Abstract
Objectives
We examined whether annual gout flare frequency, as a marker of cumulative inflammatory burden, predicts long-term major adverse cardiovascular events (MACEs).
Methods
Using the TriNetX Global Collaborative Network, we identified adults with a first EHR-recorded treated gout flare in 2017. Using a 12-month landmark exposure window, patients were classified as low (≤3), moderate (4–6) or high (≥7) flares/year. Modified MACE (myocardial infarction, stroke, heart failure) was followed from the landmark. Groups were compared using 1:1 propensity score matching. A sensitivity analysis applied a stricter flare definition (gout diagnosis with colchicine, corticosteroid or intra-articular injection within 0–3 days; NSAIDs excluded).
Results
Among 44 705 patients, matching produced balanced cohorts (high vs low, 13 179 pairs; moderate vs low, 9700 pairs). A graded dose–response was observed: at 7 years, MACE risk was higher in the high (RR 1.45; 95% CI 1.37–1.53) and moderate (RR 1.24; 1.15–1.33) groups vs low. Heart failure risk was elevated in both the groups (HR 1.27 [high]; 1.17 [moderate]); stroke and myocardial infarction were elevated only in the high group (7-year HR 1.28 each), with stroke significant from 3 years. Findings were consistent under the stricter sensitivity definition (7-year MACE HR 1.19 [high] and 1.12 [moderate]).
Conclusion
Annual gout flare frequency is a graded marker of sustained cardiovascular risk. Heart failure risk rises with both moderate and high burden, whereas atherothrombotic events emerge predominantly with high-flare frequency, suggesting they require greater cumulative inflammatory exposure.
Keywords: gout, major adverse cardiovascular events (MACE), heart failure, NLRP3 inflammasome, urate-lowering therapy, landmark design
Graphical abstract
Graphical Abstract.
Rheumatology key messages.
Annual gout flare frequency is a graded clinical marker of sustained long-term cardiovascular risk.
Heart failure risk rises with moderate and high flares; high burden causes atherothrombotic events.
A pragmatic threshold of >3 treated flares per year guides cardiovascular risk stratification.
Introduction
Gout is one of the most common forms of inflammatory arthritis worldwide. It is characterized by recurrent episodes of acute synovitis triggered by monosodium urate crystal deposition. Beyond joint disease, gout is increasingly recognized as a systemic inflammatory condition that clusters with cardiometabolic comorbidities, including hypertension, diabetes mellitus, chronic kidney disease and dyslipidaemia [1–4].
Population-based studies consistently report that gout is associated with higher risks of major cardiovascular events, including myocardial infarction and stroke [3–9]. Proposed mechanisms include chronic low-grade inflammation, endothelial dysfunction and oxidative stress, all of which are central to atherosclerosis [10–13]. Dual-energy CT (DECT) studies have reported coronary plaque signals consistent with putative monosodium urate deposition that are associated with a higher subsequent risk of major adverse cardiovascular events (MACEs), although the histologic identity of these vascular signals remains debated [14–16].
Recent studies have focused on gout flares as acute inflammatory triggers of cardiovascular events. Self-controlled case series analyses have reported markedly increased risks of myocardial infarction and ischaemic stroke immediately after flares, with the highest risk in the first 2 weeks and residual elevation for up to 120 days [6]. Similar findings have been reproduced in independent cohorts [7, 17].
These trigger-based studies treat flares as discrete, time-limited exposures and therefore cannot determine whether recurrent flares reflect sustained cardiovascular vulnerability beyond the post-flare window. Recurrent flares may signal a persistently active gout course and ongoing hyperuricemia-related disease propensity [3]. We therefore conducted a large real-world cohort study using a landmark design to evaluate whether gout flare frequency during a fixed 1-year window shows a graded association with long-term cardiovascular risk after that window, conceptualizing flare burden as a chronic exposure.
Methods
Data source
We performed a retrospective cohort study using the TriNetX Global Collaborative Network, a federated electronic health record (EHR) platform aggregating longitudinal, de-identified data from academic medical centres, specialty clinics and community hospitals worldwide [18].
The protocol was approved by the Institutional Review Board of Ditmanson Foundation Chia-Yi Christian Hospital (IRB No. 2025093). Informed consent was waived because analyses used aggregated, de-identified data.
Study population
Adults aged ≥18 years with a first EHR-recorded treated gout flare during calendar year 2017 were identified. Flares were defined by ICD-10-CM codes for gout (M10.* or M1A.*) plus receipt of flare-directed therapy within 7 days, including NSAIDs, colchicine, systemic corticosteroids or intra-articular injection (CPT 20610). The first treated flare in 2017 was the index date. This definition was intended to capture treated flares documented in routine care and does not necessarily capture untreated flares, flares self-managed with previously supplied colchicine or corticosteroids, over-the-counter NSAID use or flares treated outside participating health systems; a sensitivity analysis around this definition is described below.
