Does Limosilactobacillus reuteri reduce severe oral mucositis during radiotherapy?

Severe mucositis, not any mouth sore, was the target

Oral mucositis is the mucosal injury that radiotherapy to the head and neck produces as it passes through the mouth and throat. The trial's primary endpoint was not any mucositis but severe oral mucositis (SOM), defined as WHO Oral Toxicity grade 3 or higher — the tier associated in the paper's introduction with intense pain, impaired oral intake, difficulty talking and swallowing, and unplanned radiotherapy interruptions.

That distinction matters for reading every number below. A result about SOM incidence is not interchangeable with a result about mild mucositis, and the trial reports both separately. Secondary endpoints covered SOM onset and duration, plus the incidence, onset and duration of any-grade mucositis. Safety was monitored throughout.

160 patients, twice-daily lozenges, one hospital

The study was a double-blind, placebo-controlled, randomized phase II trial run from March 2024 to January 2025 at West China Hospital of Sichuan University. All 160 enrolled patients were included in the intention-to-treat analysis, 80 per arm, and received either L. reuteri probiotic lozenges or placebo twice daily during radiotherapy.

Baseline characteristics were reported as balanced. Mean age was 56.8 years (SD 11.8) and 80.6% of participants were male. Nasopharyngeal carcinoma was the most common tumor site (30.0%), followed by oral cavity (27.5%) and oropharynx (16.2%). Concurrent chemotherapy was given to 46.2% of the L. reuteri arm and 42.5% of the placebo arm, and median oral cavity dose was 40.3 Gy versus 40.1 Gy. Mean adherence was 94.3% and 92.6% in the intention-to-treat population.

Withdrawals left 71 and 70 patients for the per-protocol analysis. The stated reasons included disease progression, requests to add other mucositis-prevention medications early in radiotherapy, and antibiotic use toward the end of radiotherapy.

A 23.8-point absolute drop in severe mucositis

SOM occurred in 30 of 80 patients (37.5%) in the L. reuteri group and 49 of 80 (61.3%) in the placebo group. The relative risk was 0.61 (95% CI 0.44–0.85), the absolute risk reduction was 23.8 percentage points (95% CI 8.7 to 38.8), and P was 0.004.

The authors then adjusted for variables that could otherwise explain the gap. After accounting for oral cavity radiation dose, tumor site and receipt of concurrent chemotherapy, the adjusted relative risk was 0.62 (95% CI 0.45–0.84; P = 0.002). Adding smoking and alcohol history left it at 0.62 (95% CI 0.46–0.85; P = 0.003). In the per-protocol set, further adding baseline Shannon diversity and the abundance of Actinomyces graevenitzii gave an adjusted relative risk of 0.59 (95% CI 0.42–0.84; P = 0.003).

Post-hoc subgroup analyses reported the benefit as generally consistent across key strata, including patients receiving concurrent chemoradiotherapy and those on definitive radiotherapy. These are subgroup and post-hoc results, so they describe consistency rather than independent confirmation.

Later onset and shorter episodes, with a caveat

Time to SOM onset was longer with L. reuteri: the median was not reached versus 16.0 days on placebo (hazard ratio 0.43; 95% CI 0.27–0.67; P < 0.001). Mean SOM duration was 6.4 versus 14.9 days (mean ratio 0.43; 95% CI 0.21–0.66; P < 0.001).

Any-grade mucositis moved in the same direction. Incidence was 56.3% versus 78.8% (relative risk 0.71; 95% CI 0.57–0.89; P = 0.004), median onset was 14.0 versus 8.0 days (hazard ratio 0.41; 95% CI 0.28–0.60; P < 0.001), and mean duration was 19.9 versus 35.8 days (mean ratio 0.56; 95% CI 0.39–0.74; P = 0.003).

The authors report that the proportional hazards assumption was violated for the onset endpoints, possibly because mucositis is acute. They note that the overall hazard ratios for onset may therefore not be valid as standalone effect estimates, and that piecewise Cox regression still showed a significant effect in the initial segment. The onset hazard ratios should be read with that qualification attached.

