1 2 3 4 5 6 7 Q18Q3 9 10 11Q4 12 13 14 15 16 17 18 19 20 21Q5 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54
Association between serum uric acid and atrial fibrillation: A systematic review and meta-analysis Leonardo Tamariz, MD, MPH,*† Fernando Hernandez, MD,* Aaron Bush, MD,* Ana Palacio, MD, MPH,*† Joshua M. Hare, MD*‡ From the *Department of Medicine, Miller School of Medicine, University of Miami, Miami, Florida, † Veterans Affairs Medical Center, Miami, Florida, and ‡Division of Cardiology, XXXX, XX, XX. BACKGROUND Atrial fibrillation (AF) is mediated by oxidative stress, neurohormonal activation, and inflammatory activation. Serum uric acid (SUA) is a surrogate marker of oxidative stress. Xanthine oxidase produces SUA and is upregulated by inflammation and neurohormones. OBJECTIVE To perform a meta-analysis to evaluate the evidence supporting an association between AF and SUA. METHODS We searched the MEDLINE database (1966 to July 2013) supplemented by manual searches of bibliographies of key relevant articles. We selected all cross-sectional and cohort studies in which SUA was measured and AF was reported. In cross-sectional studies, we calculated the pooled standardized mean difference of SUA between those with AF and those without AF. In cohort studies, we calculated the pooled relative risk with the corresponding 95% confidence interval (CI) for incident AF by using the random effects method. RESULTS The search strategy yielded 40 studies, of which only 9 met our eligibility criteria. The 6 cross-sectional studies
Introduction Atrial fibrillation (AF) is an important cause of morbidity and mortality. Along with the increased risk of death, AF can lead to stroke and decreased quality of life.1 The pathogenesis of AF remains incompletely understood. Uric acid is produced by xanthine oxidase (XO), is the terminal breakdown product of purine nucleotides, and is a surrogate marker of oxidative stress. Recent studies have demonstrated that there is a strong association between serum uric acid (SUA) levels with an important AF mediator (such as heart failure mortality)2 and incident coronary artery disease.3 The key pathways implicated in the development of AF are neurohormonal activation,4 oxidative stress/nitroso– redox imbalance,5 and immune activation.6 There seems to be a common mechanism linking all 3 pathways since Address reprint requests and correspondence: Dr Leonardo Tamariz, Department of Medicine, Miller School of Medicine, University of Miami, 1120 NW 14th St, Suite 967, Miami, FL 33136. E-mail address: [email protected].
comprised 7930 evaluable patients with a median prevalence of heart failure of 4% (range 0%–100%). The standardized mean Q6 difference of SUA for those with AF was 0.42 (95% CI 0.27–0.58) compared with those without AF. The 3 cohort studies evaluated Q7 138,306 individuals without AF. The relative risk of having AF for those with high SUA was 1.67 (95% CI 1.23–2.27) compared with those with normal SUA. CONCLUSION High SUA is associated with AF in both crosssectional and cohort studies. It is unclear whether SUA represents a disease marker or a treatment target. KEYWORDS Atrial fibrillation; Uric acid; Epidemiology; Oxidative stress; Meta-analysis ABBREVIATIONS AF ¼ atrial fibrillation; CI ¼ confidence interval; HR ¼ hazard ratio; IQR ¼ interquartile range; SMD ¼ standardized mean difference; SUA ¼ serum uric acid (Heart Rhythm 2014;0:-1–7) Published by Elsevier Inc. on behalf of Heart Rhythm Society.
mechanical stretch mediated by neurohormones leads to oxidative stress and inflammation upregulates XO7–9; therefore, SUA may play a role in the etiology and persistence of AF. Identifying new associations and mechanisms of AF could lead to therapeutic targets in the future. Therefore, the purpose of this meta-analysis was to help define the relationship between SUA and AF in an effort to better understand the pathophysiology of the disease.
Methods Search strategy A search was conducted through the MEDLINE database by using PubMed, which contained articles from 1966 to July 2013. This search was conducted by filtering all articles except those containing key terms such as uric acid and AF. More specifically, the search was performed by entering the following: (“uric acid”[MeSH Terms] OR (“uric”[All Fields] AND “acid”[All Fields]) OR “uric acid”[All Fields]) AND (“atrial fibrillation”[MeSH Terms] OR (“atrial”[All Fields] AND “fibrillation”[All Fields]) OR “atrial fibrillation”[All
1547-5271/$-see front matter Published by Elsevier Inc. on behalf of Heart Rhythm Society.
http://dx.doi.org/10.1016/j.hrthm.2014.04.003
55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76
2 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133
Fields]). All searches were conducted in July 2013 and were supplemented by manual searches of bibliographies of key relevant articles. We also conducted a search of EMBASE, Scopus, and CINAHL and did not identify any new studies. We did not include meeting abstracts or studies in other languages.
Selection criteria The abstract of each citation identified was reviewed by 2 investigators. When either investigator selected an article for full-text review, the full text was reviewed by 2 investigators. Agreement on whether to review the full text or include the article in the evidence table was calculated by using interrater agreement. Articles were considered for inclusion if they were cross-sectional, cohort, or case-control and reported the necessary data for mathematical pooling.
Data abstraction One investigator (A.B.) was responsible for completing the evidence table, and the second investigator (F.H.) confirmed the accuracy of the data abstracted. Differences between the 2 reviewers were resolved by consensus with L.T. Relevant baseline characteristics were reported in evidence tables. The recorded information for cross-sectional studies included demographic characteristics, relevant comorbidities, and medications that can affect uric acid, such as angiotensin receptor blockers, diuretics, and uric acid–lowering medications. For cohort studies, we also collected the SUA cutoff used, the variables used in the multivariate analysis, and the follow-up period. For mathematical pooling of cross-sectional studies, we abstracted the number of patients with their mean SUA and the corresponding SD in patients with and without AF. For cohort studies, we abstracted the number of patients who developed incident AF by the SUA cutoff level.
Definition of uric acid The key exposure variable was the SUA measurement at baseline in mg/dL. If the studies reported SUA in μmol/L, we converted those values by using the following conversion equivalence: 1 mg/dL = 59.48 μmol/L. All studies measured SUA by using the uricase-peroxidase enzymatic method. For cohort studies, we dichotomized the SUA variable. For the primary analysis, we defined hyperuricemia or high SUA as an SUA 47 mg/dL; if SUA was reported in quartiles or tertiles, we selected 47 mg/dL or the highest level quartile reported or the highest cutoff and compared it with the lowest cutoff.
Definition of AF The outcome variable of interest was the incidence of AF. AF was reported in all articles defined as either electrocardiographic recording of AF or International Classification of Diseases, Ninth Revision–based diagnosis during the follow-up period. At the same time, we abstracted event data in each of the reported SUA cutoffs.
Heart Rhythm, Vol 0, No 0, Month 2014
Quality evaluation We used the 22-item STROBE checklist.10 These items relate to the article’s title and abstract (item 1), the introduction (items 2 and 3), methods (items 4–12), results (items 13–17), and discussion (items 18–21) sections and other information (item 22 on funding). We used the 22-item appraisal for our evaluation since we included both cohort and cross-sectional studies. Eighteen items are common to the 2 designs, while 4 (items 6, 12, 14, and 15) are designspecific, with different versions for all or part of the item. Two investigators were responsible for completing the quality evaluation (A.B. and F.H.). Differences between the 2 reviewers were resolved by consensus with L.T., and we calculated interrater agreement. We assigned a score of 1 to each item if the item had been met appropriately or 0 if not and then added it to a total score. For those items that had subitems, we also assigned a score for each subitem. Therefore, the maximum score for cross-sectional studies was 32 and for cohort studies it was 33.
