Diagnostic Value, Predictive Significance and Kinetics of Fecal Calprotectin in Suspected Acute Gut Gvhd / Gvhd Flare Post Allogeneic Stem Cell Transplant

Introduction:

Fecal calprotectin (FC) has been extensively studied in the setting of inflammatory bowel disease. It is also considered as a “biomarker” for acute intestinal graft-versus-host-disease (GVHD). However, data in the setting of acute gut GVHD is relatively scarce. Most studies on FC have compared median values in cohorts of patients with and without gut GVHD, with gut GVHD versus CMV colitis, etc. There is lack of a uniformly accepted cut-off to discriminate GVHD from other causes of diarrhoea. The sensitivity, specificity, and positive and negative predictive values are less well studied. Also, the kinetics and impact on prognosis are not well explored. In addition to the above gaps in knowledge, there is even more scarce data from low- and middle- income countries (LMICs) where infections with multi-drug resistant organisms, tuberculosis, salmonella, etc may confound the interpretation of FC. Therefore, we conducted this retrospective study to identify the sensitivity, specificity, negative and positive predictive values, prognostic impact and kinetics of FC in a tertiary care cancer centre in India.

Patients and methods:

This is a single-centre retrospective analysis of allogeneic transplant recipients. Consecutive patients who had their first FC done for evaluation of diarrhoea between June 2020 to June 2023 were included. Patients required to have at least 1 colonoscopy in order to be included in the analysis (since biopsy confirmation was taken as gold-standard for diagnosis of gut GVHD), but not necessarily at each episode of diarrhoea. Calprotectin was measured by chemiluminescence immunoassay with the upper limit of normal being 50 mcg / g of stool. It was monitored weekly from onset to resolution of diarrhoea. FC monitoring was stopped early if an alternative cause of diarrhoea was established. All diarrhoeal episodes were retrospectively classified as GVHD (or flares) and non-GVHD causes based on clinical presentation, results of microbiological investigations, concurrent colonoscopic and histologic findings (whenever available), response to antibiotics, need for and response to steroids, and any other interventions leading to resolution of diarrhoea. For obtaining the optimal cut-off and the sensitivity / specificity, the first FC value within +/- 5 days from the onset of diarrhoea (referred to as baseline FC) was taken into consideration. Sensitivity, specificity and predictive values were calculated using standard formulae. Kinetics of calprotectin were evaluated in patients with 3 or more values to predict steroid refractory GVHD.

Results

A total of 154 FC assessments were done during evaluation of 69 episodes of diarrhoea in 19 patients. The median age was 30 years and 11 were males. The diagnosis includes AML/MDS in 6, ALL in 7, CML in 5 and PMF in 1. Fourteen patients underwent matched related donor transplants. Of the 69 episodes of diarrhoea, 33 were because of GVHD or its flare. Colonoscopy (within 7 days of baseline FC) was done in 24 episodes; 21 of these were histologically confirmed GVHD. Overall, of the 33 GVHD / flares, 21 episodes were histologically confirmed and 12 were clinically diagnosed. The median baseline FC value was significantly higher in GVHD / flares compared with diarrhoea of other causes (33 vs 8.5, p=0.03). A cut-off of 50 mcg/g of stool gave a sensitivity and specificity of 33% and 77% respectively for GVHD / flares. Increasing the cut-off value to 100 mcg / g of stool improved the specificity to 92% (95% CI – 78-98%). At this cut-off of 100 mcg/g, the positive predictive value was 70%. Three or more values of FC were available in 16 of 69 (23%) episodes. Among these 16 episodes, 2 consecutive FC values > 100 mcg/g or a rise in FC to > 100 mcg/g after an initial decline to below 100 was present in 8 episodes. Seven of these 8 episodes were because of steroid refractory GVHD versus 1 of 8 without such kinetics (p=0.003).

Conclusions:

FC may be a non-invasive investigation to predict the probability of diarrhoea being caused by GVHD in allogeneic transplant recipients. Baseline FC > 100 mcg /g of stool was highly specific for gut GVHD / flare of GVHD. Persistent values > 100 mcg/g over 2 weeks or a rise in FC to > 100 mcg/g after an initial decline predicted the occurrence of steroid refractory GVHD. A larger cohort of prospective patients is required to validate these findings.