A prognostic tool developed to predict survival in patients with various forms of pulmonary arterial hypertension (PAH) is fairly accurate in predicting survival outcomes for many patients with PAH related to systemic sclerosis (SSc-PAH), according to a new study. However, the prognostic accuracy is less reliable for SSc-PAH patients with the highest risk of death.
That’s the conclusion of a recently published study that examined the accuracy of a prognostic equation and risk score calculator derived from the Registry to Evaluate Early and Long-Term PAH Disease Management (REVEAL) to predict survival outcomes in a specific group of patients with PAH.1 Developed from prospectively gathered data from a large cohort of patients with PAH, the REVEAL calculator is used to predict one-year survival in patients with various forms of PAH.2 The prognostic equation and risk score comprise 19 parameters identified through multivariate analysis as clinical and hemodynamic variables independently associated with survival. The prognostic equation provides an estimated prediction of one-year survival, whereas the risk score provides an integer value corresponding to predicted probability of one-year survival.
In the study, investigators examined the accuracy of the REVEAL calculator when applied specifically to patients with PAH related to SSc. According to the study’s senior author, Stephen C. Mathai, MD, MHS, associate professor of medicine in the Division of Pulmonary and Critical Care Medicine at Johns Hopkins University School of Medicine, Baltimore, specific tools to predict survival in SSc-PAH patients are lacking despite good therapies for PAH that may improve patient outcomes.
“This is the first study to examine the accuracy of survival prediction equations in SSc-PAH,” he says. “The results suggest predictors of risk may carry different weight in SSc-PAH compared [with] other forms of PAH.”
The Study
In the study, Mathai and colleagues prospectively gathered data on 292 patients with newly diagnosed SSc-PAH taken from the Johns Hopkins Pulmonary Hypertension Program (JHPHP; n=117) and Pulmonary Hypertension Assessment Recognition of Outcome in Scleroderma (PHAROS; n=176) study. Data obtained included vital signs and hemodynamic measurements, lab analyses, pulmonary test results, a six-minute walk test, computed tomography of the chest, echocardiogram reports and clinical assessment of World Health Organization Functional Class.
The investigators tabulated the REVEAL prognostic equation and risk score for each patient in this cohort to test its accuracy to predict one-year survival. One-year survival probability was calculated based on the assessment of 19 parameters for each patient at the time of and prior to the time at risk. The time at risk was defined slightly differently for the patients enrolled through the JHPHP and PHAROS registry, but basically began the date patients were diagnosed with PAH based on right heart catheterization. Death was determined by the Social Security death index and clinical records.
To calculate the REVEAL risk score, the same 19 parameters were used. Based on the predicted probability of one-year survival, patients were stratified into five risk groups: low risk (>95% survival), average risk (90–95% survival), moderately high risk (85–90% survival), high risk (70–85% survival) and very high risk (<70% survival).
Based on the predicted probability of 1-year survival, patients were stratified into 5 risk groups.
The study found an overall survival rate of 87.4% (95% confidence interval [CI], 82.9–90.7%), representing 36 deaths (14 patients were lost to follow-up before one year).
When looking at outcomes based on risk group, patients identified by either the prognostic equation or risk score as being at high risk appeared to experience the worst survival rates. The study found that although in general the REVEAL equation adequately predicts outcomes in a newly diagnosed SSc-PAH population, the model performed less well in patients at highest risk of death.
To validate the model’s predicative accuracy, the investigators assessed model fit, discrimination and calibration. Using concordance statistics (c-indexes) for both the prognostic equation and the risk score, the investigators found both demonstrated good discrimination (0.734, 95% CI, 0.652–0.816 and 0.743, 95% CI, 0.663–0.823, respectively) comparable to the model. The overall model slope indicated only a marginal fit to the model (slope of 0.707, 95% CI, 0.400–1.014; P=0.06). Less valid was the magnitude of risk assigned to a low distance on a six-minute walk test and elevated brain natriuretic peptide (BNP), and model calibration was particularly poor for the highest risk groups.
“Discrimination and calibration appear to be less accurate in patients in the highest risk groups, indicating the REVEAL prediction model may not perform adequately in separating and predicting survival in the newly diagnosed SSc-PAH patients with the high-risk features and the lowest predicted probabilities of one-year survival,” state the authors.1
Dr. Mathai says he and his colleagues think the variation from predicted survival in the high-risk patients based on the REVEAL score is likely due to the greater emphasis placed on two measures in the study, the six-minute walk distance and blood levels of N-terminal pro brain natriuretic peptide. “Previous research by our group has demonstrated that both of these measurements have different prognostic relevance in SSc-PAH patients compared to PAH patients with other forms of the disease,” he says.
Based on these findings, Dr. Mathai underscored the usefulness of the REVEAL equation to predict outcomes in SSc patients with newly diagnosed PAH, but cautioned it may not accurately reflect the risk in patients at the highest risk of worse outcomes.
“It is important for providers and SSc-PAH patients to discuss prognosis on an individual basis to inform treatment plans and goals of care,” he says.
Mary Beth Nierengarten is a freelance medical journalist based in Minneapolis.
References
- Mullin CJ, Khair RM, Damico RL, et al. Validation of the REVEAL prognostic equation and risk score calculator in incident systemic sclerosis-associated pulmonary arterial hypertension. Arthritis Rheumatol. 2019 May 8. [Epub ahead of print]
- Benza RL, Miller DP, Gombert-Maitland M, et al. Predicting survival in pulmonary arterial hypertension: insights from the Registry to Evaluate Early and Long-Term Pulmonary Arterial Hypertension Disease Management (REVEAL). Circulation. 2010 Jul 13;122(2):164–172.