Statistical analysis We reported relevant baseline characteristics as median values of the reported means or percentages with the interquartile range (IQR). Because we have no patient level data, the medians reflect only the distribution of the reported data. To assess for heterogeneity across studies, we used the Cochran Q χ2 statistic (significance level of P o .010) and the I2 statistic. For the quantitative analysis, we used Stata 12 (StataCorp LP, College Station, TX) and conducted 2 different analyses depending on the study design. For cross-sectional studies, we calculated the standardized mean difference (SMD) in SUA between those with AF and those without AF. The SMD represents the difference between the weighted mean and SD of the SUA of individuals with AF and that of the controls. For cohort studies, we calculated the relative risk (RR) of AF with the respective 95% confidence intervals (CIs) and P values. For our main analysis, we categorized the data by the incidence of AF and hyperuricemia rates. We used both fixed effects and DerSimonian and Laird random effects models to calculate the pooled RR across levels of SUA. Because of heterogeneity, we elected to use the random effects model. To assess the robustness of our findings, we conducted a series of subanalyses. First, we evaluated the effects of certain variables explaining the results and heterogeneity using weighted meta-regression. To assess the effect of variables adjusted for in the statistical models, we used the number of variables and the appropriateness of the variables and we also evaluated the effect of the level of SUA used in the analysis. Second, we conducted an analysis in crosssectional studies that included only individuals who had not used uric acid–lowering medications.
Results Literature search Our search yielded 40 abstracts (Figure 1). We excluded 21 at the abstract level because they did not met our inclusion
134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 F1189 190
Tamariz et al
Uric Acid and Atrial Fibrillation
Pubmed search result 191 192 40 abstracts 193 194 21 abstracts excluded 195 196 197 10 full-text studies 19 full-text studies excluded: 198 199 W 5 no data provided for 200 E mathemacal pooling 201 B 4 reviews 202 4 9 included studies: 203 C 1 unable to replicate SUA results 204 / 6 cross-seconal F 205 P 3 cohort 206 O 207Q15 Figure 1 PubMed search strategy and results. SUA ¼ serum uric acid. 208 209 criteria and selected 19 for full-text review; of the19 selected 210 for full-text review, we included 9 studies and excluded 10 211 for the following reasons: studies did not include data that 212 could be abstracted for mathematical pooling (n ¼ 4); 213 editorials or review articles (n ¼ 4); and studies included measurements that could not be replicated 214 uric acid 11 Q8 (n ¼ 1). Interrater agreement between reviewers on inclu215 216 sion vs exclusion of abstracts was 96% (95% CI 88%–100%), 217 and interrater agreement for full-text review was 100%. The 218 reviewers did not identify citations for review from full-text references. We therefore included 6 cross-sectional 219 article 12–17 and 3 cohort studies.18–20 studies 220 221 222 Quality of the studies 223 For the 6 cross-sectional studies, we found a median score of 224 23 (IQR 18–23). The items that lowered the quality of the 225 cross-sectional studies were the lack of a clear description of 226 the measurement of SUA in 2 studies and the lack of detail 227 regarding the method to define the type of AF in all but one 228 study. All but one study conducted multivariate analysis to 229 account for confounding effects of other cardiovascular risk 230 factors. Four studies reported on the use of medications that 231 can affect SUA such as diuretics and angiotensin receptor 232 233 Table 1 Baseline characteristics of the 6 cross-sectional studies 234 No. of patients Mean SUA of patients No. of 235 Author (year) with AF with AF controls 236 86 6.2 ⫾ 1.3 48 237 Letsas et al (2010) 238 Liu et al 1253 5.41 ⫾ 1.8 2236 239 (2010) 240 Liu et al 50 6.2 ⫾ 1.6 401 (2011) 241 7.0 ⫾ 1.7 3347 242 Hu et al (2010) 125 Tekin et al 78 6.4 ⫾ 2.0 285 243 (2012) Q16 244 Kim et al 11 4.9 ⫾ 2 10 245 (2011) 246 AF ¼ atrial fibrillation; SUA ¼ serum uric acid. 247
3 blockers, but only 2 studies excluded patients on uric acid– lowering medications. For the 3 cohort studies, we found a median score of 25 (IQR 23–25). The major concerns in the cohort studies were the residual confounding in 1 study when conducting the multivariate analysis, the lack of reporting of loss to followup and the use of logistic regression as the preferred method for statistical analysis. All studies reported on diuretics and angiotensin receptor blockers but only 1 study reported on individuals that were using uric acid–lowering medications.
Baseline characteristics of the cross-sectional studies Table 1 reports the baseline characteristics of the crosssectional studies. The 6 cross-sectional studies compared 1603 individuals with AF with 6327 individuals without AF. The median age of the population was 63 years (IQR 56–67 years); the median percentage of women was 46% (IQR 44%–49%); and the median percentage of heart failure was 4% (IQR 0%–8%).
Relationship between AF and cross-sectional studies The median SUA level was 6.2 (range 5.4–6.4) for patients with AF compared with 5.1 (IQR 4.9–5.7) for those without AF. Five of the 6 studies reported SUA as a significant predictor of AF in multivariate models. Figure 2 shows the mathematical pooling of the 6 cross-sectional studies. There was significant heterogeneity (I2 statistic 62%; P ¼ .02) mainly explained by 1 study that had an SMD of 0.93. The SMD of SUA for those with AF was 0.42 (95% CI 0.27– 0.58) compared with that for those without AF (P o .01).
Baseline characteristics of cohort studies Table 2 reports the baseline characteristics of the cohort studies. The 3 cohort studies included 138,306 individuals without AF at baseline and 3466 individuals who developed AF over a median follow-up of 12 years (IQR 10–20 years). The median age of the population was 54 years (IQR 50–64 years); the median percentage of women was 41% (IQR
Mean SUA of controls
Mean age % % Heart (y) Females failure
% Hypertension
5.1 ⫾ 1.3
67
40
0
65
4.9 ⫾ 1.6
59
44
4
44
5.2 ⫾ 1.5
56
51
0
100
6.4 ⫾ 2.4 5.7 ⫾ 1.9
67 67
49 46
8 100
100 74
3.4 ⫾ 1
56
248 249 250 251 252 253 254 255 256 257 258 259 260 261 T1 262 263 264 265 266 267 268 269 270 271 272 273 274 F2275 276 277 278 279 280 281 282 T2283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304
4 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 T3 334 QF3 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361
Heart Rhythm, Vol 0, No 0, Month 2014
W E B 4 C / F P O
Figure 2
Standardized mean difference of cross-sectional studies. CI ¼ confidence interval; SMD ¼ standardized mean difference.
26%–54%); and the median percentage of heart failure was 5% (IQR 2%–6%).
Incidence of AF in cohort studies The median percentage of AF was 10% (range 2%–26%) for those with high SUA compared with 6% (range 1.7%–7%) for those with low SUA. SUA was a significant predictor of AF in multivariate models in all 3 studies; however, 1 study did not account for all the variables included (Table 3). Figure 3 shows the mathematical pooling for the 3 cohort studies. There was significant heterogeneity (I2 statistic 92%; P o .01) mainly explained by 1 study that had an RR of 3.70. The RR of having AF was 1.67 (95% CI 1.23–2.27) for those with high SUA compared with those with low SUA (P o .01). Neither the number of variables used in the multivariate analysis (P ¼ .74) nor the SUA cutoff level (P ¼ .53) explained the heterogeneity seen in cohort studies.
Subgroup analysis in specific populations One cross-sectional study reported a subgroup analysis according to the type of AF. Patients with permanent AF had a mean SUA level of 6.7 ⫾ 1.4 compared with 5.7 ⫾ 1.1 in those with paroxysmal AF. For those cross-sectional studies (n ¼ 2) that excluded patients using uric acid– Table 2
lowering therapy, we found an SMD of 0.68 (95% CI 0.45– 0.91). All 3 cohort studies reported results in subpopulations. Valbusa et al20 found no difference in the risk of AF by SUA when stratified by demographic characteristics, comorbidities, and use of medications. Chao et al19 found differences in their subgroup analysis, demonstrating a higher risk of AF in patients younger than 40 years (hazard ratio [HR] 2.5; 95% CI 1.4–4.4) and those aged 40–60 years (HR 1.3; 95% CI 1.1–1.6) with high SUA as compared with older patients with high SUA. Tamariz et al18 found a higher risk of AF in blacks with high SUA (HR 1.5; 95% CI 1.2–1.9) and in women with high SUA (HR 1.2; 95% CI 1.0–1.4).
