Limosilactobacillus reuteri 6475 and Prevention of Early Postmenopausal Bone Loss

Key Points Question Can daily supplementation with the probiotic Limosilactobacillus reuteri 6475 vs placebo reduce or prevent age-dependent bone loss in early postmenopausal women? Findings In this randomized, placebo-controlled clinical trial with 239 early postmenopausal women, L reuteri 6475 supplementation vs placebo did not affect change in tibia volumetric bone mineral density (BMD). In addition, no effect on BMD of the lumbar spine or hip was observed over 24 months. Meaning The findings of this trial suggest that L reuteri 6475 does not affect BMD of the tibia, spine, or hip in early postmenopausal women and should not be recommended to women at this age to prevent bone loss.


Introduction
Osteoporosis, a disease characterized by low bone mineral density (BMD) and impaired bone microstructure leading to increased risk of fracture, represents a major public health concern with pronounced implications for morbidity and mortality. 1,2The total annual cost for osteoporotic fractures has been estimated to $20 billion in the US and $30 billion in the European Union. 3teoporosis treatments used today are only indicated for patients with already established low BMD, emphasizing the need for safe and effective novel treatments to prevent osteoporosis development. 4e gut microbiome has several important physiologic functions, including regulation of the immune system, protection against pathogen overgrowth, intestinal endocrine signaling, biosynthesis of vitamins, and contribution to energy biogenesis. 5[7][8] Limosilactobacillus reuteri ATCC PTA 6475 (L reuteri) is one of the few indigenous Lactobacillus species present in infants as well as adults. 9,10[16][17] A placebo-controlled randomized clinical trial (RCT) found that L reuteri reduced bone loss over 12 months in women aged 76 years by approximately half, 18 but the magnitude of effect was limited.
However, if the treatment effects increase over time, long-term treatment may result in clinically relevant differences in BMD, which are then likely to affect the risk of fracture in this population. 19In this RCT, we aimed to evaluate whether daily supplementation with L reuteri vs placebo could reduce early postmenopausal bone loss and whether the effects remained or increased during the 2year study.

Study Design
The Early Postmenopausal Bone Loss with Lactobacillus reuteri II study is a double-blind, randomized, placebo-controlled, single-center RCT performed in the greater Gothenburg area in southwestern Sweden between December 4, 2019, and October 6, 2022.The exclusion criteria, inclusion criteria, study procedures, and predefined outcomes are available in Supplement 1.The study was approved by the Swedish Ethical Review Authority.The study is reported following the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline for RCTs (Figure 1).

Participants
Study participants were recruited by online advertisements and letters that were sent to 10 062 women aged 50 to 60 years.Women who contacted the clinic (n = 752) first underwent a telephone screening process, resulting in 292 women who were invited to a screening visit.Of those who were  1).Inclusion criteria were willingness to participate, availability throughout the study period, no menstruation within the past 1 to 4 years, and serum 25-OH-vitamin D levels greater than 10 ng/mL (to convert to nanomoles per liter, multiply by 2.496).
Only randomized participants were reimbursed for travel and parking costs and received 900 Sk (US $82.62) after study completion.Exclusion criteria included: a T score less than −2.5 SDs for BMD combined with a 10-year probability of major osteoporotic fracture according to the fracture risk assessment tool (FRAX) 20 of 20% or higher; severe osteoporosis, defined as a T score less than −3.0 SDs in either total hip, femoral neck, or lumbar spine; vertebral fracture diagnosed using lateral spine imaging with dual-energy x-ray absorptiometry (DXA); and previous (within the past 5 years) use of antiresorptive treatment, including systemic hormone therapy (estrogen), bisphosphonates, strontium ranelate, or denosumab.Further Study Protocol details are provided in Supplement 1.

Randomization and Blinding
After inclusion, women were randomized to 1 of 3 treatment groups, with 2 groups receving different capsules with identical appearance of the active L reuteri treatment, high and low dose, and 1 group receiving placebo.The randomization visit took place within 4 weeks of the screening visit.The sponsor (BioGaia AB) carried out the randomization performed in blocks of 8 study participants generated using the website Randomization.com. 21The investigators had no access to the

Bone Measurements
Bone mineral density of the total hip and lumbar spine and body composition of the total body were measured at baseline, 1 year, and 2 years using scan imaging (GE Lunar iDXA; GE Lunar).The coefficients of variation for BMD measurements, assessed in 8 men and 24 women, were 0.48% for the total hip, 0.77% for the femoral neck, and 0.73% for the lumbar spine.Lateral spine imaging was used to diagnose and classify vertebral compression fractures. 25lumetric BMD (vBMD) and bone microstructure parameters were measured at baseline, after 1 year, and after 2 years in the distal tibia using high-resolution peripheral quantitative computed tomography (XtremeCT; Scanco Medical AG) using a protocol described previously. 26,27After processing the images, 28 the following variables were obtained: total vBMD (milligrams per cubic centimeter), cortical thickness (millimeters), cortical vBMD, and trabecular bone volume fraction (percent).The coefficients of variation at our unit, assessed in 30 women aged 75 to 80 years, were 0.2% (total vBMD), 0.5% cortical thickness, 0.3% (cortical vBMD), and 0.5% (trabecular bone volume fraction).

