Accuracy of ICD-9 codes for identifying children with cerebral sinovenous thrombosis

Childhood sinovenous thrombosis is rare, making it difficult to study; International Classification of Diseases, ninth revision (ICD-9), code searches across multiple hospitals would permit the identification of large numbers of children with sinovenous thrombosis. However, the accuracy of these codes for identifying childhood sinovenous thrombosis has not been established. We per-formed a retrospective search of admissions records for Riley Hospital for Children in Indianapolis, Indiana, from January 1999 to June 2005 using ICD-9 codes 325 (cerebral sinovenous thrombosis, excluding nonpyogenic cases and cases associated with pregnancy and the puerperium), 437.6 (cerebral venous thrombosis of nonpyogenic origin), and 671.5 (cerebral venous thrombosis in pregnancy or the puerperium) in any position. During this period, there were 47 042 admissions. ICD-9 code 325 identified 61 admissions on 56 children. Only 13% were of pyogenic ori gin. Fifty-two (92.9%) had "possible, probable, or definite" sinovenous thrombosis, but only 76.9% of those had "probable or definite" sinovenous thrombosis. Uncertainty in diagnoses stemmed from limitations in imaging and disagreement over interpretation of imaging studies. ICD-9 code 325 in the primary position identified 7 children; all had possible (n = 1), probable (n = 1), or definite (n = 5) sinovenous thrombosis. ICD-9 code 437.6 identified a single admission on a single case of probable cerebral venous thrombosis; it was unclear whether this case was "nonpyogenic." ICD-9 code 671.5 did not identify any children. ICD-9 code 325 is useful for identifying children likely to have sinovenous thrombosis, but it is not useful for differentiating pyogenic and nonpyogenic cases, and uncertainty in clinical diagnosis makes it difficult to gauge the true accuracy. Furthermore, it is important to search for the code in any position as limiting searches to the primary position misses most cases.