Should you take probiotics with antibiotics? Read on as I review the evidence, break down some important methodology, and discuss the best practices for supporting antibiotic recovery.
Last year, I published an article and recorded a podcast with Chris Kresser on a study published in Cell that suggested the need for caution about taking probiotics after antibiotics. In both humans and animal models, probiotics were shown to delay the return of the normal, native microbiota.
I’ve gotten a lot of questions this week about a recent blog post by popular gut health blogger and functional medicine practitioner Dr. Michael Ruscio. His article dismissed the latest study in favor of a 2014 systematic review and contended that you should take probiotics with antibiotics. He’s certainly not the only one that has dismissed this study.
To dispel any confusion, I’m sharing my take here and thought I’d walk you through my analysis of the article and its claims.
TLDR: I still believe that the evidence warrants caution about taking probiotics during or after antibiotics and stand by what I said in my original article. If you feel like you have to take a probiotic with antibiotics, Saccharomyces boulardii CNCM I-745 (Florastor) is probably the least harmful, but more research is needed. Better yet, supplement with butyrate to support gut hypoxia or consider an autologous fecal transplant!
Addressing some specific claims
A few things before I dive in:
- First, my goal here is not to try to take anyone down. I simply want to focus on the evidence. Dr Ruscio has a great deal of clinical experience helping people with gut issues, and just because I disagree with him on the evidence here does not mean that he doesn’t produce plenty of valuable content related to gut health. I have referred to his site many times and hope that we can have an open and respectful discussion about this research.
- Second, I am always willing to consider new evidence, even if it directly contradicts what I’ve said before. I’ve been quite transparent about major paradigm shifts in the past related to the optimal stool testing methodologies and whether SIBO breath tests are all they are cracked up to be. At the same time, I’m equally willing to defend my position if I believe the evidence still supports it.
With that said, let’s look at the major claims in the Ruscio article:
1) Probiotics make antibiotics more effective
“Rather than canceling each other out, research shows that taking probiotics and antibiotics together is more effective than taking antibiotics alone.”
This is absolutely true in some instances. Probiotics, especially strains of Saccharomyces var. boulardii, have been shown to improve eradication rates of H. pylori. They may also be effective for eradicating Blastocystis.
“There is also research to show that probiotics and antibiotics are more effective together for SIBO and other gut infections , , .”
I’ve written a lot recently about why we need to think about SIBO differently. The three studies cited here are based on cultured aspirates and breath testing, and therefore are not particularly convincing that we should be using probiotics and antibiotics to treat SIBO. Moreover:
- The first found that even probiotic monotherapy was effective.
- The second was not a randomized controlled trial.
- The third found only moderate quality evidence for probiotic prevention of C. diff infection in patients taking multiple antibiotics at the same time in a hospital setting. More on that in the next section.
2) Probiotics reduce antibiotic side effects
I discussed this in the takeaways section of my original article, where I outlined several studies that suggest probiotic use during antibiotics can confer certain protections to the host:
A 2012 meta-analysis published in JAMA found that probiotic use was associated with a significantly lower risk of developing antibiotic-associated diarrhea (AAD). However, the reduction in absolute risk was only 0.07. This means that 13 people would need to be treated and potentially delay normal restoration of their gut microbiome for 1 person to avoid AAD.
Subgroup analyses actually revealed a trend for an increased relative risk of AAD in patients receiving Lactobacillus (p=0.09) and Bifidobacterium (p=0.16) species. Saccharomyces and Bacillus species, on the other hand, showed a numerical but non-significant reduction in absolute risk.
Clostridioides difficile infection
A 2017 meta-analysis published in Gastroenterology found that probiotic use reduced the risk of C. difficile. difficile infection (CDI) in hospitalized adults taking antibiotics. Probiotics given within two days of starting antibiotics. were the most effective in preventing CDI. Again though, the reduction in absolute risk was small; this time, 43 people would need to be treated with probiotics and potentially delay normal restoration of their gut microbiome for 1 person to avoid C. diff. This was also performed in hospitalized patients, who are at much higher risk of CDI than individuals taking antibiotics in an outpatient setting. C. diff is certainly not something to mess around with and there may be some cases where the risk is high enough to warrant using probiotics. However, butyrate may also be protective against CDI, and may actually facilitate microbiome recovery instead of delaying it. But more on that later.
