Models to predict lupus based on genetics, lifestyle perform ‘very well in all situations’
Click Here to Manage Email Alerts
WASHINGTON — A model predicting the development of systemic lupus erythematosus, based on genetic, lifestyle and environmental factors, was successfully validated in a large cohort, according to data presented at ACR Convergence 2024.
“I have been interested in the possibility of identifying people who are developing lupus or at risk for lupus, and identifying modifiable risk factors among them,” Karen Costenbader, MD, MPH, professor of medicine at Harvard Medical School, and lupus program director at Brigham and Women’s Hospital, in Boston, told Healio. “We developed the first lupus prediction models based on genetics, lifestyle and environmental risk factors among women in the Nurses’ Health Study, published in 2023.
“It had pretty good overall predictive ability, separating those who did from those who didn’t develop lupus with an area under the curve (AUC) of about 0.76 for all lupus and 0.82 for the subtype of dsDNA-autoantibody-positive lupus,” she added. “However, the models needed to be validated in another population, and we also sought to extend them to people who were not female and not of European ancestry.”
To accomplish this, Costenbader and colleagues conducted a nested case-control study using patient data from the Mass General Brigham Biobank. The study included 535 patients with SLE (mean age at diagnosis, 32.6 years; 90% women) matched to 2,135 patients without SLE (mean age at match, 32.6 years; 90% women).
The researchers compared the performance of three models. The first model was based on a weighted genetic risk score, while the second considered ethnicity, smoking status, obesity and other demographic and lifestyle factors. The third model was a combination of the other two models.
Overall, the models “performed very well in all situations,” Costenbader said. Among women, the genetic risk model yielded an AUC of 0.69 (95% CI, 0.67-0.72), and the demographic and lifestyle model yielded AUC of 0.74 (95% CI, 0.72-0.77), according to the researchers.
Meanwhile, the combination model demonstrated the best performance among women (AUC = 0.82; 95% CI, 0.79-0.84) and among both sexes (AUC = 0.81; 95% CI, 0.78-0.83), the researchers wrote.
“The models and data still need to be cleaned up a bit, but, frankly, I was surprised at how well they really performed in this external population,” Costenbader said. “They performed even better in most cases than they did in the initial Nurses’ Health Study population.
“This validation of the models is very exciting and, since genetics are now available for more and more people, it might be feasible in the future to use these models in clinic, or even in family members of those with lupus,” she added. “It is really exciting to think that we might be able to identify people who are at the highest risk of lupus before it takes hold. This opens the door to new possibilities, such as identifying high-risk people, educating them and trying new strategies and interventions to prevent lupus onset.”
Collapse
Koopman JJE. Abstract #0843. Presented at: ACR Convergence 2024; Nov. 14-19, 2024; Washington, D.C.
Read more about
Click Here to Manage Email Alerts