The fermentation of starch in vitro produces a higher proportion of butyrat
e than the fermentation of most other substrates. The alpha-glucosidase inh
ibitor acarbose increases the amount of starch entering the colon, and has
been shown to increase faecal butyrate in humans. It is generally considere
d that colonic butyrate is quantitatively removed by the colonic mucosa and
liver and does not appear in peripheral blood. However, studies in animals
suggest that a small proportion of colonic butyrate reaches peripheral blo
od. Thus, we hypothesised that an increase in colonic butyrate production w
ould result in a rise in serum butyrate in human subjects. To test this, su
bjects with impaired glucose tolerance were randomly treated in a double-bl
ind fashion with placebo (n 11) or acarbose (n 11) (100 mg three times per
day). Serum short-chain fatty acid concentrations were measured twelve time
s over 12 h with subjects eating a standard diet before randomization and a
fter 4 months of therapy. At baseline, 12 h mean serum butyrate concentrati
ons were similar in the placebo and acarbose groups (2.8 (se 0.7) and 3.3 (
se 0.6) mu m, respectively). After 4 months on placebo, mean serum butyrate
(2.6 (se 0.5) mu m) was no different from baseline. However, after 4 month
s on acarbose, serum butyrate had increased to 4.2 (se 1.0) mu m, a value w
hich differed significantly from both the baseline value in the acarbose gr
oup and the treatment value in the placebo group. We conclude that acarbose
increased serum butyrate in subjects with impaired glucose tolerance. Thes
e results support the hypothesis that increased colonic butyrate production
in human subjects can be detected by an increase in serum butyrate.