3) Probiotics correct dysbiosis caused by antibiotics
Now the claim that makes up the majority of the substance of the article:
“…when you are looking for health insights from research, it’s important to follow the overall trends rather than focus narrowly on one study. So, if we compare the evidence for taking probiotics after antibiotic treatment:
- A meta-analysis of 63 research trials shows that 83% of subjects experienced recovery in their microbiota after taking probiotics .
- A single study found probiotics to be less effective (for 8 patients) than no treatment at all (for 7 patients) .
It’s clear that this one small study doesn’t stand up again a much larger meta-analysis of 63 studies. This is the reason a meta-analysis is the gold standard for research.”
I agree that we ideally shouldn’t rely on a single study to shape recommendations, and I discussed the limitations of the study in detail at the end of my original article. I always make a point to review the literature for all available studies in a particular area before drawing any conclusions. So how did I miss a systematic review of 63 trials?
Systematic reviews and meta-analyses are absolutely the gold-standard for research, but importantly, are only as good as the studies they include. It’s essential to consider the strength of methodology of the individual studies and the selection criteria used by the authors.
So let’s take a look at these 63 research trials. I’d highly encourage you to open the systematic review and analyze it for yourself.
The authors included three different study models. A quick look at Figure 1 reveals that the only model relevant to our discussion is Model A, in which they took healthy volunteers with no dysbiosis at baseline, gave them a probiotic during antibiotic treatment, and then measured recovery of the gut microbiome afterwards.
Notably, only 10 studies with a total of 15 treatment arms fit this model (not 63!). If you look at Table 1, you can see a brief summary of all 15 treatment arms that were included in that section of the review. Of these:
- 12 were based on culture assays alone! While culture was once the only option for studying microbes, we now know that culture detects less than 5 percent of known human gut microbes and completely fails to detect or quantify some of the most important obligate anaerobes.
- 1 was based on fluorescent in situ hybridization (FISH) in patients receiving chemotherapy (not antibiotics) and saw no benefit from taking probiotics.
- 2 studies used culture-based analyses combined with PCR-based TRFLP: both of these involved a single follow-up sample at two weeks after treatment with a broad-spectrum antibiotic. TRFLP is a PCR-based method that certainly outperforms culture-based assays, but is still prone to some bias from amplification and is now considered an outdated These two studies are certainly worth a closer look though:
- Jernberg et al: (2005) a total of eight patients received clindamycin, four of which received a probiotic yogurt and four of which received plain yogurt with no probiotics. Unsurprisingly, the authors concluded: “…due to the unique microbial community fingerprint for each subject and due to the small number of subjects used, it was not possible to use statistics to calculate significant differences between the study groups.”
- Engelbrektson et al (2006): this study had a total of 32 patients that received augmentin, but TRFLP was not even sensitive enough to detect an effect of broad-spectrum antibiotic treatment, let alone the effects of probiotics on restoration! The only “significant” differences they reported between individuals who received the probiotic and those that did not were based on culture assays.
In summary, not a single study in this systematic review reliably suggests that probiotics support recovery after antibiotics. In the discussion of the paper, the authors themselves explicitly acknowledged the significant limitations of the study methodologies:
“Most of the studies (80%) using model A to document restoration of the normal microbiota only used microbiological culturing techniques, which can only detect those organisms that grow in culture. Use of the more advanced molecular metagenomic techniques have found that culturing alone misses up to 95% of these organisms.21 22”
My best guess is that Dr. Ruscio made the claim in his article based on reading the abstract of the systematic review alone, without actually reading the entire paper and looking at the individual studies. A more accurate statement about this paper after reading it in full would be: A systematic review of 63 trials, of which 15 treatment arms were relevant, of which 12 used outdated culture-based methodologies, showed that 83 percent of probiotics resulted in partial or complete restoration of the 5 percent of bacteria that can be measured in culture.
