Diagnostic Accuracy of Fecal Calprotectin for Predicting Relapse in Inflammatory Bowel Disease: A Meta-Analysis Original paper

What was reviewed?

This systematic review and meta-analysis evaluated whether fecal calprotectin relapse prediction can accurately forecast impending relapse in adults with inflammatory bowel disease (IBD) who are clinically in remission. Calprotectin is a neutrophil- and monocyte-derived protein complex (S100A8/S100A9) that reflects intestinal neutrophilic inflammation and is stable in stool, making it practical for home sampling and longitudinal monitoring. Across 24 prospective cohorts, the authors compared baseline fecal calprotectin (FC) values during remission against subsequent relapse outcomes and used a multiple-threshold statistical model to identify the FC cut-off that best balanced sensitivity and specificity for predicting relapse. Importantly, the analysis explicitly addressed the long-standing clinical problem that FC thresholds vary widely across studies (reported cut-offs ranged from 50–500 µg/g), making implementation inconsistent in practice.

Who was reviewed?

The meta-analysis included 24 prospective studies comprising 2260 adult patients with quiescent IBD at baseline, of whom 715 relapsed during follow-up. Studies enrolled patients with Crohn’s disease (CD) only (7/24), ulcerative colitis (UC) only (7/24), or mixed UC/CD populations (10/24). Follow-up periods were heterogeneous, commonly 12 months but extending from 2 months up to 38 months in some cohorts. Relapse definitions varied and were grouped into “clinical relapse” (symptoms and/or treatment escalation) versus “endoscopic relapse” (endoscopic activity, sometimes with histology). Only 5/24 studies used endoscopy as the primary reference standard, while most relied on clinical indices (e.g., CDAI, HBI, partial Mayo) and therapy change, reflecting real-world practice but also introducing variability in outcome ascertainment.

Most important findings

Using a multiple-threshold model that incorporates all reported FC cut-offs, the authors identified an optimal threshold of 152 µg/g for fecal calprotectin relapse prediction. At this cut-off, pooled sensitivity was 0.72 (95% CI 0.528–0.856) and pooled specificity was 0.74 (0.618–0.834), with an SROC AUC of 0.794, indicating good (but not perfect) discrimination for future relapse. The diagnostic odds ratio was reported as 14, supporting meaningful clinical utility. When the authors repeated the analysis using a traditional bivariate model (choosing one threshold per study, closest to 152 µg/g), performance estimates improved (sensitivity ~0.80, specificity ~0.78, AUC ~0.85), suggesting that single-threshold approaches may overestimate accuracy because they implicitly favor “best” cut-offs within each cohort. Subgroup analyses found broadly similar accuracy whether patients had UC vs CD, shorter vs longer follow-up, and whether relapse was defined clinically or endoscopically—supporting use of a unified threshold in practice. Clinically, the Fagan nomogram analysis suggested FC meaningfully shifts post-test probabilities: in low pretest relapse settings, a negative FC reduces relapse likelihood enough to support reassurance, whereas in high pretest settings, an elevated FC markedly increases relapse probability and can justify intensified monitoring. Notably, because calprotectin is also an antimicrobial metal-sequestering innate immune protein, elevated FC may reflect active neutrophil-driven mucosal inflammation that also reshapes the gut microbial ecosystem; however, this paper does not report microbial taxa-level signatures, instead offering a host inflammatory biomarker that can be paired with microbiome profiling to contextualize dysbiosis during “silent” inflammation.

Key implications

For clinicians, fecal calprotectin relapse prediction at a pragmatic threshold near 152 µg/g offers a noninvasive, relatively inexpensive strategy to flag patients in apparent remission who are biologically “smoldering” and at higher relapse risk within the subsequent months (up to ~24 months in the authors’ framing). The major value is triage: higher FC can prompt earlier reassessment (repeat FC testing, medication optimization, or endoscopic confirmation), while lower FC supports de-escalated surveillance and reduced patient burden. However, FC is not specific to IBD—false positives may occur with infections (e.g., C. difficile), malignancy, NSAID/PPI use, and other GI inflammatory conditions—so clinicians should actively exclude alternative causes before escalating IBD therapy. From a microbiome-signatures perspective, FC functions as a host-derived inflammation anchor that can be integrated with stool metagenomics: microbial shifts associated with relapse risk may be more interpretable when stratified by FC-defined inflammatory activity, enabling databases to distinguish dysbiosis linked to active neutrophilic inflammation from compositional variation unrelated to relapse biology.

Citation

Shi J-T, Chen N, Xu J, Goyal H, Wu Z-Q, Zhang J-X, Xu H-G. Diagnostic Accuracy of Fecal Calprotectin for Predicting Relapse in Inflammatory Bowel Disease: A Meta-Analysis.Journal of Clinical Medicine. 2023;12(3):1206. doi:10.3390/jcm12031206