A randomized double-blind trial on perioperative administration of probiotics in colorectal cancer patients

Abstract

To investigate whether probiotic bacteria, given perioperatively, might adhere to the colonic mucosa, reduce concentration of pathogens in stools, and modulate the local immune function. A randomized, double-blind clinical trial was carried out in 31 subjects undergoing elective colorectal resection for cancer. Patients were allocated to receive either a placebo (group A, n = 10), or a dose of 10(7) of a mixture of Bifidobacterium longum (BB536) and Lactobacillus johnsonii (La1) (group B, n = 11), or the same mixture at a concentration of 10(9) (group C, n = 10). Probiotics, or a placebo, were given orally 2 doses/d for 3 d before operation. The same treatment continued postoperatively from day two to day four. Stools were collected before treatment, during surgery (day 0) and 5 d after operation. During the operation, colonic mucosa samples were harvested to evaluate bacterial adherence and to assess the phenotype of dendritic cells (DCs) and lymphocyte subsets by surface antigen expression (flow cytometry). The presence of BB536 and La1 was evaluated by the random amplified polymorphism DNA method with specific polymerase chain reaction probes. The three groups were balanced for baseline and surgical parameters. BB536 was never found at any time-points studied. At day 0, La1 was present in 6/10 (60%) patients in either stools or by biopsy in group C, in 3/11 (27.2%) in group B, and none in the placebo group (P = 0.02, C vs A). There was a linear correlation between dose given and number of adherent La1 (P = 0.01). The rate of mucosal colonization by enterobacteriacae was 30% (3/10) in C, 81.8% (9/11) in B and 70% (7/10) in A (P = 0.03, C vs B). The Enterobacteriacae count in stools was 2.4 (log10 scale) in C, 4.6 in B, and 4.5 in A (P = 0.07, C vs A and B). The same trend was observed for colonizing enterococci. La1 was not found at day +5. We observed greater expression of CD3, CD4, CD8, and naive and memory lymphocyte subsets in group C than in group A with a dose response trend (C > B > A). Treatment did not affect DC phenotype or activation, but after ex vivo stimulation with lipopolysaccharides, groups C and B had a lower proliferation rate compared to group A (P = 0.04). Moreover, dendritic phenotypes CD83-123, CD83-HLADR, and CD83-11c (markers of activation) were significantly less expressed in patients colonized with La1 (P = 0.03 vs not colonized). La1, but not BB536, adheres to the colonic mucosa, and affects intestinal microbiota by reducing the concentration of pathogens and modulates local immunity.