Discussion Our study reports that high SUA is associated with AF. This association was consistently seen in both cross-sectional and cohort studies, and it was also seen with different types of analyses that included univariate, multivariate, random effects mathematical pooling, and subgroup analyses. It was also seen consistently in multiple and diverse populations. The strengths that lend weight to this conclusion include the stratified analysis of the different study designs and the large sample size.
Baseline characteristics of 3 cohort studies
Author (year)
Sample size
Mean age (y)
% Female
% Heart failure
Follow-up period (y)
Chao et al (2013)
122,524
50.6
26
6
12
6
% Hypertension
Tamariz et al (2011)
15,382
54
54
5
20
NR
Valbusa et al (2013)
400
64
41
2
10
71
AF ¼ atrial fibrillation; SUA ¼ serum uric acid.
SUA cutoff level
% AF during follow-up
47.0 o7.0 47.0 o7.0 45.0 o5.0
10 6 26 7 2 1.7
362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 Q9 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418
Tamariz et al 419 420 421 422 423Q17 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475
Table 3
Uric Acid and Atrial Fibrillation
5
Confounding factors used in multivariate analysis in cohort studies
Author (year)
Variable used in multivariate analysis
Chao et al (2013) Tamariz et al (2011)
Age, sex Age, sex, race, center, BMI, alcohol use, glucose, systolic and diastolic blood pressure, LDL, prevalent heart failure, creatinine, diuretic use, P-wave duration Age, sex, BMI, hypertension, chronic kidney disease, left ventricular hypertrophy, PR period, diuretic use, allopurinol use, Framingham AF score
Valbusa et al (2013)
AF ¼ atrial fibrillation; BMI ¼ body mass index; LDL ¼ low-density lipoprotein.
There are several limitations to this study that deserve mention. First, our search strategy was focused in identifying studies where the aim was to evaluate whether uric acid is associated with AF; this strategy could have missed articles where uric acid was merely reported as a predictor. However, we did not identify articles on our review of the references for the selected articles. Second, one of the cohort studies reported their results using a lower cutoff (5 mg/dL); therefore, when combined into the categories created for our analysis, misclassification of event data could have occurred. However, we evaluated the effect of different cutoffs by using meta-regression and the difference in cutoffs did not affect the results or explain the heterogeneity. Third, the information regarding the use of medications that can affect uric acid was inconsistently reported; however, a subgroup analysis of the cross-sectional studies found the same results as the main analysis. Fourth, the selected studies used different adjustment variables in their multivariate regression analysis, which could produce some variability in the SMD and RRs; however, by using meta-regression, we did not find that the number or type of adjustmentconfounding variables affected the results of the cohort studies. Fifth, we did not include meeting abstracts and citations in languages other than English. The rationale for not including meeting abstracts is that the values required for the analysis are usually not reported consistently. Finally, the number of articles found is small. The purpose of this study was to identify whether there is an association between SUA, an easily accessible biomarker,
W E B 4 C / F P O
Figure 3
and AF. Several biomarkers have been studied in relation to AF.21 Individual markers such as brain natriuretic peptide (HR 1.62) and C-reactive protein (HR 1.25) are associated with incident AF.22,23 Schnabel et al24 identified 10 biomarkers that cover the pathways implicated in the pathophysiology of AF and predicted the incidence of AF. Multiple risk prediction models, such as the Framingham Heart Study risk score,25 and the Atherosclerosis Risk in Communities study performed well (c statistic 0.76).26 Also, a new model from the CHARGE-AF consortium uses variables readily available in primary care settings performed similarly to the others (c statistic 0.76).27 The addition of novel biomarkers improves the performance of the risk prediction rules (c statistic 0.80).28 The results of our study are comparable to those of other biomarkers, and the cost of SUA is much smaller than that of other biomarkers; therefore, adding SUA to existing clinical prediction rules is a potential strategy for future studies. The majority of the research related to uric acid has been conducted with the goal of better understanding gout. SUA has recently emerged as a marker for cardiovascular diseases. It is well documented in animal models that higher SUA is associated with higher blood pressure; in humans, SUA is associated with incident hypertension and the use of allopurinol decreases blood pressure in children.29 In heart failure, SUA is related to all-cause mortality2; however, in the OPT-heart failure trial,30 the use of oxypurinol did not change a composite heart failure outcome. In coronary artery disease, only a modest association has been documented.3
Relative risk of atrial fibrillation in cohort studies. CI ¼ confidence interval; RR ¼ relative risk.
476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 Q10 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 Q11 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532
6 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547Q12 548 549 550 551 552 553 554 555 556 557 558 559Q13 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589
Little is known about the relationship between SUA and AF. Information about this association could help understand the complex and multifactorial mechanism of AF. We know now that the key pathways implicated in the development of AF are neurohormonal activation,4 oxidative stress/nitroso– redox imbalance,5 and immune activation.6 There seems to be a common mechanism linking all 3 pathways that explains the incidence of AF and the perpetuation of AF over time. The initial cellular damage from long-standing cardiovascular risk factors such as hypertension and diabetes associated with high angiotensin II and mechanical stretch5,31 initiates the upregulation of nicotinamide adenine dinucleotide phosphate-oxidase (NOX 2/4) forming reactive oxygen species and the accumulation of reactive oxygen species reduces nitric oxide, causing oxidative stress.32,33 Oxidative stress leads to calcium overload and decreases sodium channels, leading to electrical remodeling, and at the same time, there is structural remodeling through fibroblast proliferation, inflammation, and apoptosis.5 Both electrical and structural remodeling may lead to AF.34,35 This initial oxidative injury also increases proinflammatory molecules such as tumor necrosis factor and interleukin 1β and interleukin 6 activating both the mitogen-activated protein kinase pathway and nuclear factor κB, resulting in the release of growth factors and more proinflammatory mediators, or by the deposit of SUA crystals leading to the activation of the NOD-like receptor (NLPR-3).36–38 The increase in inflammatory factors leads to further XO activation and continued apoptosis, and the structural and electrical remodeling, continued activation of neurohormones and inflammatory factors, and XO activation creates a vicious cycle that perpetuates AF. Another potential explanation for the association between SUA and AF could be the confounded relationship between cardiovascular risk factors such as hypertension and renal disease with AF and SUA. However, most of the included cohort studies have accounted for blood pressure and creatinine, indicating that SUA could be an independent risk factor. It is unclear with our work whether SUA is a treatment target or simply a marker or mediator molecule. The mechanism previously described seems to label SUA as an endogenous danger signal of cell death, and the lack of benefit of oxypurinol in the OPT-HF study also supports this.
Conclusion Despite its limitations, this study finds a clear and strong association between AF and elevated SUA. Future studies should focus on evaluating the mechanistic relationship between SUA and AF and determining whether modifying SUA on patients at risk for AF can decrease its incidence.
References 1. Kannel WB, Wolf PA, Benjamin EJ, Levy D. Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillation: population-based estimates. Am J Cardiol 1998;82:2N–9N.
