Celecoxib does not alter intestinal microbiome in a longitudinal, diet-controlled study

Sir,

Rogers and Aronoff in their recent article, “The influence of non-steroidal anti-inflammatory drugs on the gut microbiome” (1), described a study reporting the effects of several different classes of medications on the fecal microbiome. Specifically, they found that NSAIDs reproducibly alter intestinal bacterial composition, depending on which NSAID was used. They describe an ambitious cross-sectional evaluation of 155 community dwellers of all ages from Southeastern Michigan who completed a questionnaire about their medication use in the 30 days prior to a fecal collection using a home stool specimen kit. Aliquots were extracted and analyzed by 16S rRNA pyrosequencing. The authors found enrichment of several different species depending upon the NSAID used, including Acidaminococcaceae and Enterobacteriaceae species in patients taking celecoxib.

Multiple factors influence the intestinal microbiome, including the subjects’ age, gender, and diet, making cross-sectional evaluation of the impact of any single medication difficult. We believe that information about the effects of common medications on the gut microbiome are useful to better clarify the microbiome impact on human disease. Performing such studies longitudinally with both diet and environmental control can reduce the experimental variation.

We performed a longitudinal study of the effects of the celecoxib 200 mg twice daily in a homogeneous group of 10 obese post-menopausal women during analysis of the effects of this drug on adipose tissue inflammation. Recruited subjects were not permitted to consume dietary supplements during the study. Their diets were analyzed using three-day food diaries under the supervision of a research nutritionist and each subject’s individual diet was maintained throughout the study period. While subjects were admitted to the metabolic unit of the Rockefeller University Hospital for a two-day run-in period, blood, urine, and stool specimens were obtained. These same samples were obtained again after 10 days receiving celecoxib. Eight subjects also produced a fecal specimen during a visit 7–14 days after leaving the hospital while continuing to consume their individual diets. Aliquots of fecal specimens were immediately frozen at -80 degrees C and analyzed by 16S rRNA microbiome sequencing and imputed functional activity.

We found no drug treatment effects on bacterial richness, microbiome composition, or beta diversity (Figure 1). No bacterial species had significant abundance alterations from celecoxib administration either during treatment or follow-up. Inter-individual variation was the driving feature of bacterial diversity, and subjects maintained their own core microbiota throughout the observation period (Figure 1). Samples clustered by subject with no effect of treatment. PICRUSt imputed microbiome metagenome patterns were similarly unchanged. The lack of effect of celecoxib on the gut microbiome stands in contrast to diet intervention studies, in which a small but quantifiable diet effect on the gut microbiome is observed.

Figure 1.

Inter-individual variation, rather than celecoxib administration, accounts for most microbiome diversity within a longitudinal human cohort. A) Average bacterial richness (as observed operational taxonomic units, OTUs) in subjects’ stools during run-in (before), after 10 days of celecoxib administration, and 7–14 days after completing drug administration (follow-up). B) Unweighted UniFrac PCoA comparing microbiome similarity between samples. Each sample is represented by a point, color-coded by subject. Note that samples predominantly cluster by subject, indicating no consistent effect of celecoxib treatment on bacterial composition.

The possibility that celecoxib reproducibly impacts the microbiome remains promising. Celecoxib is metabolized by many microbial species (2) and has broad antimicrobial activity against Gram-positive bacteria, as well as Gram-negatives when supplemented with membrane permeabilizing agents (3). Although inter-individual differences remained following treatment, the brief administration of the drug is a weakness of our study. However, in human subjects, celecoxib is rapidly absorbed, metabolized, and excreted primarily in the feces (4). Larger-scale and longer longitudinal studies in a more diverse cohort are warranted to fully assess the effects of celecoxib on the human intestinal microbiome. If celecoxib impacts the microbiome, it is important to ascertain its effects on patient health, including the risk of Clostridium difficile infection (5). These are important possibilities to explore, and we argue that more investigation is needed.

Sincerely,

NA Bokulich, T Battaglia, J O Aleman, J Walker, MJ Blaser, and PR Holt.