Calendar year 2017 was prespecified to ensure a uniform 1-year exposure assessment window and to enable long-term follow-up for cardiovascular outcomes. Patients were required to have adequate baseline data before the index date. To evaluate incident outcomes, we excluded individuals with prior MACEs (heart failure, myocardial infarction or ischaemic stroke) before the landmark date. Patients with end-stage renal disease, dialysis dependence or kidney transplantation before the landmark date were also excluded.
Exposure assessment: gout flare burden
Recurrent flares were identified using subsequent gout diagnoses accompanied by flare-directed treatment. A 30-day washout period was used to separate discrete flare episodes. Flare burden was measured over a fixed 12-month window from the index date to 12 months after. Patients were categorized by annualized treated flares as low (≤3 flares/year), moderate (4–6 flares/year) and high (≥7 flares/year) These categories were prespecified a priori for clinical interpretability and to preserve adequate sample size for matched analyses. Although current management guidelines consider ≥2 flares/year a threshold for initiating urate-lowering therapy [19], our cut-offs were not intended to define treatment eligibility but to stratify further gradations of healthcare-recorded treated-flare burden.
Landmark design and follow-up
A landmark design was applied to minimize immortal time and exposure misclassification (Fig. 1). The landmark was set 12 months after the index date, marking completion of the flare assessment window. Patients experiencing a major cardiovascular event or death before the landmark were excluded. Follow-up began at the landmark and continued until outcome occurrence, death, loss to follow-up or end of available data. Outcomes were summarized at prespecified time points corresponding to 3, 5 and 7 years after the index date.
Figure 1.
Study flow diagram and analytic framework. Note: Adult patients with ≥1 treated gout flare were identified from the TriNetX Global Collaborative Network. After exclusion of those with prior MACE or ESRD, the index date was defined as the first treated gout flare in 2017. The landmark date marked the end of a 1-year flare assessment period. Patients were categorized into flare-frequency groups (low ≤3, moderate 4–6, high ≥7 per year). Two analytical approaches were applied: (1) a Cox proportional hazards model using the full cohort to estimate hazard ratios for time-to-event outcomes, and (2) 1:1 propensity score-matched cohort comparisons (high vs low [n = 13 179 per group] and moderate vs low [n = 9700 per group]) to estimate relative risks and risk differences
Outcomes
The primary outcome was incident modified MACEs, defined as a composite of acute myocardial infarction (AMI), stroke and heart failure. Cardiovascular death was not included because cause-of-death information is not reliably available in TriNetX. Secondary outcomes were the individual components. Heart failure was identified using the TriNetX-curated diagnosis concept set anchored on ICD-10-CM I50.* (heart failure), including the systolic (I50.2x), diastolic (I50.3x), combined (I50.4x) and unspecified (I50.1/I50.8/I50.9) subcategories. This outcome reflects clinical diagnoses recorded in routine care and was not adjudicated by chart review or imaging confirmation.
Covariates
Baseline covariates were assessed in the 12 months before the index date and included age, sex, race/ethnicity, hypertension, type 2 diabetes mellitus, hyperlipidaemia, obesity, smoking, baseline eGFR and chronic kidney disease, and medication use. Baseline cardiovascular medications included statins and antihypertensive therapies; antihypertensive treatment was captured using individual classes (RAAS inhibitors, diuretics, beta blockers and calcium channel blockers).
Statistical analysis
Baseline characteristics were summarized across the flare-frequency groups. Two complementary analytic approaches were used. First, Cox proportional hazards models were fitted in the full cohort to estimate hazard ratios (HRs) and 95% CIs for modified MACE and each component outcome from the landmark date. Second, 1:1 propensity score matching (PSM) was performed separately for the high vs low and moderate vs low flare-frequency comparisons using nearest-neighbour matching with a calliper of 0.1 standard deviations of the logit of the propensity score. Covariate balance after matching was assessed using standardized mean differences, with values <0.10 indicating adequate balance. In the PSM cohorts, Kaplan–Meier curves were used to depict event-free survival, and the TriNetX Compare tool estimated risk ratios (RRs) and risk differences (RDs) at prespecified follow-up time points.
Sensitivity analyses
A sensitivity analysis re-defined treated flares as a gout diagnosis with colchicine, systemic corticosteroid or intra-articular glucocorticoid injection within 0–3 days; NSAID-only episodes were excluded. Cox proportional hazards models were repeated for modified MACE and each component outcome at 3, 5 and 7 years using the same covariate specification as the primary analysis (Supplementary Fig. S1).
Results
Study population and baseline characteristics
A total of 44 705 patients with a first treated gout flare in 2017 met the inclusion criteria. Following stratification by annual flare frequency, the cohort included: low (n = 20 764), moderate (n = 9712) and high (n = 14 229) frequency groups (Table 1). To minimize confounding, 1:1 PSM was performed. The high vs low comparison yielded 13 179 matched pairs, while the moderate vs low comparison yielded 9700 matched pairs.
Table 1.