Heart Rhythm, Vol 0, No 0, Month 2014 2. Tamariz L, Harzand A, Palacio A, Verma S, Jones J, Hare J. Uric acid as a 590 predictor of all-cause mortality in heart failure: a meta-analysis. Congest Heart 591 Fail 2011;17:25–30. 592 3. Kim SY, Guevara JP, Kim KM, Choi HK, Heitjan DF, Albert DA. Hyperuricemia 593 and coronary heart disease: a systematic review and meta-analysis. Arthritis Care Res (Hoboken) 2010;62:170–180. 594 4. Tisdale JE, Borzak S, Sabbah HN, Shimoyama H, Goldstein S. Hemodynamic 595 and neurohormonal predictors and consequences of the development of atrial 596 fibrillation in dogs with chronic heart failure. J Card Fail 2006;12:747–751. 5. Youn JY, Zhang J, Zhang Y, Chen H, Liu D, Ping P, Weiss JN, Cai H. Oxidative 597 stress in atrial fibrillation: an emerging role of NADPH oxidase. J Mol Cell 598 Cardiol 2013;62:72–79. 599 6. Friedrichs K, Klinke A, Baldus S. Inflammatory pathways underlying atrial fibrillation. Trends Mol Med 2011;17:556–563. 600 7. Berry CE, Hare JM. Xanthine oxidoreductase and cardiovascular disease: 601 molecular mechanisms and pathophysiological implications. J Physiol 2004;555: 602 589–606. 8. Ghaemi-Oskouie F, Shi Y. The role of uric acid as an endogenous danger signal 603 in immunity and inflammation. Curr Rheumatol Rep 2011;13:160–166. 604 9. Shi Y, Evans JE, Rock KL. Molecular identification of a danger signal that alerts 605 the immune system to dying cells. Nature 2003;425:516–521. 10. von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP, 606 STROBE Initiative. The Strengthening the Reporting of Observational Studies in 607 Epidemiology (STROBE) statement: guidelines for reporting observational 608 studies. Ann Intern Med 2007;147:573–577. 11. Zhao QY, Yu SB, Huang H, Cui HY, Qin M, Huang T, Huang CX. Serum uric 609 acid levels correlate with atrial fibrillation in patients with chronic systolic heart 610 failure. Chin Med J (Engl) 2012;125:1708–1712. 611 12. Letsas KP, Korantzopoulos P, Filippatos GS, Mihas CC, Markou V, Gavrielatos G, Efremidis M, Sideris A, Kardaras F. Uric acid elevation in atrial fibrillation. 612 Hellenic J Cardiol 2010;51:209–213. 613 13. Liu T, Zhang X, Korantzopoulos P, Wang S, Li G. Uric acid levels and atrial 614 fibrillation in hypertensive patients. Intern Med 2011;50:799–803. 14. Liu Y, Liu H, Dong L, Chen J, Guo J. Prevalence of atrial fibrillation in 615 hospitalized patients over 40 years old: ten-year data from the People’s Hospital 616 of Peking University. Acta Cardiol 2010;65:221–224. 617 15. Hu YF, Chen SA, Yeh HI. Metabolic syndrome does not impose greater atrial fibrillation risk in elderly hypertensive patients. Acta Cardiol 2010;65:653–659. 618 16. Tekin G, Tekin YK, Erbay AR, Turhan H, Yetkin E. Serum uric acid levels are 619 associated with atrial fibrillation in patients with ischemic heart failure. 620 Angiology 2013;64:300–303. 621 17. Kim SM, Lee JH, Kim JR, Shin DG, Lee SH, Cho KH. Female patients with atrial fibrillation have increased oxidized and glycated lipoprotein properties and lower 622 apolipoprotein A-I expression in HDL. Int J Mol Med 2011;27:841–849. 623 18. Tamariz L, Agarwal S, Soliman EZ, Chamberlain AM, Prineas R, Folsom AR, 624 Ambrose M, Alonso A. Association of serum uric acid with incident atrial fibrillation (from the Atherosclerosis Risk in Communities [ARIC] study). Am J 625 Cardiol 2011;108:1272–1276. 626 19. Chao TF, Hung CL, Chen SJ, Wang KL, Chen TJ, Lin YJ, Chang SL, Lo LW, 627 Hu YF, Tuan TC, Chen SA. The association between hyperuricemia, left atrial Q14 size and new-onset atrial fibrillation. Int J Cardiol 2013;XX:XX–XX. 628 20. Valbusa F, Bertolini L, Bonapace S, Zenari L, Zoppini G, Arcaro G, Byrne CD, 629 Targher G. Relation of elevated serum uric acid levels to incidence of atrial 630 fibrillation in patients with type 2 diabetes mellitus. Am J Cardiol 2013;112: 499–504. 631 21. Rienstra M, McManus DD, Benjamin EJ. Novel risk factors for atrial fibrillation: 632 useful for risk prediction and clinical decision making? Circulation 2012;125: 633 e941–e946. 22. Patton KK, Ellinor PT, Heckbert SR, Christenson RH, DeFilippi C, Gottdiener 634 JS, Kronmal RA. N-terminal pro-B-type natriuretic peptide is a major predictor of 635 the development of atrial fibrillation: the Cardiovascular Health Study. Circu636 lation 2009;120:1768–1774. 23. Schnabel RB, Larson MG, Yamamoto JF, Kathiresan S, Rong J, Levy D, Keaney 637 JF Jr, Wang TJ, Vasan RS, Benjamin EJ. Relation of multiple inflammatory 638 biomarkers to incident atrial fibrillation. Am J Cardiol 2009;104:92–96. 639 24. Schnabel RB, Larson MG, Yamamoto JF, et al. Relations of biomarkers of distinct pathophysiological pathways and atrial fibrillation incidence in the 640 community. Circulation 2010;121:200–207. 641 25. Schnabel RB, Sullivan LM, Levy D, et al. Development of a risk score for atrial 642 fibrillation (Framingham Heart Study): a community-based cohort study. Lancet 2009;373:739–745. 643 26. Chamberlain AM, Agarwal SK, Folsom AR, Soliman EZ, Chambless LE, Crow 644 R, Ambrose M, Alonso A. A clinical risk score for atrial fibrillation in a biracial 645 prospective cohort (from the Atherosclerosis Risk in Communities [ARIC] study). Am J Cardiol 2011;107:85–91. 646
Tamariz et al 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663
Uric Acid and Atrial Fibrillation
27. Alonso A, Krijthe BP, Aspelund T, et al. Simple risk model predicts incidence of atrial fibrillation in a racially and geographically diverse population: the CHARGE-AF consortium. J Am Heart Assoc 2013;2:e000102. 28. Schnabel RB, Aspelund T, Li G, et al. Validation of an atrial fibrillation risk algorithm in whites and African Americans. Arch Intern Med 2010;170: 1909–1917. 29. Feig DI, Soletsky B, Johnson RJ. Effect of allopurinol on blood pressure of adolescents with newly diagnosed essential hypertension: a randomized trial. JAMA 2008;300:924–932. 30. Hare JM, Mangal B, Brown J, Fisher C Jr, Freudenberger R, Colucci WS, Mann DL, Liu P, Givertz MM, Schwarz RP; OPT-CHF Investigators. Impact of oxypurinol in patients with symptomatic heart failure: results of the OPT-CHF study. J Am Coll Cardiol 2008;51:2301–2309. 31. Yagi S, Akaike M, Aihara K, Ishikawa K, Iwase T, Ikeda Y, Soeki T, Yoshida S, Sumitomo-Ueda Y, Matsumoto T, Sata M. Endothelial nitric oxide synthaseindependent protective action of statin against angiotensin II-induced atrial remodeling via reduced oxidant injury. Hypertension 2010;55:918–923. 32. Tziomalos K, Hare JM. Role of xanthine oxidoreductase in cardiac nitroso-redox imbalance. Front Biosci (Landmark Ed) 2009;14:237–262.
7 33. Cappola TP, Kass DA, Nelson GS, Berger RD, Rosas GO, Kobeissi ZA, Marban E, Hare JM. Allopurinol improves myocardial efficiency in patients with idiopathic dilated cardiomyopathy. Circulation 2001;104:2407–2411. 34. Korantzopoulos P, Letsas KP, Liu T. Xanthine oxidase and uric acid in atrial fibrillation. Front Physiol 2012;3:150. 35. Korantzopoulos P, Kolettis TM, Galaris D, Goudevenos JA. The role of oxidative stress in the pathogenesis and perpetuation of atrial fibrillation. Int J Cardiol 2007;115:135–143. 36. Kanellis J, Watanabe S, Li JH, Kang DH, Li P, Nakagawa T, Wamsley A, Sheikh-Hamad D, Lan HY, Feng L, Johnson RJ. Uric acid stimulates monocyte chemoattractant protein-1 production in vascular smooth muscle cells via mitogen-activated protein kinase and cyclooxygenase-2. Hypertension 2003;41: 1287–1293. 37. Kang DH, Han L, Ouyang X, et al. Uric acid causes vascular smooth muscle cell proliferation by entering cells via a functional urate transporter. Am J Nephrol 2005;25:425–433. 38. Baldwin W, McRae S, Marek G, Wymer D, Pannu V, Baylis C, Johnson RJ, Sautin YY. Hyperuricemia as a mediator of the proinflammatory endocrine imbalance in the adipose tissue in a murine model of the metabolic syndrome. Diabetes 2011;60:1258–1269.
