; CABG, coronary artery bypass grafting; CAD, coronary artery disease; CAG, coronary angiography; CI, confidence interval; CIN, contrast‐induced nephropathy; CKD, chronic kidney disease; CM, contrast medium; DM, diabetes mellitus; eGFR, estimated glomerular filtration rate; ESRD, end‐stage renal disease; HF, heart failure; HTN, hypertension; IABP, intra‐aortic balloon pump; LV, left ventricle; LVEF, left ventricular ejection function; MI, myocardial infarction; N/A, not applicable; NSTEMI, non‐ST‐elevation myocardial infarction; OAC, oral anticoagulant; PCI, percutaneous coronary intervention; PTCA, percutaneous transluminal coronary angioplasty; SCr, serum creatinine; STEMI, ST‐elevation myocardial infarction; VHD, valvular heart disease; WBC, white blood cell count.

Per agreement among the authors, we defined urgent catheterization by >50% of participants in each study undergoing emergent procedure while nonurgent meant conversely.

2.5. Data extraction

A standardized data collection form was used to obtain the following information from each study: title, name of first author, year of study, year of publication, country of origin, number of participants, demographic data of participants, method used to identify cases and controls, method used to diagnose the outcomes of interest (CIN), average duration of follow‐up, adjusted and unadjusted risk ratios and their corresponding 95% CI, and list of confounders that were adjusted for multivariate analysis. To ensure accuracy, all investigators independently performed this data extraction process. Any data discrepancy was resolved by referring back to the original articles.

2.6. Statistical analysis

Meta‐analysis of the combined data was performed using a random‐effect, generic inverse variance method of DerSimonian and Laird.18 The heterogeneity of effect size estimates across these studies was quantified using the I ² statistic and Q statistic. For the Q statistic, substantial heterogeneity was defined as P < 0.10. The I ² statistic ranges in value from 0% to 100% (I ² < 25%, low heterogeneity; I ² = 25%‐50%, moderate heterogeneity; and I ² > 50%, substantial heterogeneity).19 A sequential exclusion strategy, as described by Patsopoulos and colleagues,20 was used to examine whether the overall estimates were influenced by the substantial heterogeneity observed. We sequentially and cumulatively excluded studies that accounted for the largest share of heterogeneity until I ² was less than 50%. We then examined whether relative risk estimates were consistent. In accordance with Cochrane, publication bias was assessed using funnel plot. Funnel plot asymmetry was further confirmed with Egger test if there were more than 10 available studies.21 All analysis was performed using STATA version 14.1 (StataCorp. 2016. Stata Statistical Software: Release Stata/SE 14.1 for Mac. College Station, TX: StataCorp).

3. RESULTS

3.1. Description of included studies

Our search strategy yielded 37 potentially relevant articles (15 articles from EMBASE and 22 articles from MEDLINE). After exclusion of 5 duplicated articles, 32 articles underwent title and abstract review. One was excluded at this stage since all participants undergoing surgery. Two were excluded because both were abstracts. One was excluded since it was a review article, leaving 28 articles for full‐length article review. At this stage, 20 studies were excluded, which 14 of them did not report incidence of CIN among AF patients. Three studies were excluded because of an unclear definition of CIN. Two studies were excluded due to the same database. Lastly, one study was excluded due to lack of baseline AF. Therefore, retrospective and prospective cohort studies with 1,030 AF and 15,661 without AF patients were included in this meta‐analysis (Figure 1). The clinical characteristics of included studies are described in Table 1.

3.2. Meta‐analysis result

A total of eight studies (four prospective and four retrospective cohort studies) with 16,691 participants were included in the meta‐analysis. The prevalence of AF ranged from 2.75% to 14.84%. There was a substantial association between AF and CIN in patients undergone catheterization (odd ratio [OR] = 2.17, 95% CI: 1.50‐3.14, P < 0.001) with substantial heterogeneity (I ² = 54.1%) (Figure 2A).

Figure 2.