Baseline characteristics of patients with gout categorized by annual flare frequency before propensity score matching.
| Characteristics | Low flare (≤3/year) | Moderate flare (4–6/year) | High flare (≥7/year) | Total |
|---|---|---|---|---|
| Total patients (n) | 20 764 | 9712 | 14 229 | 44 705 |
| Demographics | ||||
| Age at index, mean (S.D.) | 60.5 (13) | 58.1 (13) | 56.2 (13) | 58.5 (14) |
| Male, n (%) | 17 234 (83) | 8352 (86) | 12 522 (88) | 38 108 (85) |
| Race, n (%) | ||||
| White | 13 289 (64) | 6021 (62) | 6545 (46) | 25 855 (58) |
| Black or African American | 2492 (12) | 1263 (13) | 2134 (15) | 5889 (13) |
| Asian | 2076 (10) | 1165 (12) | 2561 (18) | 5802 (13) |
| Native Hawaiian/Other Pacific Islander | 415 (2) | 194 (2) | 285 (2) | 894 (2) |
| Others/unknown | 2492 (12) | 1069 (11) | 2704 (19) | 6265 (14) |
| Comorbidities, n (%) | ||||
| Hypertension | 14 535 (70) | 6507 (67) | 9107 (64) | 30 149 (67) |
| Hyperlipidaemia | 9136 (44) | 4273 (44) | 5692 (40) | 19 101 (43) |
| Overweight/obesity | 4983 (24) | 2622 (27) | 3700 (26) | 11 305 (25) |
| Type 2 diabetes | 5606 (27) | 2331 (24) | 3273 (23) | 11 210 (25) |
| Nicotine dependence | 1869 (9) | 1068 (11) | 1707 (12) | 4644 (10) |
| Atrial fibrillation/flutter | 1453 (7) | 583 (6) | 711 (5) | 2747 (6) |
| Chronic kidney disease | 1453 (7) | 583 (6) | 996 (7) | 3032 (7) |
| Prior renal function | ||||
| eGFR (ml/min/1.73 m²), mean (S.D.) | 81 (22) | 82 (21) | 84 (22) | 82 (22) |
| Medications, n (%) | ||||
| RAAS inhibitors | 9344 (45) | 4273 (44) | 6261 (44) | 19 878 (44) |
| Statins | 8098 (39) | 3593 (37) | 5122 (36) | 16 813 (38) |
| Diuretics | 6644 (32) | 3108 (32) | 4269 (30) | 14 021 (31) |
| Beta blockers | 6852 (33) | 3108 (32) | 4126 (29) | 14 086 (32) |
| Calcium channel blockers | 5606 (27) | 2719 (28) | 4411 (31) | 12 736 (28) |
Values are presented as mean (S.D.) or no. (%). Flare-frequency groups were defined based on the number of treated gout flares during the 12 months after the index date as low (≤3 flares/year), moderate (4–6 flares/year) and high (≥7 flares/year). Baseline characteristics and medication use were assessed during the 12 months prior to the index date. The total cohort includes distinct individuals from each flare-frequency group before propensity score matching.
eGFR, estimated glomerular filtration rate; RAAS, renin–angiotensin–aldosterone system.
Post-matching baseline variables achieved excellent balance, with all standardized mean differences (SMDs) < 0.10 (Table 2 and Supplementary Table S1). In the matched cohorts, the mean age ranged from 56.9 to 58.1 years, and the population was predominantly male (86–88%). Key cardiometabolic comorbidities were similarly distributed, including hypertension (∼65–67%), hyperlipidaemia (∼40–44%) and type 2 diabetes (∼23–24%).
Table 2.
Baseline characteristics of patients with high vs low gout flare frequency after propensity score matching.
| Characteristic | High flare (≥7×/year) (n = 13 179) | Low flare (≤3×/year) (n = 13 179) | SMD |
|---|---|---|---|
| Demographics | |||
| Age at index, years | 57.2 ± 13.1 | 56.9 ± 13.3 | 0.0231 |
| Male sex | 11 545 (87.6) | 11 582 (87.9) | 0.0086 |
| White | 6536 (49.6) | 6687 (50.7) | 0.0229 |
| Black or African American | 2011 (15.3) | 1940 (14.7) | 0.0151 |
| Asian | 2035 (15.4) | 1979 (15.0) | 0.0118 |
| Native Hawaiian/Other Pacific Islander | 261 (1.9) | 255 (1.9) | 0.0033 |
| Unknown race | 2336 (17.7) | 2318 (17.6) | 0.0012 |
| Cardiometabolic comorbidities | |||
| Hypertension | 8660 (65.7) | 8532 (64.7) | 0.0204 |
| Type 2 diabetes mellitus | 3147 (23.9) | 3054 (23.2) | 0.0166 |
| Hyperlipidaemia | 5306 (40.3) | 5359 (40.7) | 0.0082 |
| Overweight/obesity | 3323 (25.2) | 3283 (24.9) | 0.0070 |
| Nicotine dependence | 1537 (11.7) | 1516 (11.5) | 0.0050 |
| Cardiac comorbidities | |||
| Atrial fibrillation/flutter | 634 (4.8) | 592 (4.5) | 0.0151 |
| Renal function | |||
| Chronic kidney disease | 899 (6.8) | 893 (6.8) | 0.0018 |
| eGFR ≤60 (ml/min/1.73 m²) | 2747 (20.8) | 2681 (20.3) | 0.0124 |
| Baseline medications | |||
| Statin use | 4850 (36.8) | 4806 (36.5) | 0.0069 |
| RAAS inhibitors | 5833 (44.3) | 5828 (44.2) | 0.0008 |
| Diuretics | 4044 (30.7) | 3991 (30.3) | 0.0087 |
| Beta blockers | 3991 (30.3) | 3896 (29.6) | 0.0157 |
| Calcium channel blockers | 4093 (31.1) | 4017 (30.5) | 0.0125 |
Values are presented as mean ± S.D. or no. (%). Baseline characteristics were assessed after 1:1 propensity score matching at the landmark time point. Standardized mean differences (SMDs) were used to assess covariate balance, with values <0.10 indicating adequate balance. Baseline medication use was defined as prescriptions recorded within 12 months prior to the index date.