To my knowledge, the only methodologially sound study to date that fits Model A (assessing recovery from antibiotics in healthy adult volunteers) was the one I reviewed last year.
Yes, it’s a single study. Yes, it does have some limitations, which I laid out in detail. But it is the most reliable and complete study we have to date, with comparable results in humans and animals assessed using high-resolution molecular techniques, and a plausible mechanistic explanation (inhibition of the native bacteria by soluble factors secreted Lactobacillus spp.) tested in vitro.
Are there any studies in other populations?
Yep! As it turns out, a few other studies have assessed probiotic recovery after antibiotics in other populations, using metagenomics. (Thanks to Dr. Raja Dhir for sharing these with me.)
- Korpela et al. 2018 looked at 168 breastfed and 31 formula-fed infants that were born by Cesarean section and/or given antibiotics and given a four-strain probiotic and prebiotic (GOS) or placebo in a randomized placebo-controlled trial. Metagenomic analysis revealed that three months after birth, the combined probiotic/prebiotic was associated with a more “normal” microbiota composition, with increased relative abundance of Bifidobacterium and reduced Enterococcaceae and Enterobacteriaceae compared to placebo.
Given that the normal infant gut is heavily dominated by Bifidobacterium and Lactobacillus species, it is possible that probiotics containing these species could be protective against antibiotics and facilitate restoration at this stage in life. However, I do not think we can extrapolate this to adults, where these species make up only a very small percentage of the ecosystem. Moreover, there was no serial sampling done in this study, or standardization of probiotic administration relative to the antibiotics. Antibiotic exposure details were not provided in the full text or supplementary data.
- De Wolfe et al. 2018 looked at the effects of a four-strain probiotic or placebo for four weeks starting on the first day of antibiotic treatment for 31 adult patients with mild to moderate C. difficile infection. Stool was analyzed at baseline and at 4 and 8 weeks using 16S rRNA sequencing. There were no differences in alpha-diversity at 0, 4, or 8 weeks. Some differences were observed in individual taxa, with reduced Verrucomicorbiaceae (Akkermansia) in subjects administered probiotics at week 8 and reduced Ruminococcus in both groups at week 8 relative to baseline and 4 weeks.
It is possible that probiotics could be warranted with antibiotics in cases where you are starting from a place of infection or gut dysbiosis. However, with the lower resolution of 16S analysis, this study couldn’t make any major inferences. The authors concluded: “Results indicate that probiotic use may impact the microbiome function in the face of a CDI; yet, more sensitive methods with higher resolution are warranted to better elucidate the roles associated with these changes. Continuing studies are needed to better understand probiotic effects on microbiome structure and function and the resulting impact on CDI.”
I really do hope to see a lot more research in this area in a variety of different populations and will continue to update this as new evidence comes to the table. If conflicting results come out tomorrow suggesting that S. boulardii or other probiotics can help restore the gut microbiota of healthy individuals, using serial metagenomic samples, I’ll gladly change my tune. But for now, I stand by my original analysis, and in healthy adults, I still think caution is warranted.
Instead: support gut hypoxia
So, if probiotics might not be the best idea, what should we do instead?
As I discussed in my recent article on the oxygen-dysbiosis connection, perhaps the key reason why antibiotics cause dysbiosis in the first place is due to their ability to deplete important butyrate-producing microbes, stress the gut epithelium, and cause oxygenation of the gut mucosa.
This oxygen leakage into the gut results in the overgrowth of opportunistic, facultative anaerobic bacteria like Salmonella, Enterococcus, E. coli. These pro-inflammatory bacteria can use oxygen to grow and reproduce. A growing body of research in animal models has demonstrated that supporting gut hypoxia (low oxygen) with supplemental butyrate might help mitigate the overgrowth of these opportunistic pathogens.
