ISSN: 2161-0665
Mitsuru Muto, Hiroshi Matsufuji, Fuminori Muranaga, Takeshi Tomomasa, Atsushi Nakajima and Tomoaki Taguchi
Background: A former nationwide survey in Japan on pediatric chronic intestinal pseudo-obstruction (CIPO) determined the most common etiology to be 'idiopathic'. However, currently no appropriate diagnostic criteria for pediatric chronic idiopathic intestinal pseudo-obstruction (CIIP) in children has been defined. The report utilized data mining to identify key clinical features that would qualify statistically CIIP diagnostic criteria.
Methods: Sixty-nine cases were reviewed and subject to standard mining association rule. Cases were assigned to three groups: fully dependent on parenteral nutrition (PN) (group A, n = 15), on partial PN or independent from PN, but requiring permanent enterostomy for decompression (group B, n = 37), and all the other (group C, n = 17). Distinctive features related to each group assignment were then extracted from 87 reported clinical features of the CIPO. The statistical operation was performed by an expert in medical informatics using apriori algorithm.
Results: Key features of patients in group A and B, thought to be severely intractable cases, were the presence of abdominal distension, intestinal dysmotility, ileus without mechanical obstruction, and the absence of histopathological abnormalities in the intramural ganglion cells.
Conclusions: The presence of clinical features of a functional ileus, in the absence of proven histopathological abnormalities, can be used for diagnostic criteria for CIIP covering all the intractable cases in Japanese children. It is highly significant that a normal full-thickness histopathological finding was statistically extracted as one key feature for CIIP in our analysis.