In The journal of sexual medicine
BACKGROUND : Priapism, a urologic emergency, has known associations with certain medical conditions. Many cases are idiopathic, suggesting an opportunity to identify novel risk factors.
AIM : We sought to identify medical conditions and pharmaceutical treatments that are associated with priapism using data-mining techniques.
METHODS : Using deidentified data in a large insurance claims database, we identified all men (age ≥20 years) with a diagnosis of priapism from 2003 to 2020 and matched them to cohorts of men with other diseases of male genitalia: erectile dysfunction, Peyronie disease, and premature ejaculation. All medical diagnoses and prescriptions used prior to first disease diagnosis were examined. Predictors were selected by random forest, and conditional multivariate logistic regressions were applied to assess the risks of each predictor.
OUTCOMES : We identified novel relationships of HIV and some HIV treatments with priapism and confirmed existing associations.
RESULTS : An overall 10 459 men with priapism were identified and matched 1:1 to the 3 control groups. After multivariable adjustment, men with priapism had high associations of hereditary anemias (odds ratio [OR], 3.99; 95% CI, 2.73-5.82), use of vasodilating agents (OR, 2.45; 95% CI, 2.01-2.98), use of HIV medications (OR, 1.95; 95% CI, 1.36-2.79), and use of antipsychotic medications (OR, 1.90; 95% CI, 1.52-2.38) as compared with erectile dysfunction controls. Similar patterns were noted when compared with premature ejaculation and Peyronie disease controls.
CLINICAL IMPLICATIONS : HIV and its treatment are associated with priapism, which may affect patient counseling.
STRENGTHS AND LIMITATIONS : To our knowledge, this is the first study to identify risk factors for priapism utilizing machine learning. All men in our series were commercially insured, which limits the generalizability of our findings.
CONCLUSION : Using data-mining techniques, we confirmed existing associations with priapism (eg, hemolytic anemias, antipsychotics) and identified novel relationships (eg, HIV disease and treatment).
Mulloy Evan, Li Shufeng, Belladelli Federico, Del Giudice Francesco, Glover Frank, Eisenberg Michael L
2023-Mar-06
HIV, data, mining, predicting, priapism