Association between flare frequency and modified MACE
Analysis of individual MACE components via Cox proportional hazards models revealed distinct patterns of association for each outcome (Fig. 2). Heart failure demonstrated the most consistent component-specific association, with significant risk elevations observed in both moderate and high flare-frequency groups at all follow-up time points. In the high-frequency group, the risk was sustained through 7 years (HR 1.27; 95% CI 1.19–1.36), and even moderate flare activity was associated with a significant long-term risk of heart failure (7-year HR 1.17; 95% CI 1.08–1.26). AMI showed a later, threshold-dependent pattern in which significant risk elevation was restricted to the high-frequency group; the association became statistically significant at 5 years (HR 1.22; 95% CI 1.08–1.37) and persisted at 7 years (HR 1.28; 95% CI 1.16–1.41), with no significant association in the moderate group (7-year HR 1.09; 95% CI 0.97–1.23). Stroke risk was likewise confined to the high-frequency group but emerged earlier, becoming statistically significant as early as 3 years (HR 1.22; 95% CI 1.06–1.41) and persisting through 7 years (HR 1.28; 95% CI 1.16–1.41); the moderate group showed no significant excess risk for stroke (7-year HR 1.11; 95% CI 0.99–1.24).
Figure 2.
Association of annual gout flare frequency with risk of modified major adverse cardiovascular events and individual components. Note: Hazard ratios were estimated using Cox proportional hazards models in the full cohort from the landmark date. Hazard ratios (HRs; dots) and 95% CIs (horizontal lines) are shown for the high (≥7 flares per year) and moderate (4–6 flares per year) flare-frequency groups compared with the low flare-frequency group (≤3 flares per year; reference) at 3, 5 and 7 years of follow-up. The vertical dashed line indicates an HR of 1.0. Modified major adverse cardiovascular events (MACE) were defined as a composite of acute myocardial infarction, stroke and heart failure
Cumulative risk of modified MACE
A clear, graded association was observed between gout flare frequency and the long-term cumulative risk of modified MACE. Figure 3 presents Kaplan–Meier survival probability curves, demonstrating early and progressive separation between the groups beginning at the landmark date (log-rank P < 0.0001 for both comparisons). High-flare frequency was associated with a significantly higher risk of modified MACE compared with low flare frequency. The cumulative risk was higher at 3 years (8.62% vs 6.17%; RR 1.40; 95% CI 1.28–1.52) and continued to diverge through 7 years (18.61% vs 12.86%; RR 1.45; 95% CI 1.37–1.53) (Table 3A). Moderate flare activity (4–6 flares per year) was also associated with a sustained increase in cardiovascular risk compared with the low-frequency group, with RRs for modified MACE of 1.17 (95% CI 1.05–1.29) at 3 years, 1.27 (95% CI 1.17–1.38) at 5 years and 1.26 (95% CI 1.18–1.35) at 7 years (Table 3B).
Figure 3.
Kaplan–Meier survival probability curves for modified major adverse cardiovascular events. Note: Kaplan–Meier curves depict the survival probability (probability of remaining free from incident modified major adverse cardiovascular events [MACE]). The x-axis represents years since the index date, with the landmark point set at 12 months post-index. Curves were constructed in 1:1 propensity score-matched cohorts comparing (A) high (≥7 flares per year) vs low (≤3 flares per year) and (B) moderate (4–6 flares per year) vs low (≤3 flares per year) flare-frequency groups. Shaded bands indicate 95% CIs. P-values were determined by log-rank tests. Modified MACE was defined as a composite of acute myocardial infarction, stroke and heart failure. Time is displayed as years since the index date; follow-up begins at the 12-month landmark
Table 3.
