Can the effectiveness of tonsillectomy for PFAPA syndrome be predicted based on clinical factors.

Abstract

To evaluate the clinical factors associated with the outcome of tonsillectomy in children with periodic fever, aphthous stomatitis, pharyngitis, and cervical adenitis (PFAPA) syndrome, thereby clarifying who would most likely benefit from that surgery. This was a case-control study of 53 PFAPA patients who underwent tonsillectomy and were divided into a complete-resolution group and a postoperative-fever group. Logistic regression analyses were performed using 17 clinical factors as variables to identify factors associated with the surgical outcome. Hierarchical cluster analysis was also performed to evaluate for relationships between phenotypes and surgical outcomes. Thirty-nine (73.6%) patients had complete resolution after tonsillectomy. In simple logistic regression analysis, the surgical outcome showed significant positive trends with late-onset (odds ratio [OR] 7.1, P=0.02) and presence of headache (OR 6.5, P=0.01). In stepwise multiple logistic regression analysis adjusted for age at onset, presence of headache was significantly associated with complete resolution (OR 6.5, P=0.01). The complete resolution rates for each combination of headache status and age at onset were as follows: presence of headache/age at onset ≥36 months, 100% (14/14); presence of headache/age at onset <36 months, 76.9% (10/13); absence of headache/age at onset ≥36 months, 75.0% (6/8); and absence of headache/age at onset <36 months, 43.8% (7/16). In hierarchical cluster analysis, complete resolution, age at onset, and headache were in the same cluster. PFAPA patients with headache and late onset responded well to tonsillectomy. The mechanisms underlying this association may warrant further investigation.