Cardiology · Medical Literature
Dynamic Urinary Albumin/Creatinine Ratio Patterns Predict Adverse Outcomes in HFpEF: TOPCAT Cohort Analysis.
Abstract
Urinary albumin-to-creatinine ratio (uACR) is a recognized cardiovascular risk marker, but its dynamic changes in heart failure with preserved ejection fraction (HFpEF) remain underexplored. We aimed to identify distinct uACR trajectory patterns and their associations with clinical outcomes in HFpEF patients. We analyzed 746 HFpEF patients with ≥ 3 uACR measurements from the TOPCAT trial. Latent Class Trajectory Modeling identified uACR patterns. The primary outcome was cardiovascular death, aborted cardiac arrest, or heart failure hospitalization. Cox proportional hazards models assessed associations between trajectories and outcomes. Three uACR trajectory patterns emerged: stable (70.5%), decreasing (16.5%), and fluctuating high (13.0%). The primary endpoint occurred in 15.02%, 25.20%, and 36.08% of patients, respectively (p < 0.001). In adjusted models, the fluctuating high group showed increased risk for the composite endpoint (HR 1.91, 95% CI 1.25-2.91) and heart failure hospitalization (HR 1.91, 95% CI 1.19-3.06), while the decreasing group showed an association with myocardial infarction based on few events (HR 3.00, 95% CI 1.33-6.76). Adding trajectory information produced a modest improvement in risk prediction (C-index 0.758 to 0.765; AUC 0.786 to 0.792 at 4 years and 0.832 to 0.845 at 6 years, p = 0.0149). This study first identified distinct uACR trajectory patterns in HFpEF patients. The fluctuating high trajectory was associated with increased cardiovascular risk and heart failure hospitalization, while the association between the decreasing trajectory and myocardial infarction was exploratory and requires validation.
Clinical relevance
In patients with HFpEF, do distinct longitudinal urinary albumin-to-creatinine ratio (uACR) trajectory patterns — identified via latent class trajectory modeling — predict the composite outcome of cardiovascular death, aborted cardiac arrest, or heart failure hospitalization? __SNAPSHOT__ {"studyDesign":"Secondary analysis (post hoc cohort analysis) of the TOPCAT randomized controlled trial using Latent Class Trajectory Modeling to define uACR subgroups, with Cox proportional hazards models for outcome associations","setting":"Multicenter; TOPCAT trial sites (United States, Canada, Brazil, Argentina, Russia, Georgia); specific subsite distribution for this subsample not reported","sampleSize":"n = 746 HFpEF patients with ≥3 uACR measurements from the TOPCAT trial (subgroup allocation by trajectory: stable ~526, decreasing ~123, fluctuating high ~97 [calculated from reported percentages])","population":"Adults with HFpEF enrolled in the TOPCAT trial who had at least three serial uACR measurements available; detailed inclusion/exclusion criteria for this subsample beyond ≥3 measurements not reported in the abstract","intervention":"Identification of uACR trajectory class (stable, decreasing, or fluctuating high) as the primary exposure variable","comparator":"Stable uACR trajectory group (70.5% of cohort) used as the reference category in adjusted Cox models","primaryOutcome":"Composite of cardiovascular death, aborted cardiac arrest, or heart failure hospitalization","followUp":"Not reported in the abstract (aligned with TOPCAT trial follow-up, which was approximately 3.3 years in the parent trial, but subsample follow-up duration is not specified)","fundingSource":"Not reported in the abstract"}
Key findings
Latent Class Trajectory Modeling identified three distinct uACR patterns among 746 HFpEF patients: stable (70.5%), decreasing (16.5%), and fluctuating high (13.0%). The primary composite endpoint occurred in 15.02%, 25.20%, and 36.08% of patients in these groups, respectively (p < 0.001). In adjusted Cox proportional hazards models, the fluctuating high trajectory was associated with a significantly increased risk of the composite endpoint (HR 1.91, 95% CI 1.25–2.91) and heart failure hospitalization specifically (HR 1.91, 95% CI 1.19–3.06), both relative to the stable trajectory. The decreasing trajectory was associated with myocardial infarction risk (HR 3.00, 95% CI 1.33–6.76), though the abstract explicitly cautions that this finding was based on few events and should be interpreted with commensurate uncertainty. Absolute risk difference between stable and fluctuating high groups for the composite primary endpoint was 21.06 percentage points (calculated). NNH for composite outcome — fluctuating high vs. stable — not calculable from the abstract alone without event-time data. The abstract is truncated at the point of reporting incremental model performance (C-statistic or reclassification metrics for addition of trajectory to the base model), and those values are therefore not reported here.
