Refers to: https://www.sciencedaily.com/releases/2016/08/160827162200.htm

Though we’re still far from knowing enough to precisely engineer human weight loss through gut microbe manipulations, ironically, at least one case study (1) shows it’s possible to influence body weight simply by manipulating gut microbiota. Irony because the outcome in that case turned out to be inadvertent weight gain. Nevertheless this example offers

  • An important proof of principle that microbe transfers could mediate metabolic transformation in humans.
  • Insight into conditions necessary for effecting such transformation.
  • Scientific obstacles that stand in the way of using it to reverse obesity in human subjects.

In this report (1),

  • A 32-year old woman with recurrent Clostridium difficile infection (CDI) underwent Fecal microbiota transplant (FMT), her 16-year old daughter the FMT donor.
  • The FMT proceeded as planned after the daughter was screened and declared clear for HIV1 and 2, syphilis, hepatitis A, B and C, C. difficile, Giardia lamblia and other enteric pathogens,.
  • Post-FMT, the patient’s condition improved with no CDI recurrence . However, 16 months post-FMT, patient reported having gained 34 pounds. She now weighed 170 pounds with a BMI of 33, i.e., obese. This even though she hadn’t initially lost any weight over the months she underwent a series of antibiotic Rx for CDI.
  • Prior to FMT, the patient stably weighed 136 pounds with a BMI of 26 while her daughter weighed ~140 pounds with a BMI of 26.4.

How to explain why this happened?

  • The FMT donor, the patient’s daughter, provides the first clue. 140 pounds at the time of FMT, she too steadily gained weight to subsequently become 170 pounds. Obviously 16 years is the tail end of puberty, a life stage with profound hormonal and other physiological changes, which undoubtedly also impact gut microbiota composition.
  • In fact, the FMT donor’s subsequent weight gain suggests that as she completed puberty, her microbiota were changing in profile from one that supported a lighter weight to one more efficient in harvesting energy and thereby in promoting weight gain.
  • Seen in this light, this FMT recipient’s weight gain merely mirrors that of her donor’s, with the microbiota transfer the most likely change agent. Presumably post-FMT transplant, the recipient’s gut microbiota resembled those of the donor’s, i.e., more efficient in harvesting energy and thereby promoting weight gain.

Colonisation resistance Is An Important Obstacle To Microbe Transplant ‘Take’

While this FMT recipient’s weight gain can be plausibly explained as mimicking that of the donor’s, still the question needs further probing because doing so gets us to the crux of scientific obstacles in using microbe transfers to effect weight changes, be it gain or loss. Crux in this case is gut microbiota status quo in normal individuals and how it contrasts with the situation in this case report.

  • Healthy GI tracts have commensal microbes occupying the various GI tract niches as groups of specialized workers performing essential functions predicated on the needs of the niches they occupy. Thus, regardless a body is lean or obese, healthy guts have or should have niches that are microbially replete and successfully repel not only harmful invaders like C.diff but also other outsiders seeking to occupy the same niche, a process called Colonisation resistance.
  • This FMT recipient’s recurrent CDI suggests her gut microbiota was obviously already in considerable turmoil to start with. Prior to resorting to FMT, her already unstable GI tract microbiota was then subjected to antibiotic Rx consisting of metronidazole, vancomycin, amoxicillin, clarithromycin, rifamixin (1). Thus, by the time she underwent FMT, her GI tract was pretty depleted of stable microbial inhabitants, i.e., of robust colonization resistance capability. The microbes she received from her daughter’s poop would thus have been able to easily colonize the now available niches in her GI tract.
  • Genetics is another important factor, as in relatedness between donor and recipient, which presumably increased the likelihood of ‘take’ of donor’s microbes in the recipient’s GI tract (1). A 2016 study (2) was among the first to a) monitor long-term fate of FMT in human recipients, and b) observe different fates of same-donor poop transplants in different recipients, i.e., some microbial species successfully colonized some recipients but not others. Factors that determine such microbial ‘take” are still not fully clear but likely the two most important are the species and strain fitness within the donor pool on the one hand and colonization resistance in the recipient’s GI tract on the other hand. Clearly, genetic relatedness between donor and recipient is likely to play a role in how similar or different microbiota are to start with between different individuals.

Thus, this case report (1)

  • Provides preliminary proof of principle that microbiota transplant may lead to metabolic transformation in the recipient to mimic those of the donor.
  • Suggests unoccupied or available microbial niches may be a prerequisite for such possibility to convert to actuality.

