This Test Outperformed SCORE2 for Gauging Heart Risk in Rheumatic Disease

Carotid-femoral pulse wave velocity (cfPWV) measurements more accurately predicted cardiovascular (CV) events in patients with autoimmune rheumatic diseases than a commonly used CV risk scoring system, researchers said.

Among 143 patients followed for a median of 7 years — incurring a total of 20 CV events — high cfPWV values yielded an area under the receiver operating characteristic curve (AUC) of 0.84 (95% CI 0.73-0.93) versus 0.56 (95% CI 0.41-0.72) for the updated Systematic Coronary Risk Evaluation (SCORE2), according to Konstantinos Triantafyllias, MD, PhD, of Johannes Gutenberg University in Mainz, Germany, and colleagues. That value for SCORE2 is barely better than a coin flip (AUC=0.5).

Sensitivity and specificity for high cfPWV values were 0.87 and 0.73, respectively, both at least numerically superior to SCORE2, for which sensitivity and specificity were 0.60 and 0.69, respectively, the group reported in RMD Open.

These results could be a boon to rheumatologists and their patients everywhere. Increased CV risk is a hallmark of many rheumatic diseases, yet the degree of that increase has been difficult to predict for individual patients. Standard CV risk tools for primary care, such as SCORE2, look at conventional risk factors such as age and blood lipids. Those factors don’t explain the roughly 50%-100% greater risk for CV events in rheumatic disease than is the case for otherwise similar people in the general population.

One factor not included in those conventional risk tools is arterial stiffness, which is what cfPWV measures. Other studies in cardiology have established it as an independent predictor of CV events and mortality. Using ultrasound to gauge the speed at which blood pressure pulses move through the body, cfPWV is thus non-invasive and relatively simple to perform.

Triantafyllias and colleagues noted that one previous study from 2016 had shown that the test was a good predictor of CV events, but it was restricted to rheumatoid arthritis (RA) patients and didn’t include sensitivity/specificity values. Moreover, it didn’t estimate patients’ risk with standard tools for comparison.

For the current study, the group examined records for 80 patients with RA, 31 with systemic sclerosis, and 32 with spondyloarthritis, all of whom had undergone cfPWV testing from 2012 to 2017. Patients were then contacted by phone in 2023 to report CV events that occurred in the interim. Events of interest were necessarily nonfatal and included myocardial infarction, stroke, peripheral arterial disease, coronary heart disease with some kind of intervention, pulmonary embolism, deep vein thrombosis, heart failure, and atrial fibrillation. Patients were urged to check their own records or call their physicians in case of uncertainty about events they may have had. (The researchers had identified 286 patients who had undergone the pulse wave testing, succeeding to make follow-up contact in just half that number.) Patients who died during the follow-up period (n=19) were excluded from the analysis, although two of those deaths apparently stemmed from CV causes. Some 80% of the sample were women and their mean age at the time of cfPWV testing was 58.

Conventional CV risk factors were somewhat elevated in those who experienced events, but not enough to distinguish them usefully from those without events. SCORE2 results yielded a mean risk prediction of 7% in the 20 with events, versus 6% in the no-events group. Some individual components of this measure were substantially higher in the events group: body mass index values averaged 2 points greater, for example, systolic blood pressure was higher by 6 mmHg, and 55% had a previous history of CV events compared with 15% for those without events. On the other hand, mean diastolic blood pressure was identical and total cholesterol was a bit lower in the events group.

But while the SCORE2 values didn’t differ significantly between the two groups (P=0.262), baseline cfPWV did: 10.06 m/s in the events group versus 8.8 m/s in the no-events group. The latter did not seem to be affected by patients’ anti-rheumatic drug regimens at the time, nor did they differ much between RA, systemic sclerosis, and spondyloarthritis.

When the endpoint was restricted to “hard” events, i.e., myocardial infarction and stroke, the AUC for cfPWV was a bit lower at 0.78 than when softer events were included. Meanwhile, AUCs for SCORE2 were poor in every subgroup.

As well, even when adjusting for standard risk factors, cfPWV remained a significant predictor of CV events — indicating that it is indeed an independent risk factor rather than just a correlate of already established factors.

The main takeaway from the study, according to Triantafyllias and colleagues, is that “cfPWV could serve as a more accurate indicator of residual risk that is not captured by [original] SCORE or SCORE2 in this population.” They added, “cfPWV may be incorporated into the assessment of CV risk in everyday clinical practice to assist CV screening and ultimately improve prognosis. However, prospective studies with a larger number of cases for different rheumatic diseases are warranted to validate these results.”

Limitations included reliance on participants’ memory of CV events, the small number of events occurring in the study, and the fact that cfPWV testing was not universal but rather for some specific clinical purpose, with the attendant possibility of selection bias. Other standard CV risk tools besides SCORE2, such as Framingham score, were not evaluated. Also, patients’ complete medication history was not available.