Data-Driven symptom clustering identifies four potential phenotypes of self-reported postorgasmic illness syndrome:...

Postorgasmic illness syndrome (POIS) is a rare and heterogeneous condition characterized by local and/or systemic symptoms following ejaculation. However, whether distinct clinical phenotypes exist remains unclear. We aimed to identify symptom-dominant phenotypes among men with self-reported POIS-like symptoms and to explore variations in the history of allergic diseases across the resulting phenotypes. An online cross-sectional survey was conducted via social media from December 2020 to August 2021 using the WenJuanXing web application. A two-step cluster analysis was performed using 26 variables from POIS-related symptoms and history of allergic diseases. A total of 2,443 men completed the survey, including 770 with self-reported POIS-like symptoms. A four-cluster solution (BIC = 39,404.98; ratio of distance measures = 1.42; ratio of BIC changes = 0.34) was selected. Cluster 1 represented the generalized type (81 patients, 10.5%) characterized by longer spontaneous recovery time with widespread disorders, including mental, flu-like, head, muscle symptoms and mostly reported atopic status. Cluster 2 represented the neuropsychiatric type (235 patients, 30.5%) comprised subjects with fewer atopic symptoms and predominantly mental symptoms, particularly concentration and memory difficulties. Cluster 3, or the nasal type (185 patients, 24.0%) mainly showed atopic and nasal symptoms, while cluster 4, or the oligosymptomatic type (269 patients, 34.9%) had fewer symptoms, faster symptom onset after ejaculation, shorter spontaneous recovery time, and fewer atopic symptoms. All 26 variables included in the clustering model differed significantly across clusters (P <0.001). Recruitment via social media may have introduced selection bias, potentially limiting the generalizability of these findings. The four identified phenotypes reflect the clinical variability of POIS and offer a preliminary framework for future phenotype-based investigations.