Some Scientific Obstacles To Successful Weight Loss Through Microbe Transplants In The Obese

  • Unless obese recipients’ GI tracts are prepared prior to microbe transplants to create niches that can accept them, such transplants may not work in reducing obesity. In other words, depleting recipients of their indigenous gut microbes may be a necessary preliminary step. Problem is antibiotics, easiest tools available to do this, are blunt instruments and each such antibiotic’s effect on indigenous gut microbes will differ from person to person since gut microbial populations vary. Which antibiotics optimally prepare a recipient GI tract for optimal ‘take’ is currently unknown as also whether indigenous microbe-depleting effect of any one antibiotic is even generalizable across different individuals.
    • However, antibiotics may not be the only approach to deplete recipients of indigenous microbes prior to microbe transplant since the 2016 study (2) found bowel lavage alone without prior antibiotic Rx allowed stable donor microbe colonization in FMT recipients. By monitoring recipient gut microbes from 84 up to 400 days, this study showed durable co-existence of some donor and recipient microbial species.
    • Much work still needs to be done to understand how to effect efficient ‘take’ of microbe transplants.
  • Obviously obesity is outcome of diet, microbes and genetics. While interventions such as prior antibiotic Rx and bowel lavage may prepare available GI tract niches for microbial transplants to successfully occupy them initially, can sustainable microbial ‘take’ be assured without more profound, long-term habit changes? Doesn’t the principle of colonization resistance suggest that continuing post-transplant with the same diets that sustained their obesity only increase the likelihood their post-transplant GI niches would continue to preferentially support such obesity-associated microbes? Doesn’t that suggest microbe engineering alone may not suffice unless accompanied by diet change?

Misuse Of A Statistical Tool Is An Obvious Weakness Of Mouse Model Studies Of This Kind

Finally, a couple of points to add to Drew Smith‘s thorough analysis of the study quoted in the question.

One, though mouse is the most prevalent preclinical animal model, the travesty is rarely do its findings translate to humans.

Two, important to note the blatant data manipulation all too common in such mouse studies. While the article in question refers to an as-yet unpublished study presented during an August 2016 conference, this group has published on this NAPE mouse model-associated weight loss in 2014 (reference 6 in Drew Smith’s answer, 3 here). The important bit is in the figure legend. This particular experiment has 4 groups with 10 mice per group. So far so good. Not so good? That the authors chose to show not individual data points for each mouse in each group or mean +/- SD (standard deviation) but rather mean +/- SEM (standard error of the mean).

SEM is derived by dividing the standard deviation (SD) by the square root of the number tested. Let’s say SD in one group was 4.7. By dividing this SD by the square root of 10, the number of mice in the group, one can artificially reduce the variation within this group down to 1.6, i.e., its SEM. Such manipulations make the data look much cleaner and clearly separate the trends between the groups but the fact that these authors had to resort to this gimmick in such a small data set suggests much greater actual variation, i.e., considerable variation within groups and therefore considerable overlap between groups.

Variation is a given in biology especially when experiments involve such complex entities as live animals and human beings. This is compounded by less than optimal precision and accuracy of many biological assays, and widely variable skill and rigor of experimenters. As with any statistical tool, SEM has value when used appropriately as for example when trying to account for inevitable variations between experiments. Repeat the same experiment over time with experimental animals divided into the same 4 groups and there’s likely to be some variation even in the same group across experiments. SEM can help offset such variation and its use in such circumstances is not only appropriate but also tempered by the fact that combining data from different experiments adds more statistical power to the dataset not only by simply increasing the number of subjects per group but also by accounting for inter-experimental variations. That was clearly not the case here. The authors themselves describe this experiment in their paper’s Materials and Methods (3) as one experiment of 40 male mice divided into 4 groups of 10 mice each. In other words, this is misuse of the SEM statistical tool.

If within-group variations are larger than between-group variations, obviously we can’t conclude much especially in small studies where each group has only 10 subjects. Obviously such studies couldn’t get published. With the Publish or perish imperative only strengthening not weakening in recent decades, resorting to Data dredging is also at epidemic proportions. As gatekeepers, scientific peer reviewers and journal editors are responsible for stemming the tide of such abuse of statistics, a factor that also plays an important role in the current biomedicine data irreproducibility crisis. Obviously and dismayingly this example shows that even respectable scientific journals with quite high impact factors like Journal of Clinical Investigation (JCI) still aren’t performing due diligence on the data they choose to publish. No wonder the data irreproducibility crisis shows no sign of abating.


1. Alang, Neha, and Colleen R. Kelly. “Weight gain after fecal microbiota transplantation.” Open forum infectious diseases. Vol. 2. No. 1. Oxford University Press, 2015. Weight Gain After Fecal Microbiota Transplantation

2. Li, Simone S., et al. “Durable coexistence of donor and recipient strains after fecal microbiota transplantation.” Science 352.6285 (2016): 586-589.

3. Chen, Zhongyi, et al. “Incorporation of therapeutically modified bacteria into gut microbiota inhibits obesity.” The Journal of clinical investigation 124.8 (2014): 3391-3406. Incorporation of therapeutically modified bacteria into gut microbiota inhibits obesity