July 16, 2018
The vast majority of research studies on probiotics, prebiotics, and other interventions to modify the human microbiome report virtually no specific data on possible harms, according to a meta-analysis published online today in the Annals of Internal Medicine.
Even among studies that reported harms, most were incomplete or otherwise inadequate in doing so, "raising doubts about the confidence we can have in conclusions about the safety of these interventions," write Aïda Bafeta, PhD, from the Université Paris Descartes-Sorbonne Paris Cité in France.
In fact, just 2% of the trials adequately reported all the parameters recommended in guidelines for reporting harms, including number of participant withdrawals resulting from harms, definition of adverse events (AEs) and serious AEs (SAEs), number of AEs per study group, and the grade, type, and seriousness of the events with denominators.
In addition, 6% of the trials adequately reported safety findings in their results, including number of participant withdrawals resulting from harms and number of AEs and SAEs. Two percent of trials adequately covered harms assessment in their methods sections.
"The inadequacy in reporting harms-related results may lead to an inaccurate safety profile and erroneous decision making, with major consequences for patients," Bafeta and colleagues write. "For example, a generic statement, such as 'well-tolerated,' may be misinterpreted as a lack of AEs reported for a trial, but the reader cannot determine whether no AEs occurred in any participants, the AEs were not measured or reported, or both. Readers should be able to estimate the number of AEs, even if no events occur."
For the current analysis, the researchers searched the Cochrane Central Register of Controlled Trials, PubMed, EMBASE, and Web of Science and identified 384 randomized controlled trials, including 136 that enrolled healthy volunteers and 248 with participants who had at least one medical condition being studied. Among the trials of participants with medical conditions, 195 were outpatient trials and 53 involved patients in the hospital or a critical care setting.
Most trials (70%) were at a single center, and 42% had at least one private funding source. Probiotics were the most commonly studied intervention (69% of trials), and 71% of the trials used a placebo as a comparator. About a third of the trials (35%) investigated treatment for gastrointestinal diseases and endocrine and metabolic conditions.
Bafeta and colleagues looked for any information related to harms reported in the abstract, methods, results, or discussion sections, including toxicity, AEs, or other safety issues. They also looked for specifics on AEs, including patient withdrawals resulting from harms, and assessed the quality of reporting on any harms that were discussed.
"We believe that researchers must clearly describe the incidence and severity of AEs related to probiotics, prebiotics, and synbiotics, particularly when they are used to treat severe disease or are used by high-risk patients (such as preterm infants or persons who are receiving ventilation or are critically ill)," the authors write.
Yet more than a quarter of the studies (28%) did not report any harms-related data at all, and more than a third (37%) did not report safety results. Eighty percent of the trials did not provide the number of SAEs that occurred, if any.
Bafeta told Medscape Medical News that some clinical trials and case reports have shown that probiotics, for instance, have been associated with a wide range of health complications, including "gastrointestinal events such as diarrhea, constipation, nausea, or abdominal pain; respiratory infection, inappropriate immune response in susceptible individuals, deleterious metabolic activities and gene transfer."
Even death has been reported as a SAE related to microbiome modification, she said. "The PROPATRIA trial reported by Besselink et al (2008) reported a higher mortality rate in the probiotic treatment group than in the placebo-control group in patients with acute pancreatitis," she noted. "The big issue is the huge uncertainty about the safety of these interventions in particular when given to vulnerable patients, such as those who are immune-compromised, postsurgery, critically ill, or elderly." Critically sick infants and those hospitalized for a long time are also especially vulnerable, she told Medscape Medical News.
Among the 242 (63%) of trials that mentioned harms in their results, 37% only reported them using generic statements, and 16% lacked adequate metrics.
"There is strong belief about security of these ingredients because of the long history of their use as ingredients in food," Bafeta told Medscape Medical News. "The use of probiotics, prebiotics, and synbiotics is new in therapeutic evaluation."
Although relatively new, use of these products already well outstrips the research, which already occurs at a rate of approximately 40,000 new studies a year when one includes laboratory, animal, and human research, according to Vince Young, MD, PhD, a professor of infectious disease at the University of Michigan Medical School in Ann Arbor, who studies the microbiome.
"I think one of the main reasons there's not much adverse effect reporting is that these products kind of fly under the radar" when it comes to regulation, Young told Medscape Medical News. Probiotics and prebiotics are considered supplements, not drugs, and are therefore not usually subject to the same level of scrutiny as pharmaceuticals. (Some exceptions exist, such as fecal transplants.)
In addition, as most trials are not intended for use in a new drug application, it's not surprising that harms are so rarely reported, he said. Further, many researchers conducting the trials likely believe reporting harms is unnecessary because the products often fall under the FDA designation of "generally regarded as safe."
In reality, however, few data exist on the safety, or even the effectiveness, of microbiome-modifying interventions, Young said. Doctors may suggest patients try probiotics for a wide range of concerns, particularly as they are available over the counter, but "we don't actually have data about whether manipulating the microbiome is a reasonable strategy or not," he continued.
Many trials are probably premature, he said, as studies showing associations between bacteria make-up and health conditions do not provide enough information to the mechanisms behind the association.
"We have this association, but if we fix the microbiome, do we really fix the condition?" he said.
Young also points out that the microbiome is often blamed for a wide range of health problems today, but changing it may not be the solution.
"If making changes to it got you in trouble in the first place, and we don't know how to change it, then if we're going to modify someone's microbiota for something in the short-term, could we cause some long-term problems?" This may be particularly troublesome because some conditions potentially influenced by microbiota are ones that take a long time to develop, such as cardiovascular disease, metabolic conditions, or depression and other psychiatric conditions.
Young emphasizes that he's not suggesting there are major harms, or that prebiotics and probiotics might not be useful, but simply that not enough is known yet.
"How much basic research should you do before you take it into people?" he said. "I don't know what the answer is, but it seems like something that should be considered."
The research did not use external funding. The authors have disclosed no relevant financial relationships. Young reported receiving compensation for consulting for Vedanta Biosciences, Inc.