664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679
Association between serum uric acid and atrial fibrillation: A systematic review and meta-analysis Leonardo Tamariz, MD, MPH,*† Fernando Hernandez, MD,* Aaron Bush, MD,* Ana Palacio, MD, MPH,*† Joshua M. Hare, MD*‡ From the *Department of Medicine, Miller School of Medicine, University of Miami, Miami, Florida, † Veterans Affairs Medical Center, Miami, Florida, and ‡Division of Cardiology, XXXX, XX, XX. BACKGROUND Atrial fibrillation (AF) is mediated by oxidative stress, neurohormonal activation, and inflammatory activation. Serum uric acid (SUA) is a surrogate marker of oxidative stress. Xanthine oxidase produces SUA and is upregulated by inflammation and neurohormones. OBJECTIVE To perform a meta-analysis to evaluate the evidence supporting an association between AF and SUA. METHODS We searched the MEDLINE database (1966 to July 2013) supplemented by manual searches of bibliographies of key relevant articles. We selected all cross-sectional and cohort studies in which SUA was measured and AF was reported. In cross-sectional studies, we calculated the pooled standardized mean difference of SUA between those with AF and those without AF. In cohort studies, we calculated the pooled relative risk with the corresponding 95% confidence interval (CI) for incident AF by using the random effects method. RESULTS The search strategy yielded 40 studies, of which only 9 met our eligibility criteria. The 6 cross-sectional studies
Introduction Atrial fibrillation (AF) is an important cause of morbidity and mortality. Along with the increased risk of death, AF can lead to stroke and decreased quality of life.1 The pathogenesis of AF remains incompletely understood. Uric acid is produced by xanthine oxidase (XO), is the terminal breakdown product of purine nucleotides, and is a surrogate marker of oxidative stress. Recent studies have demonstrated that there is a strong association between serum uric acid (SUA) levels with an important AF mediator (such as heart failure mortality)2 and incident coronary artery disease.3 The key pathways implicated in the development of AF are neurohormonal activation,4 oxidative stress/nitroso– redox imbalance,5 and immune activation.6 There seems to be a common mechanism linking all 3 pathways since Address reprint requests and correspondence: Dr Leonardo Tamariz, Department of Medicine, Miller School of Medicine, University of Miami, 1120 NW 14th St, Suite 967, Miami, FL 33136. E-mail address: [email protected].
comprised 7930 evaluable patients with a median prevalence of heart failure of 4% (range 0%–100%). The standardized mean Q6 difference of SUA for those with AF was 0.42 (95% CI 0.27–0.58) compared with those without AF. The 3 cohort studies evaluated Q7 138,306 individuals without AF. The relative risk of having AF for those with high SUA was 1.67 (95% CI 1.23–2.27) compared with those with normal SUA. CONCLUSION High SUA is associated with AF in both crosssectional and cohort studies. It is unclear whether SUA represents a disease marker or a treatment target. KEYWORDS Atrial fibrillation; Uric acid; Epidemiology; Oxidative stress; Meta-analysis ABBREVIATIONS AF ¼ atrial fibrillation; CI ¼ confidence interval; HR ¼ hazard ratio; IQR ¼ interquartile range; SMD ¼ standardized mean difference; SUA ¼ serum uric acid (Heart Rhythm 2014;0:-1–7) Published by Elsevier Inc. on behalf of Heart Rhythm Society.
mechanical stretch mediated by neurohormones leads to oxidative stress and inflammation upregulates XO7–9; therefore, SUA may play a role in the etiology and persistence of AF. Identifying new associations and mechanisms of AF could lead to therapeutic targets in the future. Therefore, the purpose of this meta-analysis was to help define the relationship between SUA and AF in an effort to better understand the pathophysiology of the disease.
Methods Search strategy A search was conducted through the MEDLINE database by using PubMed, which contained articles from 1966 to July 2013. This search was conducted by filtering all articles except those containing key terms such as uric acid and AF. More specifically, the search was performed by entering the following: (“uric acid”[MeSH Terms] OR (“uric”[All Fields] AND “acid”[All Fields]) OR “uric acid”[All Fields]) AND (“atrial fibrillation”[MeSH Terms] OR (“atrial”[All Fields] AND “fibrillation”[All Fields]) OR “atrial fibrillation”[All
1547-5271/$-see front matter Published by Elsevier Inc. on behalf of Heart Rhythm Society.
http://dx.doi.org/10.1016/j.hrthm.2014.04.003
55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76
2 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133
Fields]). All searches were conducted in July 2013 and were supplemented by manual searches of bibliographies of key relevant articles. We also conducted a search of EMBASE, Scopus, and CINAHL and did not identify any new studies. We did not include meeting abstracts or studies in other languages.
Selection criteria The abstract of each citation identified was reviewed by 2 investigators. When either investigator selected an article for full-text review, the full text was reviewed by 2 investigators. Agreement on whether to review the full text or include the article in the evidence table was calculated by using interrater agreement. Articles were considered for inclusion if they were cross-sectional, cohort, or case-control and reported the necessary data for mathematical pooling.
Data abstraction One investigator (A.B.) was responsible for completing the evidence table, and the second investigator (F.H.) confirmed the accuracy of the data abstracted. Differences between the 2 reviewers were resolved by consensus with L.T. Relevant baseline characteristics were reported in evidence tables. The recorded information for cross-sectional studies included demographic characteristics, relevant comorbidities, and medications that can affect uric acid, such as angiotensin receptor blockers, diuretics, and uric acid–lowering medications. For cohort studies, we also collected the SUA cutoff used, the variables used in the multivariate analysis, and the follow-up period. For mathematical pooling of cross-sectional studies, we abstracted the number of patients with their mean SUA and the corresponding SD in patients with and without AF. For cohort studies, we abstracted the number of patients who developed incident AF by the SUA cutoff level.
Definition of uric acid The key exposure variable was the SUA measurement at baseline in mg/dL. If the studies reported SUA in μmol/L, we converted those values by using the following conversion equivalence: 1 mg/dL = 59.48 μmol/L. All studies measured SUA by using the uricase-peroxidase enzymatic method. For cohort studies, we dichotomized the SUA variable. For the primary analysis, we defined hyperuricemia or high SUA as an SUA 47 mg/dL; if SUA was reported in quartiles or tertiles, we selected 47 mg/dL or the highest level quartile reported or the highest cutoff and compared it with the lowest cutoff.
Definition of AF The outcome variable of interest was the incidence of AF. AF was reported in all articles defined as either electrocardiographic recording of AF or International Classification of Diseases, Ninth Revision–based diagnosis during the follow-up period. At the same time, we abstracted event data in each of the reported SUA cutoffs.
Heart Rhythm, Vol 0, No 0, Month 2014
Quality evaluation We used the 22-item STROBE checklist.10 These items relate to the article’s title and abstract (item 1), the introduction (items 2 and 3), methods (items 4–12), results (items 13–17), and discussion (items 18–21) sections and other information (item 22 on funding). We used the 22-item appraisal for our evaluation since we included both cohort and cross-sectional studies. Eighteen items are common to the 2 designs, while 4 (items 6, 12, 14, and 15) are designspecific, with different versions for all or part of the item. Two investigators were responsible for completing the quality evaluation (A.B. and F.H.). Differences between the 2 reviewers were resolved by consensus with L.T., and we calculated interrater agreement. We assigned a score of 1 to each item if the item had been met appropriately or 0 if not and then added it to a total score. For those items that had subitems, we also assigned a score for each subitem. Therefore, the maximum score for cross-sectional studies was 32 and for cohort studies it was 33.