Risk of modified major adverse cardiovascular events by gout flare frequency after propensity score matching.
| Follow-up | Events/N (high) | Events/N (low) | Risk, % (high vs low) | Risk ratio (95% CI) | Risk difference, % | P-value |
|---|---|---|---|---|---|---|
| A. High flare frequency (≥7 flares/year) vs low flare frequency (≤3 flares/year) | ||||||
| 3 years | 1136/13 179 | 813/13 179 | 8.62 vs 6.17 | 1.40 (1.28–1.52) | +2.45 | <0.0001 |
| 5 years | 1798/13 179 | 1266/13 179 | 13.64 vs 9.61 | 1.42 (1.33–1.52) | +4.04 | <0.0001 |
| 7 years | 2453/13 179 | 1695/13 179 | 18.61 vs 12.86 | 1.45 (1.37–1.53) | +5.75 | <0.0001 |
| Follow-up | Events/N (moderate) | Events/N (low) | Risk, % (moderate vs low) | Risk ratio (95% CI) | Risk difference, % | P-value |
|---|---|---|---|---|---|---|
| B. Moderate flare frequency (4–6 flares/year) vs low flare frequency (≤3 flares/year) | ||||||
| 3 years | 741/9700 | 635/9700 | 7.64 vs 6.54 | 1.17 (1.05–1.29) | +1.10 | 0.0035 |
| 5 years | 1221/9700 | 961/9700 | 12.59 vs 9.91 | 1.27 (1.17–1.38) | +2.68 | <0.0001 |
| 7 years | 1641/9700 | 1302/9700 | 16.92 vs 13.42 | 1.26 (1.18–1.35) | +3.50 | <0.0001 |
Data are shown after 1:1 propensity score matching. Risks, risk ratios and risk differences were calculated using cumulative incidence at each prespecified follow-up time point. Major adverse cardiovascular events were defined as a composite of acute myocardial infarction, stroke and heart failure (modified MACE). Follow-up began after completion of the 12-month flare assessment window (landmark design). P-values are two-sided. Follow-up time points are presented as years since the index date, with follow-up starting at the 12-month landmark.
Sensitivity analyses
Under the stricter sensitivity definition (gout diagnosis plus colchicine, systemic corticosteroid or intra-articular injection within 0–3 days; NSAID-only episodes excluded), the graded dose–response pattern for modified MACE was preserved (Supplementary Fig. S1). Modified MACE remained significantly elevated in the high-frequency group at 3 years (HR 1.24, 95% CI 1.14–1.35), 5 years (HR 1.22, 95% CI 1.14–1.30) and 7 years (HR 1.19, 95% CI 1.13–1.26), and in the moderate-frequency group at 5 years (HR 1.09, 95% CI 1.01–1.17) and 7 years (HR 1.12, 95% CI 1.05–1.19). Component-specific patterns were also preserved: heart failure was the most robust outcome, reaching statistical significance in both the moderate (HR 1.14–1.15 across time points) and high (HR 1.25–1.30) groups across all follow-up windows. AMI and stroke associations remained confined to the high-frequency group and emerged from 5 years onward (AMI: 5-year HR 1.15, 95% CI 1.01–1.31; 7-year HR 1.20, 95% CI 1.08–1.34; stroke: 5-year HR 1.18, 95% CI 1.04–1.35; 7-year HR 1.17, 95% CI 1.05–1.31). These results indicate that the main findings were not driven by NSAID-only episodes or by a permissive 7-day temporal window between the gout diagnosis and flare-directed treatment.
Discussion
In this large, multinational real-world cohort, gout flare frequency during the 12-month exposure window after the index treated flare in 2017 identified a high-flare phenotype with sustained cardiovascular vulnerability after the 12-month landmark. We observed a clear graded association between flare frequency and long-term modified MACE. In propensity score-matched analyses, compared with patients with ≤3 flares/year, those with 4–6 flares/year had a modest but consistent excess risk (RR 1.27 at 5 years and RR 1.26 at 7 years), whereas those with ≥7 flares/year had substantially higher risk (RR 1.42 at 5 years and RR 1.45 at 7 years). These findings suggest that flare frequency above three episodes per year is a practical threshold for long-term cardiovascular risk stratification in gout. These findings were preserved in a sensitivity analysis using a stricter flare definition (Supplementary Fig. S1).
Use of modified MACE warrants comment. Cardiovascular death was excluded because cause-of-death data are not reliably available in TriNetX. However, AMI, stroke and heart failure are clinically meaningful endpoints commonly used in EHR-based cardiovascular research [1, 4, 5, 20]. Consistent associations across components support the validity of this composite outcome.
In our data, risk rose once flare frequency exceeded three episodes per year, suggesting a clinically actionable threshold for risk stratification. This supports flare frequency as a practical marker of sustained cardiovascular vulnerability and may help prioritize patients for closer review of gout management and cardiovascular prevention. The highest-risk group (≥7 flares/year) likely reflects persistent high-intensity systemic inflammation with cumulative vascular consequences, placing these patients on a risk spectrum comparable to other immune-mediated inflammatory diseases recognized in cardiovascular risk stratification [12, 21]. The threshold of >3 treated flares/year should be interpreted as a pragmatic risk-stratification threshold in this EHR-based study; it does not replace the existing guideline-recommended threshold of ≥2 flares/year for initiating urate-lowering therapy [19], but rather identifies, among gout patients already managed in routine care, a subgroup with more-than-quarterly healthcare-recorded treated flares and a higher subsequent cardiovascular event burden.
