Table 2.
Between-group comparisons of faecal bacterial abundances at baseline and 4-week follow-up in the control and LFD group†
Control group (n 36) | LFD group (n 33) | ANCOVA (n 69) | ||||||||
---|---|---|---|---|---|---|---|---|---|---|
Baseline | Follow-up | Baseline | Follow-up | |||||||
Taxon | Median | 25th, 75th percentile | Median | 25th, 75th percentile | Median | 25th, 75th percentile | Median | 25th, 75th percentile | Estimate‡ | P value‡ |
c_Actinobacteria | 0·52 | 0·02, 1·07 | 0·41 | 0·04, 0·97 | 0·29 | 0·13, 1·05 | 0·16 | 0·04, 0·57 | – | – |
f_Rikenellaceae§ | 3·36 | 2·58, 5·87 | 3·70 | 2·93, 6·05 | 3·62 | 2·49, 6·04 | 4·15 | 2·63, 6·64 | 0·08 | 0·48 |
g_Alistipes§ | 3·25 | 2·41, 5·87 | 3·70 | 2·93, 5·70 | 3·62 | 2·40, 6·04 | 4·15 | 2·63, 6·64 | 0·08 | 0·51 |
g_Anaerostipes§ | 0·86 | 0·45, 2·00 | 1·14 | 0·65, 1·43 | 1·01 | 0·41, 1·90 | 0·36 | 0·18, 0·73 | –0·31 | < 0·001* |
g_Lachnospiraceae UCG-008 | 0·01 | 0·00, 0·03 | 0·03 | 0·01, 0·04 | 0·02 | 0·01, 0·04 | 0·03 | 0·01, 0·06 | – | – |
g_Lachnospiraceae UCG-010§ | 0·01 | 0·00, 0·02 | 0·01 | 0·00, 0·02 | 0·01 | 0·00, 0·02 | 0·02 | 0·00, 0·04 | 0·04 | 0·03* |
g_Tyzzerella3 | 0·01 | 0·00, 0·06 | 0·00 | 0·00, 0·07 | 0·01 | 0·00, 0·12 | 0·00 | 0·00, 0·07 | – | – |
g_CandidatusSoleaferrea | 0·00 | 0·00, 0·01 | 0·00 | 0·00, 0·01 | 0·00 | 0·00, 0·01 | 0·00 | 0·00, 0·02 | – | – |
g_Flavonifractor§ | 0·03 | 0·02, 0·08 | 0·03 | 0·01, 0·08 | 0·03 | 0·01, 0·06 | 0·01 | 0·01, 0·05 | –0·03 | 0·19 |
c_Erysipelotrichia§, || | 0·39 | 0·24, 0·58 | 0·40 | 0·24, 0·69 | 0·48 | 0·30, 0·82 | 0·80 | 0·35, 1·16 | 0·16 | 0·02* |
P < 0·05.
Bacterial abundances (%) are presented as medians with 25th and 75th percentile. LFD, low FODMAP diet; FODMAP, fermentable oligosaccharides, disaccharides, monosaccharides and polyols.
Between-group comparison: ANCOVA models were fitted with follow-up values as response variable and with baseline values and diet group as explanatory variables. Estimated difference between the diet groups at follow-up (control as reference) with corresponding P group value from ANCOVA. Only taxa which had shown a change or tendency for change within the LFD group (within-group comparison using Wilcoxon signed-rank test, online Supplementary Table 6) were subject to between-group comparison. ANCOVA could not be performed for four out of ten taxa due to strong deviations from parametric model assumptions, low abundance and/or low detectability.
ANCOVA performed on square root transformed baseline and follow-up values.
Completely dominated by family Erysipelotrichaceae.