Statistical analysis We reported relevant baseline characteristics as median values of the reported means or percentages with the interquartile range (IQR). Because we have no patient level data, the medians reflect only the distribution of the reported data. To assess for heterogeneity across studies, we used the Cochran Q χ2 statistic (significance level of P o .010) and the I2 statistic. For the quantitative analysis, we used Stata 12 (StataCorp LP, College Station, TX) and conducted 2 different analyses depending on the study design. For cross-sectional studies, we calculated the standardized mean difference (SMD) in SUA between those with AF and those without AF. The SMD represents the difference between the weighted mean and SD of the SUA of individuals with AF and that of the controls. For cohort studies, we calculated the relative risk (RR) of AF with the respective 95% confidence intervals (CIs) and P values. For our main analysis, we categorized the data by the incidence of AF and hyperuricemia rates. We used both fixed effects and DerSimonian and Laird random effects models to calculate the pooled RR across levels of SUA. Because of heterogeneity, we elected to use the random effects model. To assess the robustness of our findings, we conducted a series of subanalyses. First, we evaluated the effects of certain variables explaining the results and heterogeneity using weighted meta-regression. To assess the effect of variables adjusted for in the statistical models, we used the number of variables and the appropriateness of the variables and we also evaluated the effect of the level of SUA used in the analysis. Second, we conducted an analysis in crosssectional studies that included only individuals who had not used uric acid–lowering medications.
Results Literature search Our search yielded 40 abstracts (Figure 1). We excluded 21 at the abstract level because they did not met our inclusion
134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 F1189 190
Tamariz et al
Uric Acid and Atrial Fibrillation
Pubmed search result 191 192 40 abstracts 193 194 21 abstracts excluded 195 196 197 10 full-text studies 19 full-text studies excluded: 198 199 W 5 no data provided for 200 E mathemacal pooling 201 B 4 reviews 202 4 9 included studies: 203 C 1 unable to replicate SUA results 204 / 6 cross-seconal F 205 P 3 cohort 206 O 207Q15 Figure 1 PubMed search strategy and results. SUA ¼ serum uric acid. 208 209 criteria and selected 19 for full-text review; of the19 selected 210 for full-text review, we included 9 studies and excluded 10 211 for the following reasons: studies did not include data that 212 could be abstracted for mathematical pooling (n ¼ 4); 213 editorials or review articles (n ¼ 4); and studies included measurements that could not be replicated 214 uric acid 11 Q8 (n ¼ 1). Interrater agreement between reviewers on inclu215 216 sion vs exclusion of abstracts was 96% (95% CI 88%–100%), 217 and interrater agreement for full-text review was 100%. The 218 reviewers did not identify citations for review from full-text references. We therefore included 6 cross-sectional 219 article 12–17 and 3 cohort studies.18–20 studies 220 221 222 Quality of the studies 223 For the 6 cross-sectional studies, we found a median score of 224 23 (IQR 18–23). The items that lowered the quality of the 225 cross-sectional studies were the lack of a clear description of 226 the measurement of SUA in 2 studies and the lack of detail 227 regarding the method to define the type of AF in all but one 228 study. All but one study conducted multivariate analysis to 229 account for confounding effects of other cardiovascular risk 230 factors. Four studies reported on the use of medications that 231 can affect SUA such as diuretics and angiotensin receptor 232 233 Table 1 Baseline characteristics of the 6 cross-sectional studies 234 No. of patients Mean SUA of patients No. of 235 Author (year) with AF with AF controls 236 86 6.2 ⫾ 1.3 48 237 Letsas et al (2010) 238 Liu et al 1253 5.41 ⫾ 1.8 2236 239 (2010) 240 Liu et al 50 6.2 ⫾ 1.6 401 (2011) 241 7.0 ⫾ 1.7 3347 242 Hu et al (2010) 125 Tekin et al 78 6.4 ⫾ 2.0 285 243 (2012) Q16 244 Kim et al 11 4.9 ⫾ 2 10 245 (2011) 246 AF ¼ atrial fibrillation; SUA ¼ serum uric acid. 247
3 blockers, but only 2 studies excluded patients on uric acid– lowering medications. For the 3 cohort studies, we found a median score of 25 (IQR 23–25). The major concerns in the cohort studies were the residual confounding in 1 study when conducting the multivariate analysis, the lack of reporting of loss to followup and the use of logistic regression as the preferred method for statistical analysis. All studies reported on diuretics and angiotensin receptor blockers but only 1 study reported on individuals that were using uric acid–lowering medications.
Baseline characteristics of the cross-sectional studies Table 1 reports the baseline characteristics of the crosssectional studies. The 6 cross-sectional studies compared 1603 individuals with AF with 6327 individuals without AF. The median age of the population was 63 years (IQR 56–67 years); the median percentage of women was 46% (IQR 44%–49%); and the median percentage of heart failure was 4% (IQR 0%–8%).
Relationship between AF and cross-sectional studies The median SUA level was 6.2 (range 5.4–6.4) for patients with AF compared with 5.1 (IQR 4.9–5.7) for those without AF. Five of the 6 studies reported SUA as a significant predictor of AF in multivariate models. Figure 2 shows the mathematical pooling of the 6 cross-sectional studies. There was significant heterogeneity (I2 statistic 62%; P ¼ .02) mainly explained by 1 study that had an SMD of 0.93. The SMD of SUA for those with AF was 0.42 (95% CI 0.27– 0.58) compared with that for those without AF (P o .01).
Baseline characteristics of cohort studies Table 2 reports the baseline characteristics of the cohort studies. The 3 cohort studies included 138,306 individuals without AF at baseline and 3466 individuals who developed AF over a median follow-up of 12 years (IQR 10–20 years). The median age of the population was 54 years (IQR 50–64 years); the median percentage of women was 41% (IQR
Mean SUA of controls
Mean age % % Heart (y) Females failure
% Hypertension
5.1 ⫾ 1.3
67
40
0
65
4.9 ⫾ 1.6
59
44
4
44
5.2 ⫾ 1.5
56
51
0
100
6.4 ⫾ 2.4 5.7 ⫾ 1.9
67 67
49 46
8 100
100 74
3.4 ⫾ 1
56
248 249 250 251 252 253 254 255 256 257 258 259 260 261 T1 262 263 264 265 266 267 268 269 270 271 272 273 274 F2275 276 277 278 279 280 281 282 T2283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304
4 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 T3 334 QF3 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361
Heart Rhythm, Vol 0, No 0, Month 2014
W E B 4 C / F P O
Figure 2
Standardized mean difference of cross-sectional studies. CI ¼ confidence interval; SMD ¼ standardized mean difference.
26%–54%); and the median percentage of heart failure was 5% (IQR 2%–6%).
Incidence of AF in cohort studies The median percentage of AF was 10% (range 2%–26%) for those with high SUA compared with 6% (range 1.7%–7%) for those with low SUA. SUA was a significant predictor of AF in multivariate models in all 3 studies; however, 1 study did not account for all the variables included (Table 3). Figure 3 shows the mathematical pooling for the 3 cohort studies. There was significant heterogeneity (I2 statistic 92%; P o .01) mainly explained by 1 study that had an RR of 3.70. The RR of having AF was 1.67 (95% CI 1.23–2.27) for those with high SUA compared with those with low SUA (P o .01). Neither the number of variables used in the multivariate analysis (P ¼ .74) nor the SUA cutoff level (P ¼ .53) explained the heterogeneity seen in cohort studies.
Subgroup analysis in specific populations One cross-sectional study reported a subgroup analysis according to the type of AF. Patients with permanent AF had a mean SUA level of 6.7 ⫾ 1.4 compared with 5.7 ⫾ 1.1 in those with paroxysmal AF. For those cross-sectional studies (n ¼ 2) that excluded patients using uric acid– Table 2
lowering therapy, we found an SMD of 0.68 (95% CI 0.45– 0.91). All 3 cohort studies reported results in subpopulations. Valbusa et al20 found no difference in the risk of AF by SUA when stratified by demographic characteristics, comorbidities, and use of medications. Chao et al19 found differences in their subgroup analysis, demonstrating a higher risk of AF in patients younger than 40 years (hazard ratio [HR] 2.5; 95% CI 1.4–4.4) and those aged 40–60 years (HR 1.3; 95% CI 1.1–1.6) with high SUA as compared with older patients with high SUA. Tamariz et al18 found a higher risk of AF in blacks with high SUA (HR 1.5; 95% CI 1.2–1.9) and in women with high SUA (HR 1.2; 95% CI 1.0–1.4).