Our findings extend prior trigger-based studies by shifting the focus from short-term post-flare triggering to longer-term risk stratification. While self-controlled designs demonstrate a transient increase in cardiovascular risk after flares [6, 17], our landmark approach uses flare frequency during the first year as an early clinical signal of a patient’s subsequent disease course. Importantly, the observed associations should be interpreted primarily as the long-term cardiovascular risk of a high-flare phenotype, rather than a causal effect attributable solely to inflammation accumulated during the first 12 months. The graded risk persisting over 5–7 years supports the concept that recurrent flares identify sustained vulnerability over time [1, 10, 11].
Our component-specific analysis delineates distinct risk profiles for hemodynamic vs atherothrombotic outcomes. Heart failure emerged as the most sensitive indicator, driving much of the observed association. This finding aligns with previous population studies and meta-analyses reporting a stronger link between gout and heart failure than other cardiovascular outcomes [4, 5, 9, 20]. However, the heart-failure end point should be interpreted cautiously: it was ascertained using routine EHR diagnosis codes anchored on ICD-10-CM I50.* and was not adjudicated by chart review or imaging confirmation. This definition is heterogeneous and does not distinguish heart failure with preserved vs reduced ejection fraction, acute decompensated events from chronic diagnostic coding or differences in coding intensity across health systems. Nevertheless, the consistency of the heart-failure association across all follow-up time points and across both the primary and sensitivity analyses (Supplementary Fig. S1) argues against the signal being an artefact of a single coding pattern.
Recurrent flares may repeatedly activate the NLRP3 inflammasome via MSU crystals, sustaining elevations of IL-1β, IL-6 and TNF-α and promoting endothelial dysfunction and oxidative stress [3]. Experimental data also implicate inflammatory pathways relevant to heart failure with preserved ejection fraction (HFpEF), including an ATP–P2X7–NLRP3 axis that promotes cardiomyocyte hypertrophy, interstitial fibrosis and diastolic dysfunction [22]. These effects may be amplified by coexisting chronic kidney disease, a common comorbidity in this population [23]. Furthermore, NLRP3 may directly modulate collagen synthesis and myofibroblast activity, providing a biological bridge between systemic inflammation and cardiac fibrotic remodelling [24].
In contrast, risks for stroke and myocardial infarction exhibited a threshold effect, being restricted to the high-frequency group (≥7 flares/year). However, their temporal onset differed. Stroke risk was evident as early as 3 years in the high-frequency group, whereas myocardial infarction became statistically significant only after 5 years of follow-up. This divergence suggests that while frequent flares correlate with earlier cerebrovascular risk, the association with overt coronary events may require a longer duration of cumulative exposure. Imaging evidence provides additional support: DECT-coded signals consistent with putative MSU deposition have been described within coronary plaques and vasculature and have been associated with a higher subsequent risk of MACE in cohort studies [16, 25]. Nevertheless, the histologic identity of these vascular signals remains debated, and ongoing studies are needed to confirm whether they represent true crystal deposition or imaging artefacts related to low-density vascular calcification [14, 15]. Collectively, these findings suggest that frequent flares maintain a proinflammatory environment that contributes to both earlier hemodynamic impairment (heart failure) and progressive atherothrombosis.
Clinically, incorporating flare frequency into cardiovascular risk assessment may help identify patients with a persistently high-risk phenotype who warrant intensified prevention. Given the substantial cardiometabolic comorbidity burden in gout and its association with coronary outcomes, closer cardiovascular risk assessment and optimization may be particularly important in patients with frequent flares [1, 2]. These patients may also benefit from timely optimization of urate-lowering therapy and appropriate flare prophylaxis to reduce subsequent flare burden. Colchicine, used for gout flare treatment and prophylaxis, has also been explored for cardiovascular prevention in broader cardiovascular populations [26]; however, evidence that urate- or flare-targeted strategies translate into lower cardiovascular mortality remains uncertain [27]. Integrating flare frequency into established cardiovascular risk frameworks may support multidisciplinary care consistent with recommendations for other immune-mediated inflammatory diseases [21].