Severity trajectories also differed: mean WHO scores and SOM incidence were consistently lower in the L. reuteri group at all evaluated time points, while both groups worsened toward a peak in the final week of radiotherapy.

Less parenteral nutrition and analgesia, no severe safety signal

Exploratory endpoints pointed to less intensive supportive care in the L. reuteri arm: parenteral nutrition use was 23.9% versus 54.3% (relative risk 0.44; 95% CI 0.28–0.70; P < 0.001) and analgesic use was 7.0% versus 21.4% (relative risk 0.33; 95% CI 0.13–0.87; P = 0.03). The authors describe these as nominal quality-of-life improvements and suggest the low analgesic use may partly reflect stoic attitudes toward pain among Chinese patients.

On safety, three patients in the L. reuteri group had mild-to-moderate gastrointestinal symptoms (diarrhea, abdominal distension, dyspepsia) and one placebo patient reported dyspepsia; all resolved spontaneously without intervention. No severe adverse events linked to L. reuteri or placebo were observed.

Not every exploratory measure separated. Grade ≥2 hyposalivation at the end of radiotherapy was 81.69% (58 of 71) versus 85.71% (60 of 70), and BMI reduction was comparable at 1.22 versus 1.49 kg/m².

Microbiome shifts are a hypothesis, not a proven mechanism

The trial collected 423 saliva samples for 16S rDNA sequencing, of which 420 passed quality control. In the L. reuteri group, alpha diversity (Sobs and Shannon indices) stayed stable across time points, while beta diversity showed specific shifts in community structure during radiotherapy (P = 0.001) that were not seen with placebo.

Radiotherapy shifted the oral microbiome in both arms, with opportunistic pathobionts such as Fusobacterium nucleatum subsp. animalis and Prevotella nigrescens rising from baseline to radiotherapy completion in both groups. Patients who never developed mucositis had different microbial profiles from those who did, and among patients with mucositis the placebo arm showed higher abundance of pathobionts while the L. reuteri arm showed higher abundance of commensals.

The authors are explicit that species-level 16S annotations and PICRUSt2 functional predictions are putative, and that no mechanistic, functional or causal link between specific microbial changes and clinical outcomes has been demonstrated. The microbiome results support a proposed model; they do not establish that microbial modulation caused the mucositis reduction.

One center, one population, a phase II question

Several features bound how far the 23.8-point reduction travels. The trial was single-center and did not use stratified randomization, which the authors say may limit generalizability and leave prognostic imbalance across heterogeneous tumor sites and radiation doses. Patients receiving perioperative antibiotics were excluded, so L. reuteri's antibiotic tolerance — relevant to broader clinical use — was not assessed.

The cohort's etiological profile also differs from Western head and neck cancer populations: the authors describe a high prevalence of EBV and low rates of tobacco and alcohol use, which may not represent HPV-dominant or lifestyle-associated populations. They call for multicenter trials across more diverse clinical and demographic settings.

The authors also note that clinical efficacy was more modest than in murine models, likely reflecting translational differences in relative probiotic dosing and biologically effective radiation doses, and that baseline microbiome divergence between arms may still confound interpretation despite adjustment.

A preventive signal worth testing, not a settled standard

The decision-relevant comparison in this paper is not whether L. reuteri helped at all, but how large and how well-supported the severe-mucositis reduction is relative to the alternatives described. The authors frame the context as thin: palifermin is approved for mucositis prophylaxis only in hematological malignancy patients receiving stem cell transplantation, and benzydamine mouthwash carries Level I evidence but is recommended only for head and neck radiotherapy below 50 Gy. Against that background, a 37.5% versus 61.3% SOM incidence in a phase II trial is a signal for confirmatory multicenter work rather than a change in standard prophylaxis.

The most useful way to read the result is by endpoint tier. The incidence, onset and duration of severe mucositis all favored L. reuteri and survived adjustment for radiation dose, tumor site, chemotherapy, smoking and alcohol; the exploratory supportive-care and microbiome findings are weaker, nominal or explicitly associative. A reader weighing this evidence should treat the SOM incidence reduction as the trial's primary claim, the onset hazard ratios as qualified by the proportional hazards violation, and the microbial story as a hypothesis the authors themselves say remains untested causally.

Sources used in this answer