Discussion Our study reports that high SUA is associated with AF. This association was consistently seen in both cross-sectional and cohort studies, and it was also seen with different types of analyses that included univariate, multivariate, random effects mathematical pooling, and subgroup analyses. It was also seen consistently in multiple and diverse populations. The strengths that lend weight to this conclusion include the stratified analysis of the different study designs and the large sample size.
Baseline characteristics of 3 cohort studies
Author (year)
Sample size
Mean age (y)
% Female
% Heart failure
Follow-up period (y)
Chao et al (2013)
122,524
50.6
26
6
12
6
% Hypertension
Tamariz et al (2011)
15,382
54
54
5
20
NR
Valbusa et al (2013)
400
64
41
2
10
71
AF ¼ atrial fibrillation; SUA ¼ serum uric acid.
SUA cutoff level
% AF during follow-up
47.0 o7.0 47.0 o7.0 45.0 o5.0
10 6 26 7 2 1.7
362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 Q9 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418
Tamariz et al 419 420 421 422 423Q17 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475
Table 3
Uric Acid and Atrial Fibrillation
5
Confounding factors used in multivariate analysis in cohort studies
Author (year)
Variable used in multivariate analysis
Chao et al (2013) Tamariz et al (2011)
Age, sex Age, sex, race, center, BMI, alcohol use, glucose, systolic and diastolic blood pressure, LDL, prevalent heart failure, creatinine, diuretic use, P-wave duration Age, sex, BMI, hypertension, chronic kidney disease, left ventricular hypertrophy, PR period, diuretic use, allopurinol use, Framingham AF score
Valbusa et al (2013)
AF ¼ atrial fibrillation; BMI ¼ body mass index; LDL ¼ low-density lipoprotein.
There are several limitations to this study that deserve mention. First, our search strategy was focused in identifying studies where the aim was to evaluate whether uric acid is associated with AF; this strategy could have missed articles where uric acid was merely reported as a predictor. However, we did not identify articles on our review of the references for the selected articles. Second, one of the cohort studies reported their results using a lower cutoff (5 mg/dL); therefore, when combined into the categories created for our analysis, misclassification of event data could have occurred. However, we evaluated the effect of different cutoffs by using meta-regression and the difference in cutoffs did not affect the results or explain the heterogeneity. Third, the information regarding the use of medications that can affect uric acid was inconsistently reported; however, a subgroup analysis of the cross-sectional studies found the same results as the main analysis. Fourth, the selected studies used different adjustment variables in their multivariate regression analysis, which could produce some variability in the SMD and RRs; however, by using meta-regression, we did not find that the number or type of adjustmentconfounding variables affected the results of the cohort studies. Fifth, we did not include meeting abstracts and citations in languages other than English. The rationale for not including meeting abstracts is that the values required for the analysis are usually not reported consistently. Finally, the number of articles found is small. The purpose of this study was to identify whether there is an association between SUA, an easily accessible biomarker,
W E B 4 C / F P O
Figure 3
and AF. Several biomarkers have been studied in relation to AF.21 Individual markers such as brain natriuretic peptide (HR 1.62) and C-reactive protein (HR 1.25) are associated with incident AF.22,23 Schnabel et al24 identified 10 biomarkers that cover the pathways implicated in the pathophysiology of AF and predicted the incidence of AF. Multiple risk prediction models, such as the Framingham Heart Study risk score,25 and the Atherosclerosis Risk in Communities study performed well (c statistic 0.76).26 Also, a new model from the CHARGE-AF consortium uses variables readily available in primary care settings performed similarly to the others (c statistic 0.76).27 The addition of novel biomarkers improves the performance of the risk prediction rules (c statistic 0.80).28 The results of our study are comparable to those of other biomarkers, and the cost of SUA is much smaller than that of other biomarkers; therefore, adding SUA to existing clinical prediction rules is a potential strategy for future studies. The majority of the research related to uric acid has been conducted with the goal of better understanding gout. SUA has recently emerged as a marker for cardiovascular diseases. It is well documented in animal models that higher SUA is associated with higher blood pressure; in humans, SUA is associated with incident hypertension and the use of allopurinol decreases blood pressure in children.29 In heart failure, SUA is related to all-cause mortality2; however, in the OPT-heart failure trial,30 the use of oxypurinol did not change a composite heart failure outcome. In coronary artery disease, only a modest association has been documented.3
Relative risk of atrial fibrillation in cohort studies. CI ¼ confidence interval; RR ¼ relative risk.
476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 Q10 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 Q11 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532
6 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547Q12 548 549 550 551 552 553 554 555 556 557 558 559Q13 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589
Little is known about the relationship between SUA and AF. Information about this association could help understand the complex and multifactorial mechanism of AF. We know now that the key pathways implicated in the development of AF are neurohormonal activation,4 oxidative stress/nitroso– redox imbalance,5 and immune activation.6 There seems to be a common mechanism linking all 3 pathways that explains the incidence of AF and the perpetuation of AF over time. The initial cellular damage from long-standing cardiovascular risk factors such as hypertension and diabetes associated with high angiotensin II and mechanical stretch5,31 initiates the upregulation of nicotinamide adenine dinucleotide phosphate-oxidase (NOX 2/4) forming reactive oxygen species and the accumulation of reactive oxygen species reduces nitric oxide, causing oxidative stress.32,33 Oxidative stress leads to calcium overload and decreases sodium channels, leading to electrical remodeling, and at the same time, there is structural remodeling through fibroblast proliferation, inflammation, and apoptosis.5 Both electrical and structural remodeling may lead to AF.34,35 This initial oxidative injury also increases proinflammatory molecules such as tumor necrosis factor and interleukin 1β and interleukin 6 activating both the mitogen-activated protein kinase pathway and nuclear factor κB, resulting in the release of growth factors and more proinflammatory mediators, or by the deposit of SUA crystals leading to the activation of the NOD-like receptor (NLPR-3).36–38 The increase in inflammatory factors leads to further XO activation and continued apoptosis, and the structural and electrical remodeling, continued activation of neurohormones and inflammatory factors, and XO activation creates a vicious cycle that perpetuates AF. Another potential explanation for the association between SUA and AF could be the confounded relationship between cardiovascular risk factors such as hypertension and renal disease with AF and SUA. However, most of the included cohort studies have accounted for blood pressure and creatinine, indicating that SUA could be an independent risk factor. It is unclear with our work whether SUA is a treatment target or simply a marker or mediator molecule. The mechanism previously described seems to label SUA as an endogenous danger signal of cell death, and the lack of benefit of oxypurinol in the OPT-HF study also supports this.
Conclusion Despite its limitations, this study finds a clear and strong association between AF and elevated SUA. Future studies should focus on evaluating the mechanistic relationship between SUA and AF and determining whether modifying SUA on patients at risk for AF can decrease its incidence.
References 1. Kannel WB, Wolf PA, Benjamin EJ, Levy D. Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillation: population-based estimates. Am J Cardiol 1998;82:2N–9N.
