Table 1.
Demographic information for study participants (Children and caregivers; n = 375 at study start, except where otherwise noted).
Table 2.
Frequencies of biomarker testing, including 13 tests on plasma, 4 on fecal and L/M absorption testing on urine as shown.
*Of 326 children with samples obtained within 1 month of study start.
Fig 1.
Heat map showing all significant partial Pearson correlations of barrier and systemic biomarkers with HAZ or WAZ at enrollment (ie. study start, ss) or with changes in (Δ) HAZ or WAZ.
Significant correlations (at p<0.05; * = p<0.01) between biomarkers and growth, controlling for child age and gender. HAZ = height for age Z score; WAZ = weight for age z score; ss = study start; Δ = change in HAZ or WAZ at 2-6m followup; numbers range from 230 to 292 except for zonulin at age >12m where n = 172. Full r, p and df values are provided in S1 Table.
Fig 2.
Fecal MPO, Fecal alpha-1-antitrypsin (A1AT), and plasma LPS, FABP and SAA each predicts subsequent growth impairment.
a: For MPO, p = 0.028; n = 266 when correcting for age and gender, and independent of breastfeeding status (that showed no correlation in these 6-26m old children) and of age. b: For A1AT, n = 237; p = 0.042; and A1AT also correlates with “catchup WAZ” as well, p = 0.035 after correcting for age and gender. c: For urine L/M, higher values correlated (controlling for age and gender) with impaired growth (delta HAZ) (r = -0.173; p = 0.009; n = 230). d: For plasma LPS (ie lower LUM), higher values correlated with impaired growth (delta HAZ) (r = 0.151; p = 0.017; n = 251). e: For plasma FABP, higher values correlated with impaired growth (delta HAZ) (r = -0.134; p = 0.042; n = 231). f: For plasma SAA, higher values correlated with impaired growth (delta HAZ) (r = -0.132; p = 0.046; n = 231).
Fig 3.
Plasma citrulline (in girls) and tryptophan (in boys) predicts subsequent better growth.
A. citrulline (for which gender differs significantly) in girls (p<0.001; n = 131; median citrulline = 23.97 umol/L) and B. tryptophan in boys (p = 0.010; n = 114; median tryptophan = 66umol/L) predicts subsequent better growth.
Fig 4.
Repeated measures MANOVAs show interactions in predicting growth between MPO and Neo: high MPO when combined with high neopterin associate with poorest growth.
Fig 5.
Cluster dendrogram with Pearson correlations among those biomarkers with ≥274 values showing three main groups: 1) translocation markers, LPS and IgA and IgG anti-LPS or FliC as well as zonulin and the 2 potential predictors of subsequent growth, tryptophan and citrulline; 2) predominantly systemic responses to disrupted barrier function and translocation; and 3) markers of specific intestinal barrier disruption or local intestinal inflammation.
As discussed, groups 1 and 3 may predispose to group 2 systemic responses in associating with each other as shown in the heat map as would fit our concept of the pathogenesis of enteropathy.
Fig 6.
Path model, using Principle Components Analyses (Equamax rotation solution maximizing independence of groups) showing associations among 1) Barrier (green), 2) Local Gut (orange) and 3) Systemic (pink) sets of biomarkers, as well as their predictive utility regarding linear growth.