This study has limitations inherent to EHR-based observational research, including residual confounding and outcome misclassification despite excellent covariate balance after matching. Several limitations relate specifically to flare ascertainment. Our EHR-based algorithm identified treated flares recorded in routine care and therefore likely underestimates the true frequency of all patient-experienced flares. Patients may self-treat using previously supplied colchicine or corticosteroids, may use over-the-counter NSAIDs without a recorded prescription, or may not seek care for milder flares. Conversely, medications used for other indications could lead to misclassification, although requiring a temporally linked gout diagnosis was intended to improve specificity. Reassuringly, a sensitivity analysis that simultaneously excluded NSAID-only episodes and narrowed the gout-diagnosis-to-treatment window from 7 to 0–3 days yielded effect estimates closely consistent with the primary analysis (Supplementary Fig. S1), suggesting that the main findings are not driven by reliance on NSAID capture or by a permissive 7-day temporal window. Flare frequency in this study should therefore be interpreted as a pragmatic clinical marker of healthcare-recorded treated-flare burden rather than a complete measure of biological inflammatory activity. A key limitation is that we did not model post-landmark flare activity or treatment changes (e.g. initiation or intensification of urate-lowering therapy, colchicine prophylaxis or corticosteroid use) as time-varying covariates. Individuals classified as having high-flare frequency during the first year are likely to continue experiencing frequent flares after the landmark and may also be more likely to undergo treatment escalation; both processes could influence cardiovascular risk. Consequently, time-varying confounding may have contributed to the observed associations, and our estimates may reflect a composite of persistent disease activity and subsequent treatment trajectories rather than an independent causal effect of first-year cumulative inflammation. In addition, serum urate values were not sufficiently complete or standardized across participating health systems to be incorporated into the primary analysis, and post-landmark urate-lowering therapy initiation, dose escalation, adherence, colchicine prophylaxis and corticosteroid exposure were not modelled as time-varying covariates. These factors may lie in the causal pathway between flare burden and cardiovascular risk or may act as time-varying confounders. Accordingly, the observed associations should be interpreted as the cardiovascular risk associated with a high-flare clinical phenotype rather than as an isolated causal effect of first-year flare burden. In addition, flare frequency captured in routine care may partly reflect relatively stable behavioural factors and healthcare-seeking patterns over time (e.g. diet, adherence, access to care), which could also help explain the consistent risk gradient. Flare ascertainment relied on coding plus treatment within 7 days and may miss untreated or mild flares, or misclassify episodes.
Cardiovascular death was not available, and regional variation in coding practices and healthcare representation may limit generalizability. Restricting the index period to 2017 may reduce external validity; however, it ensured a uniform exposure assessment window and strengthened internal validity. Strengths of this study include the large multinational cohort, robust PSM with standardized mean differences <0.10, the landmark design to reduce immortal time bias and component-specific analyses distinguishing myocardial infarction, stroke and heart failure. Future work should incorporate time-dependent approaches (e.g. time-dependent Cox models) and/or marginal structural models with inverse probability weighting to better address time-varying confounding and should incorporate longitudinal serum urate and treatment trajectories.
In conclusion, gout flare frequency appears to be a clinically useful marker of a high-risk phenotype with sustained cardiovascular vulnerability. A threshold of >3 flares/year may aid risk stratification and may prompt earlier review of gout management and cardiovascular prevention. Future interventional studies are needed to determine whether reducing flare burden can directly lower MACE incidence.
Supplementary Material
Acknowledgements
AI Disclosure: Gemini 3.0 Pro (Google LLC) was utilized solely for language polishing, grammatical correction and enhancing the clarity and flow of the manuscript’s English text. The AI tool was not used to generate original research ideas, interpret data, perform statistical analysis or draft the initial scientific content of the paper. All research findings and the final version of the text were reviewed and approved by the authors, who remain fully responsible for the accuracy and integrity of the work.
Contributor Information
Kuei-Ting Tung, Division of Nephrology, Department of Internal Medicine, Ditmanson Medical Foundation Chia-Yi Christian Hospital, Chiayi City, Taiwan.
Solomon Chih-Cheng Chen, Department of Pediatrics, Ditmanson Medical Foundation Chia-Yi Christian Hospital, Chiayi City, Taiwan; Department of Pediatrics, College of Medicine, Taipei Medical University, Taipei, Taiwan.
Chun Lee, Clinical Data Center, Ditmanson Medical Foundation Chia-Yi Christian Hospital, Chiayi City, Taiwan.
Supplementary material
Supplementary material is available at Rheumatology online.
Data availability
This study utilized de-identified data from the TriNetX Global Collaborative Network. Due to strict third-party data-use agreements and privacy protections, the authors are legally prohibited from distributing or sharing individual patient-level datasets. To facilitate reproducibility while respecting these restrictions, comprehensive details regarding study design, inclusion/exclusion criteria and statistical methodologies are provided within the manuscript. Researchers interested in accessing the raw data may apply for a licence directly through TriNetX (www.trinetx.com).
Funding
No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this article.
Disclosure statement: The authors have declared no conflicts of interest.
