Frozen sections are unreliable for the diagnosis of necrotizing soft tissue infections

Abstract

Necrotizing soft tissue infections are rare but are associated with high rates of morbidity and mortality. The use of bedside or intraoperative frozen sections has been reported to be associated with faster diagnosis and better outcomes; however, to date no large studies have been published to determine the sensitivity and specificity of frozen sections in this setting. Twenty years of cases suspicious for necrotizing soft tissue infection at a large academic referral center were reviewed, blinded to the final clinical diagnosis (gold standard). Cases were assessed for the number of neutrophils, extent of necrosis, presence of thrombi, bacteria, karyorrhexis, and fibrin, and concordance with permanent sections. A total of 166 cases suspicious for necrotizing soft tissue infection had frozen section slides available for review. Sixty-three cases were clinically determined to be positive and 103 negative. Neutrophils, necrosis, thrombi, bacteria, karyorrhexis, and fibrin were present in both positive and negative cases; however, no histological feature or combination of features was found to be both sensitive and specific for necrotizing soft tissue infection. The combined presence of necrosis and frequent neutrophils was 73% sensitive and 68% specific, with a 58% positive predictive value and 80% negative predictive value. The additional observation of bacteria decreased sensitivity to 32%, whereas raising specificity to 91%, with 69% positive predictive value and 68% negative predictive value. Thirty-two cases (19%) contained findings identified on permanent sections (eg, bacteria) not observed on frozen section slides, highlighting the risk of false negatives owing to technical limitations or sampling errors. Frozen sections in necrotizing soft tissue infections and negative cases may show similar histological findings. Combined with the risk of false negatives, these results suggest that frozen sections are likely to be of limited clinical utility due to lack of sensitivity and specificity, and risk for delayed diagnosis and treatment.