OR WAIT null SECS
Combining algorithms SCORE2 and QRISK3 created the best predictive model to predict the cardiovascular risk for patients with PsA.
A new study found 4 risk chart algorithms help identify low and high cardiovascular (CV) risk in patients with psoriatic arthritis (PsA).1
This study, led by Jessica Polo y La Borda, from Hospital Universitario Rey Juan Carlos in Spain, marked the first time comparing the strengths and limitations of 4 separate risk scales. Investigators aimed to see how well the scales predicted CV events and mortality, such as from ischemic heart disease, stroke, peripheral arterial disease, and heart failure.
“In addition to several versions of the European SCORE, the algorithm traditionally used to determine CV risk in the Spanish population, we also analyzed in our cohort the QRISK3, used in the UK to evaluate CV risk,” investigators wrote.
CV mortality and events have a greater prevalence in patients with PsA than in the general population.2 However, research underscored the need for improved algorithms to predict CV risk in patients with PsA, since the Systematic Coronary Risk Assessment (SCORE) often underestimated the CV risk.
The team leveraged patients with PsA from the Spanish prospective project CARdiovascular in RheuMAtology who had no history of CV events at the baseline visit and visited rheumatology outpatient clinics at tertiary centers for 7.5 years.1 The team aimed to predict the patients’ CV risk with the algorithms Systematic Coronary Risk Assessment (SCORE), the modified version of the European Alliance of Rheumatology Associations (mSCORE EULAR 2015/2016), the SCORE2, and the QResearch risk estimator version 3 (QRISK3).
The SCORE calculation included age, sex, smoking, systolic blood pressure, and total cholesterol. SCORE2 calculated the CV risk with nearly all the same factors, although it included non-HDL cholesterol instead of total cholesterol. SCORE evaluated only the 10-year risk of death from CV disease, but SCORE2 estimates the 10-year risk of fatal and non-fatal CV disease events for people aged 40 – 69 years. For people aged ≥ 70 years, SCORE2 estimates the 5-year risk of death from CV disease and the 10-year risk of non-fatal CV event.
Furthermore, QRISK3 estimates an individual’s 10-year risk of having a heart attack or stroke. It calculates conventional CV risk factors, diabetes mellitus, chronic kidney disease, emerging CV risk factors, and other inflammatory rheumatic diseases. As for mSCORE EULAR, the scale followed the recommendations from 2015/2016.
The 4 CV risk algorithms demonstrated significant associations with CV events (P < .001), with the scales predicting risks of 1.07 (95% confidence interval [CI], 1.05 – 1.08) for QRISK3, 1.06 (95% CI, 1.04 – 1.09) for SCORE, 1.04 (95% CI, 1.03 – 1.06) for mSCORE EULAR 2015/2016, and 1.17 (95% CI, 1.12 to 1.23) for SCORE2. SCORE, EULAR, and QRISK3 could effectively differentiate between low and high CV risk patients.
After 7.5 years, 34 CV events occurred, with a CV linearized rate of 7.10 per 1000 person-years (95% confidence interval [CI], 4.92 – 9.92). The cumulative rate of CV events observed was lower than predicted by the risk scales. However, the team noted that since it was a 7.5-year follow-up, and they were estimating 10-year risks, this may have resulted in finding fewer CV events than predicted.
The team found the best predictive model was combining QRISK3 with any other scale, specifically QRISK3 with SCORE2, creating the lowest Akaike information criterion (411.15) and Bayesian information criterion (420.10). On its own, SCORE2 was better at estimating the CV risk than the other algorithms.
“This observation suggests that non-HDL cholesterol might play a crucial role in influencing CV events in PsA, given the prevalent metabolic risk in these patients,” investigators wrote. “However, we feel that this insight may be extrapolated to the general population.”
Despite SCORE2 doing well in predicting the risk, none of the 4 algorithms provided all the necessary information about the CV risk, which is why the team found combining QRISK3 with SCORE2 yielded the best results.
“The integration of QRISK3 and SCORE2 in a comprehensive model demonstrated an optimal combination, leveraging QRISK3’s discrimination capability and SCORE2’s calibration accuracy,” Investigators concluded.
References