Heart Rhythm, Vol 0, No 0, Month 2014 2. Tamariz L, Harzand A, Palacio A, Verma S, Jones J, Hare J. Uric acid as a 590 predictor of all-cause mortality in heart failure: a meta-analysis. Congest Heart 591 Fail 2011;17:25–30. 592 3. Kim SY, Guevara JP, Kim KM, Choi HK, Heitjan DF, Albert DA. Hyperuricemia 593 and coronary heart disease: a systematic review and meta-analysis. Arthritis Care Res (Hoboken) 2010;62:170–180. 594 4. Tisdale JE, Borzak S, Sabbah HN, Shimoyama H, Goldstein S. Hemodynamic 595 and neurohormonal predictors and consequences of the development of atrial 596 fibrillation in dogs with chronic heart failure. J Card Fail 2006;12:747–751. 5. Youn JY, Zhang J, Zhang Y, Chen H, Liu D, Ping P, Weiss JN, Cai H. Oxidative 597 stress in atrial fibrillation: an emerging role of NADPH oxidase. J Mol Cell 598 Cardiol 2013;62:72–79. 599 6. Friedrichs K, Klinke A, Baldus S. Inflammatory pathways underlying atrial fibrillation. Trends Mol Med 2011;17:556–563. 600 7. Berry CE, Hare JM. Xanthine oxidoreductase and cardiovascular disease: 601 molecular mechanisms and pathophysiological implications. J Physiol 2004;555: 602 589–606. 8. Ghaemi-Oskouie F, Shi Y. The role of uric acid as an endogenous danger signal 603 in immunity and inflammation. Curr Rheumatol Rep 2011;13:160–166. 604 9. Shi Y, Evans JE, Rock KL. Molecular identification of a danger signal that alerts 605 the immune system to dying cells. Nature 2003;425:516–521. 10. von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP, 606 STROBE Initiative. The Strengthening the Reporting of Observational Studies in 607 Epidemiology (STROBE) statement: guidelines for reporting observational 608 studies. Ann Intern Med 2007;147:573–577. 11. Zhao QY, Yu SB, Huang H, Cui HY, Qin M, Huang T, Huang CX. Serum uric 609 acid levels correlate with atrial fibrillation in patients with chronic systolic heart 610 failure. Chin Med J (Engl) 2012;125:1708–1712. 611 12. Letsas KP, Korantzopoulos P, Filippatos GS, Mihas CC, Markou V, Gavrielatos G, Efremidis M, Sideris A, Kardaras F. Uric acid elevation in atrial fibrillation. 612 Hellenic J Cardiol 2010;51:209–213. 613 13. Liu T, Zhang X, Korantzopoulos P, Wang S, Li G. Uric acid levels and atrial 614 fibrillation in hypertensive patients. Intern Med 2011;50:799–803. 14. Liu Y, Liu H, Dong L, Chen J, Guo J. Prevalence of atrial fibrillation in 615 hospitalized patients over 40 years old: ten-year data from the People’s Hospital 616 of Peking University. Acta Cardiol 2010;65:221–224. 617 15. Hu YF, Chen SA, Yeh HI. Metabolic syndrome does not impose greater atrial fibrillation risk in elderly hypertensive patients. Acta Cardiol 2010;65:653–659. 618 16. Tekin G, Tekin YK, Erbay AR, Turhan H, Yetkin E. Serum uric acid levels are 619 associated with atrial fibrillation in patients with ischemic heart failure. 620 Angiology 2013;64:300–303. 621 17. Kim SM, Lee JH, Kim JR, Shin DG, Lee SH, Cho KH. Female patients with atrial fibrillation have increased oxidized and glycated lipoprotein properties and lower 622 apolipoprotein A-I expression in HDL. Int J Mol Med 2011;27:841–849. 623 18. Tamariz L, Agarwal S, Soliman EZ, Chamberlain AM, Prineas R, Folsom AR, 624 Ambrose M, Alonso A. Association of serum uric acid with incident atrial fibrillation (from the Atherosclerosis Risk in Communities [ARIC] study). Am J 625 Cardiol 2011;108:1272–1276. 626 19. Chao TF, Hung CL, Chen SJ, Wang KL, Chen TJ, Lin YJ, Chang SL, Lo LW, 627 Hu YF, Tuan TC, Chen SA. The association between hyperuricemia, left atrial Q14 size and new-onset atrial fibrillation. Int J Cardiol 2013;XX:XX–XX. 628 20. Valbusa F, Bertolini L, Bonapace S, Zenari L, Zoppini G, Arcaro G, Byrne CD, 629 Targher G. Relation of elevated serum uric acid levels to incidence of atrial 630 fibrillation in patients with type 2 diabetes mellitus. Am J Cardiol 2013;112: 499–504. 631 21. Rienstra M, McManus DD, Benjamin EJ. Novel risk factors for atrial fibrillation: 632 useful for risk prediction and clinical decision making? Circulation 2012;125: 633 e941–e946. 22. Patton KK, Ellinor PT, Heckbert SR, Christenson RH, DeFilippi C, Gottdiener 634 JS, Kronmal RA. N-terminal pro-B-type natriuretic peptide is a major predictor of 635 the development of atrial fibrillation: the Cardiovascular Health Study. Circu636 lation 2009;120:1768–1774. 23. Schnabel RB, Larson MG, Yamamoto JF, Kathiresan S, Rong J, Levy D, Keaney 637 JF Jr, Wang TJ, Vasan RS, Benjamin EJ. Relation of multiple inflammatory 638 biomarkers to incident atrial fibrillation. Am J Cardiol 2009;104:92–96. 639 24. Schnabel RB, Larson MG, Yamamoto JF, et al. Relations of biomarkers of distinct pathophysiological pathways and atrial fibrillation incidence in the 640 community. Circulation 2010;121:200–207. 641 25. Schnabel RB, Sullivan LM, Levy D, et al. Development of a risk score for atrial 642 fibrillation (Framingham Heart Study): a community-based cohort study. Lancet 2009;373:739–745. 643 26. Chamberlain AM, Agarwal SK, Folsom AR, Soliman EZ, Chambless LE, Crow 644 R, Ambrose M, Alonso A. A clinical risk score for atrial fibrillation in a biracial 645 prospective cohort (from the Atherosclerosis Risk in Communities [ARIC] study). Am J Cardiol 2011;107:85–91. 646
Tamariz et al 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663
Uric Acid and Atrial Fibrillation
27. Alonso A, Krijthe BP, Aspelund T, et al. Simple risk model predicts incidence of atrial fibrillation in a racially and geographically diverse population: the CHARGE-AF consortium. J Am Heart Assoc 2013;2:e000102. 28. Schnabel RB, Aspelund T, Li G, et al. Validation of an atrial fibrillation risk algorithm in whites and African Americans. Arch Intern Med 2010;170: 1909–1917. 29. Feig DI, Soletsky B, Johnson RJ. Effect of allopurinol on blood pressure of adolescents with newly diagnosed essential hypertension: a randomized trial. JAMA 2008;300:924–932. 30. Hare JM, Mangal B, Brown J, Fisher C Jr, Freudenberger R, Colucci WS, Mann DL, Liu P, Givertz MM, Schwarz RP; OPT-CHF Investigators. Impact of oxypurinol in patients with symptomatic heart failure: results of the OPT-CHF study. J Am Coll Cardiol 2008;51:2301–2309. 31. Yagi S, Akaike M, Aihara K, Ishikawa K, Iwase T, Ikeda Y, Soeki T, Yoshida S, Sumitomo-Ueda Y, Matsumoto T, Sata M. Endothelial nitric oxide synthaseindependent protective action of statin against angiotensin II-induced atrial remodeling via reduced oxidant injury. Hypertension 2010;55:918–923. 32. Tziomalos K, Hare JM. Role of xanthine oxidoreductase in cardiac nitroso-redox imbalance. Front Biosci (Landmark Ed) 2009;14:237–262.
7 33. Cappola TP, Kass DA, Nelson GS, Berger RD, Rosas GO, Kobeissi ZA, Marban E, Hare JM. Allopurinol improves myocardial efficiency in patients with idiopathic dilated cardiomyopathy. Circulation 2001;104:2407–2411. 34. Korantzopoulos P, Letsas KP, Liu T. Xanthine oxidase and uric acid in atrial fibrillation. Front Physiol 2012;3:150. 35. Korantzopoulos P, Kolettis TM, Galaris D, Goudevenos JA. The role of oxidative stress in the pathogenesis and perpetuation of atrial fibrillation. Int J Cardiol 2007;115:135–143. 36. Kanellis J, Watanabe S, Li JH, Kang DH, Li P, Nakagawa T, Wamsley A, Sheikh-Hamad D, Lan HY, Feng L, Johnson RJ. Uric acid stimulates monocyte chemoattractant protein-1 production in vascular smooth muscle cells via mitogen-activated protein kinase and cyclooxygenase-2. Hypertension 2003;41: 1287–1293. 37. Kang DH, Han L, Ouyang X, et al. Uric acid causes vascular smooth muscle cell proliferation by entering cells via a functional urate transporter. Am J Nephrol 2005;25:425–433. 38. Baldwin W, McRae S, Marek G, Wymer D, Pannu V, Baylis C, Johnson RJ, Sautin YY. Hyperuricemia as a mediator of the proinflammatory endocrine imbalance in the adipose tissue in a murine model of the metabolic syndrome. Diabetes 2011;60:1258–1269.
664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679
