Poor nutritional intake prior to allogeneic hematopoietic cell transplantation is associated with higher incidence of lower-gastrointestinal graft-versus-host disease

Abstract

Abstract Introduction. Nutritional intake during allogeneic hematopoietic cell transplantation (allo-HCT) may influence clinical outcomes and gut microbiome ecology, though longitudinal studies are limited. With a cohort of patients from our Center, enrolled on a Clostridium butyricum MIYAIRI 588 (CBM588) pilot trial (NCT03922035), we explored how calorie and protein intake during allo-HCT correlate with microbiome composition /and incidence of lower gastrointestinal (GI) graft-versus-host disease (GVHD). Methods. We evaluated trial participants (n=30) at our Center from 4/2018 to 1/2020 with dietary intake data collected from day -7 to day 30 relative to transplant. Patients enrolled on the trial underwent reduced-intensity allo-HCT and, after a 6 patient safety lead-in, were randomized to either the treatment (standard of care with daily CBM588 administration from admission to discharge) or control arm (standard of care alone). Daily oral calorie and protein intake were recorded by clinical dietitians and diet technicians using standardized inpatient calorie count procedures. Calorie and protein intake were evaluated at multiple timepoints: baseline (day-7 to 0), week 1 (day 0 to 7), week 2 (day 7 to 14), and week 3 (day 14 to discharge). Calorie and protein intake were evaluated relative to estimated energy and protein needs (calculated by dietitians using the Mifflin-St Jeor equation and adjustment factors based on age, gender, weight, height, BMI, and clinical status). Weekly stool samples (baseline, days 0, 7, 14, 21, and 28 post-HCT) were collected from the participants until discharge. Shotgun and 16S-amplicon sequencing were performed to estimate microbiome composition. Sample alpha-diversity was estimated by the Simpson's reciprocal (S) at genus level. Results. Oral nutritional intake was below 90% of estimated energy requirements for most participants most of the admission time for allo-HCT. At baseline, protein intake was below estimated needs for 90% of patients consuming less than 90% of their requirements regardless of treatment or control arm assignment. In contrast, the median baseline calorie intake met 95% of estimated needs, with 48% of patients consuming less than this threshold. Both calorie and protein intake declined significantly during HCT in both the control and treatment groups: calorie intake decreased from baseline to day 7 (from a median of 95% of estimated requirement to 40%; p=6.3e-07) and protein intake decreased (from a median of 55% of estimated requirement to 20%; p=1.997e-06). Lower calorie intake was observed in CBM588-treated patients compared to control at week 2 (p=5.8e-04) and week 3 (p=0.01) and lower protein intake at week 1 (p=0.038), week 2 (p=0.002), and week 3 (p=0.029). Baseline oral nutritional intake was associated with patient outcomes. A Kaplan Meier survival analysis revealed patients who consumed <95% estimated calorie requirements at baseline were more likely to develop lower GI-GVHD (p=0.023). Patients who consumed <50% estimated protein needs at baseline were more likely to develop lower GI-GVHD (p=0.048). We hypothesized that poor intake and consequent outcomes were due to comorbidity. However, a Cox proportional hazard ratio analysis showed that calorie (p=0.035, hazard ratio=0.034-0.88) and protein (p=0.019, hazard ratio=0.005-0.62) intake association with lower GI GVHD onset hold after controlling for patient comorbidity. We also evaluated the association between nutrition intake and patient microbiota. The median protein and calorie intake value in the microbiome sample window were correlated with alpha diversity (R=0.37; p=0.0034, and R=0.392; p=0.0019, respectively). When focusing on baseline calorie and protein intake, patients with greater intake (>95% of estimated calorie requirements and >50% of estimated protein) had increased alpha diversity at day 14 relative to transplant. Patients that consumed less than the median for protein at baseline were more likely to have Enterococcus domination at day 28 relative to transplant. Conclusions. Poor nutritional intake during allo-HCT, particularly pre-transplant and early peri-transplant, may contribute to the development of lower GI GVHD and microbial dysbiosis. These findings underscore the need to investigate nutritional intake and clinical outcomes in larger cohorts. It highlights the potential therapeutic opportunity of early nutritional intervention on microbiome and transplant outcomes.