References
- 1. Choi HK, Curhan G. Independent impact of gout on mortality and risk for coronary heart disease. Circulation 2007;116:894–900. [DOI] [PubMed] [Google Scholar]
- 2. Kuo CF, Grainge MJ, Mallen C, Zhang W, Doherty M. Comorbidities in patients with gout prior to and following diagnosis: case-control study. Ann Rheum Dis 2016;75:210–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Burnier M. Gout and hyperuricaemia: modifiable cardiovascular risk factors? Front Cardiovasc Med 2023;10:1190069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Colantonio LD, Saag KG, Singh JA et al. Gout is associated with an increased risk for incident heart failure among older adults: the REasons for Geographic And Racial Differences in Stroke (REGARDS) cohort study. Arthritis Res Ther 2020;22:86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Seminog OO, Goldacre MJ. Gout as a risk factor for myocardial infarction and stroke in England: evidence from record linkage studies. Rheumatology 2013;52:2251–9. [DOI] [PubMed] [Google Scholar]
- 6. Cipolletta E, Tata LJ, Nakafero G et al. Association between gout flare and subsequent cardiovascular events among patients with gout. JAMA 2022;328:440–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Lopez D, Dwivedi G, Nossent J et al. Risk of major adverse cardiovascular event following incident hospitalization for acute gout: a Western Australian population‐level linked data study. ACR Open Rheumatol 2023;5:298–304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Ferguson LD, Molenberghs G, Verbeke G et al. Gout and incidence of 12 cardiovascular diseases: a case-control study including 152 663 individuals with gout and 709 981 matched controls. Lancet Rheumatol 2024;6:e156–67. [DOI] [PubMed] [Google Scholar]
- 9. Kuo C-F, Yu K-H, See L-C et al. Risk of myocardial infarction among patients with gout: a nationwide population-based study. Rheumatology 2013;52:111–7. [DOI] [PubMed] [Google Scholar]
- 10. Feig DI, Kang D-H, Johnson RJ. Uric acid and cardiovascular risk. N Engl J Med 2008;359:1811–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Ridker PM. Anticytokine agents: targeting interleukin signaling pathways for the treatment of atherothrombosis. Circ Res 2019;124:437–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Krishnan E. Inflammation, oxidative stress and lipids: the risk triad for atherosclerosis in gout. Rheumatology 2010;49:1229–38. [DOI] [PubMed] [Google Scholar]
- 13. Libby P, Soehnlein O. Inflammation in atherosclerosis: lessons and therapeutic implications. Immunity 2025;58:2383–401. [DOI] [PubMed] [Google Scholar]
- 14. Dalbeth N, Alhilali M, Riordan P et al. Vascular deposition of monosodium urate crystals in gout: analysis of cadaveric tissue by dual‐energy computed tomography and compensated polarizing light microscopy. Arthritis Rheumatol 2022;74:1295–6. [DOI] [PubMed] [Google Scholar]
- 15. Pascart T, Budzik J-F. Does monosodium urate crystal vascular deposition exist? Review of the evidence. Gout, Urate, Crystal Deposition Dis 2023;1:208–16. [Google Scholar]
- 16. Held J, Schwabl C, Haschka D et al. Major cardiovascular events in patients with cardiovascular monosodium urate deposits in atherosclerotic plaques. Rheumatology 2025;64:1443–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Cipolletta E, Nakafero G, Richette P et al. Short‐term risk of cardiovascular events in people newly diagnosed with gout. Arthritis Rheumatol 2025;77:202–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Olaker VR, Fry S, Terebuh P et al. With big data comes big responsibility: strategies for utilizing aggregated, standardized, de‐identified electronic health record data for research. Clin Transl Sci 2025;18:e70093. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. FitzGerald JD, Dalbeth N, Mikuls T et al. 2020 American College of Rheumatology Guideline for the management of gout. Arthritis Care Res (Hoboken) 2020;72:744–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Wang P, Yang H. Risk of myocardial infarction and heart failure in gout patients: a systematic review and meta-analysis. J Cardiothorac Surg 2025;20:69. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Agca R, Smulders Y, Nurmohamed M. Cardiovascular disease risk in immune-mediated inflammatory diseases: recommendations for clinical practice. Heart 2022;108:73–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Li Z, Sun X, Wang Y et al. Renal denervation attenuates cardiac dysfunction in HFpEF by inhibiting the ATP–P2X7–NLRP3 inflammasome axis. Basic Res Cardiol 2025;120:1225–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Lee DY, Moon JS, Jung I et al. Risk acceleration by gout on major adverse cardiovascular events and all‐cause death in patients with diabetes and chronic kidney disease. Diabetes Obes Metab 2025;27:1554–63. [DOI] [PubMed] [Google Scholar]
- 24. Artlett CM. The role of the NLRP3 inflammasome in fibrosis. Open Rheumatol J 2012;6:80–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Klauser AS, Halpern EJ, Strobl S et al. Dual-energy computed tomography detection of cardiovascular monosodium urate deposits in patients with gout. JAMA Cardiol 2019;4:1019–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Punzi L, Paola G, Roberto L et al. Gout: one year in review 2023. Clin Exp Rheumatol 2024;42:1–9. [DOI] [PubMed] [Google Scholar]
- 27. Saag KG, Becker MA, White WB et al. ; CARES Investigators. Evaluation of the relationship between serum urate levels, clinical manifestations of gout, and death from cardiovascular causes in patients receiving febuxostat or allopurinol in an outcomes trial. Arthritis Rheumatol 2022;74:1593–601. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
This study utilized de-identified data from the TriNetX Global Collaborative Network. Due to strict third-party data-use agreements and privacy protections, the authors are legally prohibited from distributing or sharing individual patient-level datasets. To facilitate reproducibility while respecting these restrictions, comprehensive details regarding study design, inclusion/exclusion criteria and statistical methodologies are provided within the manuscript. Researchers interested in accessing the raw data may apply for a licence directly through TriNetX (www.trinetx.com).