The future: stool banking and autologous FMT
Even better: bank your stool before antibiotics and use it to re-inoculate your gut with a personal “autologous” fecal microbiota transplant (aFMT)! In the study I discussed last year, this approach resulted in near complete restoration of the gut microbiota in less than a single day.
Currently this is a bit complicated and potentially risky to do at home, but I have no doubt that this will one day be the standard of care and hope to contribute to research efforts that will help make this happen.
An analogy: thinking of the gut as an ecosystem
In case this is all a bit too confusing, let’s think about this on a macro-scale. I’ve written before about the importance of thinking about the gut from an ecological perspective. (Credit: Dr. David Relman for the original analogy and thanks to Stephanie Welch for helping me develop it at the PAH retreat!)
Think of your gut as a rainforest. Here’s your gut before antibiotics. It’s a dense, complex network of hundreds of species and their interactions.
And here is your gut after antibiotics: virtually all of the species have been knocked down, with perhaps a few seeds left in the soil, and the terrain is looking pretty depleted.
So what’s the best way to restore the ecosystem to its former state?
- Introduce a few invasive species (probiotics)? This may possibly prevent the overgrowth of a particularly nasty native species that was more resilient to the disturbance (C. diff) or prevent soil erosion (antibiotic-associated diarrhea). But introducing a large amount of one species is likely going to produce a monoculture that will fill the niche meant to be filled by the native species. It’d be like re-seeding the rainforest with a field of kale, spinach, and arugula – all “good” species, but not really the true biodiversity we’re looking for. The native species will likely still squeeze their way back in over time, but the invasive species will likely slow their progress and potentially skew the end result.
- Support the health of the soil (butyrate, ketones, nutrient-dense diet). A healthy environment (gut mucosa) will naturally select for the species that it is evolutionarily designed to support. Support the environment and provide the right substrates, and the native species should grow back over time.
- Re-seed a little bit of ALL of the native species (autologous FMT). Transplanting a little bit of all of the native species back to the area will help it return to its former state much faster. The result may not look exactly like the original ecosystem but will probably come pretty darn close!
Lastly, a few words on probiotic strain
The Ruscio article ends with:
“It’s not necessary to find the one right probiotic strain for your specific health condition. That’s because all probiotics have a similar synergistic effect of balancing the gut microbiota, modulating the immune system and reducing inflammation.”
This really surprised me, as it plainly contradicts all of the research to date.
It’s true that for a few purposes, differences between strains may not matter as much. But in a great deal of cases, strain specificity is extremely important.
I’ve highlighted this before, but here are just a few examples to illustrate my point:
- Lactobacillus rhamnosus GG has shown some promise in randomized, placebo-controlled trials of mild-to-moderate ulcerative colitis, but has been linked to nausea, vomiting, epigastric pain, and other adverse events in Crohn’s disease.
- Lactobacillus plantarum 299v has been shown to reduce IBS symptoms in a number of trials, while Lactobacillus plantarum MF1298 has been shown to worsen IBS symptoms.
In fact, a 2018 systematic review and meta-analysis concluded:
“Strong evidence was found supporting the hypothesis that the efficacy of probiotics is both strain-specific and disease-specific. […] The clinical relevance of these findings indicates that health-care providers need to take these two factors into consideration when recommending the appropriate probiotic for their patient.”
I’d really like to think this discrepancy isn’t because Dr. Ruscio sells probiotics in his supplement store, which currently do not provide their strain information and were linked in his article. I’d love to see him come around on this and hope he will be open to considering this research. For example, his Lacto-Bifido blend currently contains both L. rhamnosus and L. plantarum, but I wouldn’t know whether to recommend it for someone with Crohn’s or IBS because I don’t know what strains it contains.
This is one of the major reasons why I continue to crowdfund my work through Patreon. I truly have nothing against practitioners who choose to rely on affiliate marketing or selling supplements, but personally prefer abstaining from this realm so that I can be sure I’m providing you with my most objective, unbiased take on the evidence. I’ll certainly continue to update this article as we learn more!
That’s all for now! Hope you enjoyed this article. I’d love to hear your take on this research in the comments below!