Background
Recent studies have attempted to verify the predictive capability of the Laboratory Risk Indicator for Necrotizing Fasciitis (LRINEC) for cervical necrotizing fasciitis (CNF).
Methods
The two authors independently reviewed six databases (PubMed, Cochrane, Embase, Web of Science, SCOPUS and Google Scholar databases were searched). Seven cohort studies were included in the analysis. Sensitivity and specificity were determined by extracting items from binary classification from each paper. The diagnostic accuracy of the included studies was evaluated using QUADAS ver. 2.
Results
The diagnostic odds ratio (OR) of LRINEC for CNF with the cutoff value of 6 was 13.9952 (95% CI, 3.8537; 50.8255, I2 = 76.7%). The area under the SROC curve was 0.842, suggesting acceptable diagnostic accuracy. The correlation between sensitivity and false positive rate was 0.055, indicating that it was not heterogeneous. The sensitivity, specificity, and negative predictive values were 0.7503 ([0.4637; 0.9126], I2 = 79.1%), 0.8455 ([0.7084; 0.9250], I2 = 96.0%) and 0.9829 ([0.9089; 0.9970], I2 = 93.7%), respectively. In the comparison of subgroups according to a LRINEC score (6–8, and), the cutoff value of 6 showed moderate sensitivity (75%) and high specificity (85%) and greater diagnostic power than other cutoff values.
Conclusions
LRINEC is a useful adjunctive tool for predicting CNF in patients with a soft tissue infection. In addition, a more accurate diagnosis is possible by using the LRINEC score with a cutoff value of 6.