A, Forest plot of studies comparing the occurrence of contrast‐induced nephropathy in patients with and without atrial fibrillation horizontal lines represent the 95% CIs with marker size reflecting the statistical weight of the study using random‐effects model. A diamond data marker represents the overall adjusted OR and 95% CI for the outcome of interest. B, Subgroup analyses by the adjustment vs no adjustment horizontal lines represent the 95% CIs with marker size reflecting the statistical weight of the study using random‐effects model. A diamond data marker represents the overall adjusted OR and 95% CI for the outcome of interest. C, Subgroup analyses by the urgency of catheterization and adjustment horizontal lines represent the 95% CIs with marker size reflecting the statistical weight of the study using random‐effects model. A diamond data marker represents the overall adjusted OR and 95% CI for the outcome of interest. AF, atrial fibrillation; CAD, coronary artery disease; CI, confidence interval; CIN, contrast‐induced nephropathy

Since we were aware that the overall result was derived from both adjusted and nonadjusted OR, we therefore examined and have found that both were still statistically associated with CIN (adjusted: OR = 3.00, CI: 1.77‐5.08, P < 0.001, I ² = 47.9% and nonadjusted: OR = 1.81, CI: 1.44‐2.27, P < 0.001, I ² = 35.6%). Our results still showed that both groups were associated with the occurrence of CIN (shown in Figure 2B). Moreover, to evaluate the effect of adjusted studies on overall outcome, we then conducted meta‐regression between multivariate adjusted and nonadjusted studies. We found no significant difference among both groups (P = 0.123).

Concerning different indications for catheterization among studies, we performed subgroup analyses to determine whether urgent or nonurgent catheterization could affect different outcomes. Both groups demonstrated the same trend, which urgent group showed higher risk of CIN (OR = 2.44, 95% CI: 1.66‐3.60, P < 0.001 with I ² = 51.7%) compared to the nonurgent group (OR = 1.73, 95% CI: 1.14‐2.63, P = 0.01 with I ² = 43.3%) (shown in Figure 2C).

To explore other sources of heterogeneity in our studies, other possible factors which may affect our studies including demographics data, and the definition of CIN were further analyzed. The results were all statistically significant for the occurrence of CIN (Supporting Information Data S1).

Publication bias was not found from Egger test (P > 0.05 for all) and funnel plot (Figure 3). Sensitivity analysis to explore heterogeneity showed no significant change in our findings when each study was separately omitted.

Figure 3.

Funnel plot of AF and CIN. Circles represent published studies. AF, atrial fibrillation; CIN, contrast‐induced nephropathy

4. DISCUSSION

Our meta‐analysis demonstrated that the presence of AF in patients undergoing cardiac catheterization with or without stent/angioplasty was significantly associated with an increased risk of CIN. We also found the prevalence of AF ranging from 2.75% to 14.84%. Additionally, our subgroup analyses remained statistically significant still.

Considering substantial heterogeneity in our study, we further explored the reasons behind. First of all, the combination of both adjusted and unadjusted OR is likely the main source of heterogeneity in our analysis. We found that either group is still related to increased risk of CIN with higher extent in adjusted group compared to the other. Meta‐regression analysis also demonstrated no significant difference among both groups. These imply that participants' characteristics of both groups were relatively similar when not taking urgency of catheterization, demographics and the definition of CIN into account. Likewise, other subgroup analyses exhibited same trends of higher susceptibility to CIN. Aside from the aforementioned, we believe there were still under‐covered factors contributing to the heterogeneity such as medications, the amount of contrast used in the procedure, stage of CKD, etc. Unfortunately, we were not able to determine these elements due to insufficient data.

There have been observational reports illustrating that CIN could lead to five times in‐hospital mortality rate,22 20%23 worsening renal function and 0.7%22 up to 7%24 requiring renal replacement therapy among sufferers. Given high burden of healthcare cost,25 risk stratifications and preventive strategies to identify high‐risk groups are highly recommended and recognized as an effective method to reduce CIN.26, 27, 28 Many CIN risk factors have been noted so far including CKD, DM, age, recent MI, CHF, female gender and concomitant medication use such as nonsteroidal anti‐inflammatory drugs.29, 30, 31, 32, 33, 34, 35 In addition to previous factors, our study also suggests that AF could play a role in the development of CIN.