Practice implications
This secondary analysis provides hypothesis-generating evidence that serially measured uACR trajectories — rather than single-point values — may carry differential prognostic information in HFpEF. The fluctuating high pattern, present in approximately 1 in 8 patients, was associated with nearly twice the adjusted risk of the composite cardiovascular outcome and of HF hospitalization compared with the stable pattern, a difference that may be clinically meaningful if replicated. The decreasing trajectory's association with myocardial infarction (HR 3.00) is intriguing but severely constrained by the acknowledged small number of events; this estimate should be treated as exploratory and hypothesis-generating only. Important limitations of applicability include: (1) this analysis is restricted to TOPCAT enrollees who had ≥3 uACR measurements, a selected minority of the parent trial and a potential source of significant selection bias; (2) TOPCAT's well-documented regional heterogeneity (Americas vs. Russia/Georgia) and concerns about trial fidelity in the latter region are not addressed in the abstract; (3) how the trajectory model performs in contemporary HFpEF populations receiving SGLT2 inhibitors, finerenone, or GLP-1 receptor agonists — none of which were used in TOPCAT — is unknown; (4) the abstract is truncated, preventing evaluation of whether trajectory assignment improved discrimination or reclassification beyond established risk markers. Because only the abstract is available, the full covariate adjustment strategy, handling of missing data, sensitivity analyses, and C-statistic results cannot be evaluated. Current evidence does not support routine serial uACR trajectory monitoring as a clinical decision-making tool in HFpEF; confirmation in prospective, contemporary cohorts is required. __LIMITATIONS__ Post hoc secondary analysis of an RCT not designed to test this hypothesis; trajectory classification is inherently exploratory and data-driven (latent class models require pre-specification of class number, which may overfit small samples).,The analytic cohort (n = 746) represents a highly selected fraction of the TOPCAT parent trial (n = 3,445), limited to patients with ≥3 uACR measurements; this may systematically exclude the sickest or least engaged patients, limiting generalizability.,TOPCAT's well-documented regional integrity concerns (substantial differences in event rates and suspected non-adherence in Russia/Georgia) are not addressed in the abstract; the composition of this subsample by region is not reported.,The myocardial infarction association in the decreasing trajectory group is explicitly acknowledged by the authors as based on few events (exact n not reported), rendering the HR 3.00 estimate highly unreliable and subject to extreme instability.,The abstract is truncated at the point of reporting incremental model statistics; it cannot be determined whether uACR trajectory improves discrimination (C-statistic) or net reclassification beyond established clinical predictors.,TOPCAT enrolled patients without contemporary guideline-directed therapies for HFpEF (SGLT2 inhibitors, finerenone, GLP-1 receptor agonists), substantially limiting applicability to current clinical practice.,Follow-up duration for this subsample is not reported in the abstract.,Funding source is not reported, precluding assessment of potential conflicts of interest.,Causal inference is not possible; this design establishes association, not causation, between uACR trajectory and outcomes. __BOTTOM_LINE_CATEGORY__ promising_but_preliminary __BOTTOM_LINE__ In this post hoc analysis of 746 TOPCAT HFpEF patients, a fluctuating high uACR trajectory (present in ~13%) was associated with approximately twice the adjusted risk of cardiovascular death, cardiac arrest, or HF hospitalization compared with a stable trajectory. These findings are hypothesis-generating only and are substantially constrained by selection bias, the TOPCAT regional integrity concerns, the absence of contemporary HFpEF therapies, and a truncated abstract that precludes full evaluation of model performance. Serial uACR trajectory assessment should not yet influence HFpEF management; prospective validation in contemporary, well-characterized cohorts is needed.