Simple biomarker-based risk score for prediction of incident HF in (pre-)diabetes
Biomarker-Based Risk Prediction of Incident Heart Failure in Pre-Diabetes and DiabetesLiterature - Pandey A, Vaduganathan M, Patelet KV et al., - JACC Heart Fail. 2020 Dec 24;S2213-1779(20)30591-6. doi: 10.1016/j.jchf.2020.10.013.
Introduction and methods
Patients with diabetes and pre-diabetes are at high risk for HF [1,2]. It has been demonstrated that SGLT2i reduces the risk of HF among patients with diabetes . However, prescription rates remain low among diabetic patients with an indication for an SGLT2i . A risk stratification tool to identify patients with (pre-)diabetes at highest risk for HF could help to allocate SGLT2i to patients who would derive the greatest benefit from treatment. This study evaluated the application of a simple biomarker-based risk score for the stratification of HF risk in patients with diabetes and pre-diabetes.
Data from patients with dysglycemia (n=6,799, 33.2% with diabetes, 66.8% with pre-diabetes) free of CVD were pooled from 3 cohort studies (Atherosclerois Risk in Communities [ARIC], Dallas Heart Study [DHS], and Multi-Ethnic Study of Atherosclerosis [MESA]) [5-7]. Elevated levels of 4 cardiac and inflammatory biomarkers that were associated with higher risk of HF among healthy adults were included in an integer-based risk score: hs-cTnT ≥6 ng/l; NT-proBNP ≥125 pg/ml, hs-CRP ≥3 mg/l, and left ventricular hypertrophy by ECG, with 1 point for each abnormal biomarker [8-11]. Risk of incident HF was assessed in 4 score groups: very low (0, reference group), low (1), intermediate (2), and high (≥3). The primary outcome was incident HF. Median follow-up was 17 years. 5- and 10-year risk of incident HF was calculated for patients with diabetes and pre-diabetes across the 4 risk score groups. In addition, the number of HF events that could be prevented using a SGLT2i per 1000 treated participants over 5 and 10 years was estimated.
- Among patients with dysglycemia without CVD, elevated levels of hs-cTnT (≥6 ng/l), NT-proBNP (≥125 pg/ml), and hs-CRP (≥3 mg/l), and left ventricular hypertrophy were independently associated with higher risk of HF.
- A multivariable-adjusted Cox analysis showed that risk of HF increased with increasing biomarker score group among participants with diabetes (low score : HR 1.82, 95%CI 1.31-2.53; intermediate score : HR 2.42, 95%CI 1.71-3.43; high score [≥3]: HR 4.72, 95%CI 3.16-7.04) and pre-diabetes (low score : HR 1.40, 95%CI 1.09-1.80; intermediate score : HR 1.83, 95%CI 1.37-2.45; high score [≥3]: HR 3.68, 95%CI 2.53-5.34).
- The 5-year risk of incident HF among participants with a biomarker score of 0 or 1 was very low (diabetes: 0.85% to 1.29%; pre-diabetes: 0.49 to 0.67%) and comparable to the risk in participants with euglycemia (0.78%). The 5-year risk for incident HF increased with increasing biomarker score, with the highest risk among those with scores of ≥3 (diabetes: 12.0%; pre-diabetes: 7.8%). 10-year HF risk showed a similar pattern with increasing risk with increasing biomarker score groups.
- The estimated numbers of HF events that could be prevented with an SGLT2i for every 1000 subjects with diabetes increased with increasing biomarker score groups. Per 1000 treated participants with diabetes, treatment with SGLT2i could prevent 4 HF events in the very low score group (0), 5 HF events in the low score group (1) and 44 HF events over 5 years in the high score group (≥3). Over 10 years, 12, 20 and 103 HF events could be prevented with SGLT2i treatment in the very low (0), low (1) and high (≥3) score groups, respectively.
This study showed that a simple integer-based biomarker score could identify patients with (pre-)diabetes who are at an increased risk of incident HF. Risk of incident HF increased with increasing biomarker score group among participants with diabetes and pre-diabetes. The 5-year risk of incident HF among participants with a biomarker score of 0 or 1 was comparable to the risk in participants with euglycemia, while 5-year risk of HF in participants with a high biomarker score was 15-fold (in diabetes) and 10-fold (in pre-diabetes) higher compared with subjects with euglycemia. The estimated numbers of HF events that could be prevented with an SGLT2i for every 1000 subjects with diabetes increased with increasing biomarker score groups.