This study found that fecal calprotectin measurement is very helpful for diagnosis of inflammatory bowel disease in children referred to pediatric gastroenterologists. Does it work as well for children evaluated in primary care settings?
Source: Holtman GA, Lisman-van Leeuwen Y, Day AS, et al. Use of laboratory markers in addition to symptoms for diagnosis of inflammatory bowel disease in children: A meta-analysis of individual patient data. JAMA Pediatr. 2017; 171:984-991. doi: 10.1001/jamapediatrics.2017.1736. See AAP Grand Rounds commentary by Dr. Philip Rosenthal (subscription required).
The short answer is that I don't know, and this article is a great example of a cornerstone principle in deciding how to use results of published studies. The article is an individual patient data (IPD) meta-analysis of 8 studies, comprising 1120 patients, evaluating various laboratory markers of inflammatory bowel disease (IBD). They found that fecal calprotectin was the most helpful diagnostic test in making clinical decisions where IBD was being considered as a diagnosis; in particular, a negative fecal calprotectin improved classification of patients into lower risk categories, and thus could prevent some unnecessary endoscopy procedures.
However, the degree of benefit of this testing may not be as great for children being considered for IBD diagnosis in a primary care setting. This is an example of spectrum bias. Patients referred to a gastroenterologist for IBD likely have more severe disease, or more typical features, of IBD, than do children presenting to their primary care provider with symptoms suggestive of IBD. To state more simply, these 2 groups of patients can be thought of as representing different diseases, and diagnostic testing may perform differently. So, until someone performs a study of laboratory markers of IBD in primary care, be careful about applying the current study's results in other settings. The authors of this current study were careful to stress this point.
You might also be wondering how IPD meta-analysis differs from standard meta-analysis. First of all, a meta-analysis is simply (though not so simple in practice!) a standardized review of published studies of a particular condition, utilizing statistical methods to "combine" the results of these multiple studies into a summary statistical evaluation. Whereas individual studies may not be powered enough to detect significant findings, combining the results of multiple studies might reveal significance. Much controversy exists in exactly how to choose studies to combine (e.g. high-enough quality, similarities among patients in different studies) and what statistical methods to utilize. I often joke that meta-analyses are the most dangerous form of research studies, because it's much tougher to spot or even agree on errors in analysis.
IPD meta-analysis is an attempt to correct some of the deficiencies of standard (sometimes termed aggregate) meta-analysis. In the latter, researchers are looking only at published results of studies, which may suffer from all the usual problems such as publication bias (tendency for negative studies not to be published), missing data, and inconsistency across different studies. IPD meta-analysis researchers, by having access to all patient data, can determine exactly where missing data occur and use the same statistical analyses for all the studies. Of course, cooperation from the authors of the original studies, to provide the raw patient data, is necessary for an IPD meta-analysis.
If you have some time over the holidays, look at an article from the UK a few years ago to see examples of how IPD meta-analysis can change conclusions of prior meta-analyses.
