Background
The most commonly used and accepted definition of probiotics defines them as “live microorganisms, which are administered in adequate quantity and that confer benefits to the host’s health” [
1].
In the last decades, several studies have demonstrated the benefit of probiotics for various diseases, including critically ill patients, in systematic reviews or meta-analyses [
2‐
6].
In parallel, in some countries such as Finland, there has been a significant increase in the consumption of probiotics - from 1 l per person per year of
Lactobacillus GG to 6 l per person per year [
7].
According to the
Southern California Evidence-Based Practice Center, infectious complications related to the use of probiotics are infrequently assessed in the outcomes and follow-ups of randomized clinical trials and when assessments are made, these are inconsistent. Therefore, current literature is not able to answer questions about the safety of probiotics in intervention studies [
8].
Our study is a systematic review of case reports and case series evaluating serious infections such as bacteremia, fungemia, sepsis, endocarditis, abscess and other complications in the context of probiotic ingestion. We believe it may bring new evidence to clinical practice and possibly a more conscientious prescribing practice regarding antibiotics and probiotics.
Therefore, this systematic review aims to evaluate the most frequently reported infectious complications associated with the use of probiotics; the genera of probiotic microorganisms most frequently associated with infectious complications; populations and specific conditions most frequently affected with infectious complications; and finally, the factors associated with all-cause mortality among the reported cases.
Discussion
This systematic review of original articles published between 1976 and 2018 brought together 93 cases of patients who developed infectious complications related to probiotic ingestion.
The geographical distribution of cases identified by this review covers the 5 continents, and included developed as well as developing countries. We believe that these cases represent only a small proportion of existing cases, since case reports are increasingly difficult to publish. Not only the geographical dispersion of cases but also the increasing number of cases over time are noteworthy, possibly due to the more intensive use of probiotics around the world.
The most frequent infectious complication was fungemia, with 37.6% of the cases.
Saccharomyces were the most frequently reported probiotic microorganisms, accounting for nearly half of the cases. When evaluating sepsis only, 39.3% were caused by these fungi. The incidence of fungemia by
S. cerevisiae is unknown, although population studies associate it with 0.1 to 3.6% of all episodes of fungemia [
15]. Currently,
S. boulardii, which is marketed in many countries as a medicine to treat gastroenteritis, is considered identical to a
S. cerevisiae strain; this fact corroborates the findings of some authors about the compatibility between
S. cerevisiae identified in the biological material of the patients and
S. boulardii isolated from the probiotics of the respective cases. A review of fungemia [
15] by
S. cerevisiae showed that of the 60 cases found in the literature, 26 (43%) had previously used probiotics. Treatment of
Saccharomyces infections is based on expert opinion, and most clinical experience exists with fluconazole and amphotericin B; besides systemic antifungals, it is strongly recommended that probiotics are discontinued and indwelling foreign bodies are removed [
16]. In our review, 12 samples of
Saccharomyces spp. had antimicrobial susceptibility testing and 10 of them were sensitive to azoles. It is interesting to note that in the systematic review on use of probiotics in critical illness, only the older studies (from the 80’s and 90’s) had
Saccharomyces as the prescribed probiotic [
2].
Bacteremia accounted 20.4% of the infectious complications and the genus
Lactobacillus was responsible for 42% of these.
Lactobacillus bacteremia has an annual incidence of approximately 0.1 to 0.3%, [
7] in Finland. The 26% mortality at 30 days described in a recent review [
17] was credited to patients’ underlying disease, rather than bacteremia alone. The risk factors described for the occurrence of lactobacillemia are the use of broad-spectrum antimicrobials, invasive procedures of the gastrointestinal or respiratory tract, immunosuppressive conditions and selective intestinal decontamination [
18,
19]. We observed that of the 8 patients with
Lactobacillus bacteremia, 4 were immunosuppressed - 2 with HIV infection; 3 underwent intestinal invasive procedures and 1 had diverticular disease of the colon. There are some clinical studies with the use of probiotics in HIV patients where no infectious adverse effects have been reported. The risk of infection by probiotics in HIV patients appears to be small, but attention should be paid to patients with low CD4 count and disease or manipulation of the intestinal tract [
20,
21].
Boyle et al. [
22], evaluating children, proposed some probiotic risk factors for sepsis, and impaired immune system and preterm birth were considered the most important risk factors. In our review, 26.7% of the patients were younger than 1 year of age and practically two-thirds of them premature. Ten preterm infants had extremely low birth weight and half had sepsis, four patients had bacteremia and only one had fungemia. In 10 patients the indication for the use of probiotic was to prevent necrotizing enterocolitis (NE) and at least 4 of them ended up developing this complication, despite probiotic prophylaxis. NE is a serious condition, characterized by breakage of the intestinal barrier, dysbiosis and persistent inflammation of the colon; surgery may be necessary in 20 to 40% of cases and mortality may reach 30% [
23]. The systematic review and meta-analysis of Sawh et al. [
24], which evaluated the efficacy and safety of probiotics in more than 5000 infants less than 37 weeks of gestation or weighing less than 2500 g, observed a reduction in the incidence of severe NE and overall mortality compared to the placebo group. However, in the extremely low birthweight group, no significant difference was demonstrated for these outcomes. In a recently published large (4556 subjects), double-blind, placebo-controlled trial of an oral synbiotic preparation (
Lactobacillus plantarum plus fructooligosaccharide) in rural Indian newborns, a significant reduction in the primary outcome (combination of sepsis and death) was found in the treatment arm (risk ratio 0.60, 95% confidence interval 0.48–0.74). Also significant reductions were also observed for culture-positive and culture-negative sepsis and lower respiratory tract infections. The authors suggest that a large proportion of neonatal sepsis in developing countries may be avoided through this strategy [
25], although we believe safe drinking water and sanitation are more important for global health.
In our study, all-cause mortality was associated with age over 60 years, C. difficile colitis, and antibiotic use at the time of the probiotic-related infectious complication. The genus Saccharomyces was also positively associated with mortality, since 3/4 of patients who died had probiotics belonging to this genus as the etiological agent of the infectious complications found. Although the analysis suggests that the genus Saccharomyces is associated with higher overall mortality, we know that other factors were not considered, such as the numbers and types of comorbidity.
This systematic review is not intended to discourage the use of probiotics, which have been shown to be effective in many situations in clinical practice. Nonetheless, most trials evaluating probiotics included patients that were not severely ill, and therefore serious adverse events were not expected, which is often not a real-life scenario. On the other hand, the largest to date systematic review on the use of probiotics in critically ill patients failed to demonstrate an effect on ICU or hospital mortality. The number of patients included per trial was small, and the variety of probiotic strains, wide range of daily doses, and length of administration among the different trials weakened any possible clinical conclusions and recommendations [
2] . Therefore, to assume that probiotic intake is completely risk-free is not true. The proportion of cases of infectious complications is small when the total number of people who use probiotics is considered. However, the cases described here are infections with high mortality rates such as endocarditis and sepsis. So, although on one hand there is the possibility of publication bias, with more serious cases having been published, on the other, due to the mentioned limitation for the publication of case reports, several other serious cases may not have reached public knowledge.
Conclusion
The most frequent probiotic-related infectious complications were fungemia and sepsis and the most frequent probiotic microorganisms were of the genus Saccharomyces, a fungus.
Mortality was associated with age > 60 years, C. difficile colitis, current antimicrobial use and Saccharomyces infection.
Probiotics were often used in the context of excessive antibiotic use, and a more judicious use of antibiotics is critical, as the use of probiotics cannot be considered risk free and should be carefully evaluated for high- risk groups of patients.