Preoperative serum pattern analysis to predict the outcome of tonsillectomy in adults with chronic tonsillitis

The method of classification and tree analysis (CART) was used to predict the outcome of tonsillectomy for chronic tonsillitis (CHT) analyzing patterns of serological markers. In a prospective case study of 24 adult patients with CHT in comparison to 24 patients with acute peritonsillar abscess (PTA) blood samples were assessed 1 day before (T-1) and 3 days after tonsillectomy. Outcome 6 months later (T180) was documented using the Glasgow Benefit Inventory (GBI) and the Specific Benefits from Tonsillectomy Inventory (SBTI). In comparison to PTA, patients with CHT were at best classified by C-reactive protein with a cut-off value of <16.735 mg/dl. For CHT, immunoglobulin E ≤ 144.65 kU/l and the combination of monocytes ≤ 0.565 Gpt/l plus leucocytes >5.855 Gpt/l at T-1 were the best classificators for higher SBTI overall score and symptom score symptom score, respectively, at T180. A higher benefit subscore at T180 was associated to γ-globulin >15.85 % plus α2-globulin >8.950% at T-1. The best classificator for better GBI overall score at T180 was an ASL titer >169.0 IU/ml or the combination of an ASL titer ≤ 169.0 IU/ml with lymphocytes ≤ 2.195 Gpt/l. Lymphocytes ≤ 2.195 Gpt/l were associated with higher GBI general subscore. Leukocytes ≤ 6.780 Gpt/l were related to higher GBI social support subscore. The combination of immunoglobulin A >1.360 g/l with procalcitonin level >0.058 ng/ml was the best combination to classify for higher physical health score. Instead of looking on isolated serologic markers, CART of multiple parameters seems to be more effective to predict the outcome of tonsillectomy for CHT.