Wash, Rinse, Repeat: Moving Galectin-3 From Research to Clinical Practice

To the Editor:

Reproducibility is a cornerstone of clinical research. Whether evaluating a biomarker or therapeutic, consistent findings across populations are necessary before clinicians can trust and apply results at the bedside. B-type natriuretic peptide (BNP) provides a useful historical example. BNP was rapidly adopted for evaluation of dyspnea, yet only later did we appreciate substantial variation in its diagnostic and prognostic performance based on age, sex, race, obesity, atrial fibrillation, kidney function, and heart failure phenotype.^1,2 Experiences such as these, combined with variable findings and uncertainty about clinical utility, have contributed to the slower and less straightforward trajectory of galectin-3 from observational research to clinical practice. In this context, we appreciate the important caveats raised by Dr Dobre about our recent analysis of galectin-3 in a Veteran population and that our findings highlight the importance of reproducibility.

A key inconsistency highlighted by Dr Dobre is the lack of association between galectin-3 levels and incident heart failure (HF) in our study, in contrast to prior reports.3, 4, 5, 6 Several features of our cohort likely contributed. Our population was a bit more homogeneous than others in that it consisted primarily of male Veterans referred for echocardiography. In addition, there was also substantial clinical heterogeneity compared with prior community cohorts as individuals ranged from young Veterans evaluated for benign findings, such as murmurs or atypical chest pain, to patients with established cardiovascular disease, including prior coronary artery bypass grafting. Importantly, our findings are not completely inconsistent with prior studies as Ho et al⁴ observed the association between galectin-3 levels and incident HF was attenuated after adjustment for chronic kidney disease. Thus, we agree with Dr Dobre that this discrepancy is informative but believe it highlights an area for further research to clarify when and in whom galectin-3 may identify future HF risk and when it may not.

Another important question is whether galectin-3 provides prognostic value beyond established kidney biomarkers, particularly albuminuria. We agree that this is a major limitation of our analysis given the strong prognostic association of albuminuria with cardiovascular events. Similar arguments could be extended to established cardiac biomarkers such as BNP or troponin, and even emerging tubular injury markers.⁷ Although other studies similarly lacked albuminuria or other key biomarkers, there is some evidence to suggest galectin-3 contributes incremental information as the addition of galectin-3 levels to risk tools such as the Pooled Cohort Equation and the Atherosclerosis Risk in Communities Study HF Risk Calculator improved prediction, particularly for incident HF.⁵ This is a finding although that awaits reproducibility and still does not address concerns about residual confounding.

However, even if galectin-3 levels were attenuated after accounting for multiple biomarkers, this would not necessarily negate clinical usefulness. The study and application of biomarkers require consideration of practicality in clinical use. If it requires the simultaneous measurement of creatinine, cystatin C, albuminuria, BNP, troponin, and C-reactive protein levels to achieve the same estimated risk as what a single biomarker approximates, measuring galectin-3 levels alone may be reasonable. Additionally, galectin-3 may ultimately provide information distinct from risk prediction, such as guiding therapeutic choice or monitoring treatment response, questions that remain unanswered. These uncertainties support continued investigation of galectin-3 in prospective implementation studies similar to the Breathing Not Properly, STOP-HF (St. Vincent’s Screening to Prevent Heart Failure), and GUIDE-IT (Guiding Evidence-Based Therapy Using Biomarker-Intensified Treatment in Heart Failure).^1,8,9

Dr Dobre also provides a strong mechanistic link between galectin-3 levels and cardiac and kidney outcomes and raises whether galectin-3 should be pursued as a therapeutic target. This argument holds merit; however, many widely used biomarkers are not therapeutic targets but instead function as clinical decision triggers. Biomarkers such as BNP, troponin, lactate, C-reactive protein, or d-dimer exemplify this. They guide diagnostic testing, treatment initiation and risk stratification despite not being directly targeted therapies.^8,10,11 Indeed, attempts to use BNP itself as a treatment target, and treatment, have largely been unsuccessful.^9,12 Thus, although exploring galectin-3 as a therapeutic target is reasonable, its value as a prognostic and clinical decision biomarker should not be overlooked. Galectin-3 may help identify patients most likely to benefit from cardiorenal protective therapies such as sodium/glucose cotransporter 2 inhibitors, renin–angiotensin–aldosterone system inhibitors, and glucagon-like peptide-1 receptor agonists.

Ultimately, we agree with Dr Dobre that our findings represent history repeating itself, but repetition is necessary to confirm validity across populations. Science must “wash, rinse, and repeat.” However, galectin-3 now risks becoming trapped in that cycle. The field has largely demonstrated a consistent association between galectin-3 and cardio–kidney outcomes. What remains unknown is if this is clinically actionable. The time seems to have come to take the next step with prospective studies designed to determine (1) whether galectin-3 adds clinically meaningful information beyond established risk factors and biomarker and (2) whether modifying care based on galectin-3 levels improves patient outcomes. Trials such as Breathing Not Properly, STOP-HF, and GUIDE-IT have taught us how to implement BNP into clinical care. The evidence suggests it is time to do the same for galectin-3.