It has not been clearly stated for the relationship between CIN and AF. Nevertheless, there are several possible mechanisms that were referred to. First, AF could lead to hemodynamic disturbance by virtue of rapid ventricular rates, irregular heartbeat, and loss of atrioventricular synchrony and atrial contraction.36 As a result, insufficient cardiac output might occur, further contributing to CIN, despite normal left ventricular ejection fraction. Second, patients with AF burden tend to have more expression of angiotensin‐converting enzyme levels37 as well as more angiotensin II receptor levels,38 which implies that there is more activity of renin‐angiotensin‐aldosterone cascade. These result in further damage to the kidney.39 Third, it is commonly known that AF is associated with higher inflammatory state, reflected by fibrotic changes in both myocardium and kidney.40, 41 The fibrotic state itself could lead to kidney impairment due to high injury susceptibility. Fourth, renal microemboli could result from AF and lead to CIN. In addition, patients with AF usually have more severe medical condition and have higher concomitant comorbidity which, as a result, cause these patients more vulnerable to CIN.42

To the best of our knowledge, this is the first meta‐analysis conducted to determine the association between AF in patients undergone cardiac catheterization with/without PCI and CIN. We believe that the presence of AF can be helpful for risk stratification to predict CIN among these patients.

4.1. Limitation

We recognize that there are limitations to our meta‐analysis. Studies with different methodology and demographic data were included and thus might be potential sources of heterogeneity. First, since all included studies were observational in nature, the influence of residual confounders could not be completely excluded. Second, we did not further determine whether types of AF (paroxysmal, persistent and permanent) have different impact on the outcomes of interest due to insufficient data. Likewise, the presence of AF before, during or after hospitalization was not identified. Thus, we could not asses the relationship between them and CIN. In addition, there are several CIN definitions in each study which could also be possible confounder. Patients are also from different indication and urgency to undergo PCI. However, we conducted subanalyses to determine any difference in the outcome of interest which demonstrated the same tendency for the occurrence of CIN. Lastly, both adjusted and nonadjusted OR were included in our study which might result in inconclusive endpoint. Therefore, we performed the subgroup analysis which, albeit, demonstrated the same trend for CIN as well.

4.2. Conclusion

In summary, the presence of AF in patients undergoing cardiac catheterization is associated with CIN. According to our results, it is worthwhile to add the presence of AF as a part of risk stratification tool to predict CIN.

Author contribution

Narut Prasitlumkum performed the conception design, data interpretation, and draft manuscript corresponding, Veraprapas Kittipibul performed data acquisition and draft manuscript; Chanavuth Kanitsoraphan performed data acquisition and draft manuscript; Pattara Rattanawong performed data acquisition and statistical analysis; Pakawat Chongsathidkiet performed data acquisition; Wisit Cheungpasitporn performed data interpretation and critical reading.

CONFLICTS OF INTEREST

The authors declare no potential conflict of interests.

Supporting information

Appendix S1

Data S1 Subgroups analyses of the association between atrial fibrillation and contrast‐induced nephropathy (CIN) and search term

Table S1 Meta‐analysis of observational studies in epidemiology (MOOSE) checklist for meta‐analyses of observational studies

Table S2 Newcastle‐Ottawa quality assessment scale of included studies in meta‐analysis

Prasitlumkum N, Kanitsoraphan C, Kittipibul V, Rattanawong P, Chongsathidkiet P, Cheungpasitporn W. Baseline atrial fibrillation is associated with contrast‐induced nephropathy after cardiac catheterization in coronary artery disease: Systemic review and meta‐analysis. Clin Cardiol. 2018;41:1555–1562. 10.1002/clc.23100