Blood Biochemistry
Fasting morning blood samples were collected from all study participants.Aliquots of serum and plasma were stored at -80 °C until analysis.Serum type I procollagen intact N-terminal propeptide (PINP), C-terminal telopeptide cross-links of collagen type I (CTX), 25-OH-vitamin D and calcium, and plasma short-chain fatty acids butyrate, acetate, and propionate were analyzed (eMethods in Supplement 3).

Outcomes
The primary outcome was the relative change in tibia total vBMD after 2 years of treatment compared with placebo.Effects of treatment vs placebo were also investigated after 1 year and 2 years for the following secondary outcomes: BMD at the total hip and lumbar spine; tibia trabecular bone volum fraction; tibia cortical area; tibia cortical vBMD; tibia total vBMD (12 months); CTX and PINP; and plasma butyrate, acetate, and propionate.

Statistical Analysis
The study had the power of more than 99% to detect differences between groups based on Bonferroni-Holm correction, an anticipated dropout rate of 15% at 2 years, and a presumed mean (SD) decrease in tibia total vBMD of −3.50% (1.6%) in the placebo group vs 1.75% (1.6%) in the highdose group or −2.10% (1.6%) in the low-dose group.A statistical analysis plan was developed, dated, and signed by the principal investigator (M.L.) and study statistician (A.P.) before unblinding (Supplement 2).The primary and all secondary variables were analyzed both for the intention-totreat (ITT) population (all randomized participants) and per-protocol (PP) population (those who completed the study and were not in violation of the exclusion criteria).The PP population excluded women who during the study used osteoporosis medication, menopausal hormone therapy (protocol amendment: December 1, 2022, as described) and described in the statistical analysis plan), oral glucocorticoids for more than 2 weeks, or other probiotic supplements.All tests were 2-tailed.To account for type 1 error, a preplanned α level of .025was used for the 2 active comparisons against the placebo with respect to the primary end point.Bonferroni-Holm adjustment was planned to be applied to the secondary end points.
Additional exploratory subgroup analyses that were not predefined and described in the statistical analysis plan (Supplement 2) were performed using logistic regression, to investigate which baseline variables were associated with a positive treatment response (defined as a percent change in total vBMD at 2 years of more than −1.0).The relation between 2 continuous variables was described and tested by the Pearson correlation coefficient in the case of normally distributed data and otherwise by the Spearman correlation coefficient.
In the ITT population, tibia total vBMD, lumbar spine BMD, total hip BMD, tibia cortical area, and tibia cortical vBMD decreased significantly in all 3 investigated groups at both 1 and 2 years.There were no group-to-group differences in percent change in tibia vBMD high dose vs placebo (leastsquares mean, −0.08 [95% CI, −0.85 to 0.69]) and low dose vs placebo (least-square mean, −0.22 [95% CI, −0.99 to 0.55]).There were no significant treatment effects on any other predefined bone outcomes (Table 2).A significantly higher percent change in PINP was observed at 1 year for the high-dose group compared with placebo and at 2 years for both the high-and low-dose groups compared with placebo.Following adjustment for multiple comparisons, these differences were not significant.Multiple imputation was used to allow efficacy analysis and group-to-group comparison between the change in tibia total vBMD in all randomized participants (n = 239).In this analysis, no significant difference was found between either L reuteri dose and placebo (eTable 4 in Supplement 3).Similarly, no group-to-group differences were found for any of the studied bone characteristics (primary and secondary) at year 1 or at year 2 in the PP population following adjustment for multiple comparisons (eTable 5 in Supplement 3).Overall, mean (SD) adherence to the study product was high, ranging from 87.5% (24.6%) in the high-dose L reuteri group to 93.6% (12.9%) in the placebo group (eTable 6 in Supplement 3).
During the study duration, there were no significant differences between the groups in any adverse event, in any severe adverse event, or in an adverse event leading to discontinuation of the study product.A higher proportion of participants reported any treatment-related adverse event in the placebo group than in the low-dose L reuteri group.There were no group-to-group differences in adverse events leading to discontinuation of the study product (Table 3).A detailed description of the distribution of adverse events and serious adverse events per treatment group is presented in JAMA Network Open | Diabetes and Endocrinology eTable 7 in Supplement 3. The change in gastrointestinal symptoms was also studied using the GSRS scale.There were no significant differences between treatment groups in the change in any of the GSRS subdomains (reflux, abdominal pain, indigestion, diarrhea, and constipation) or in the total GSRS score (eTable 8 in Supplement 3)..03 Abbreviations: BMD, bone mineral density; CTX, C-terminal telopeptide cross-links of collagen type I; LS, least-squares; PINP, procollagen intact N-terminal propeptide.
a Mixed models for repeated measures were applied with percent difference as the outcome variable, visit, treatment group, interaction visit × treatment group as the main fixed effects; and baseline value as a covariate.An unstructured covariance pattern is used for correlated data repeated over time.Diagnostic plots of residuals were investigated and found satisfactory.No between-group differences were significant following Bonferroni-Holm adjustment.
In an exploratory but predefined analysis, the group-to-group changes in serum levels of the short-chain free fatty acids acetic acid, propionic acid, and butyric acid were investigated.Over 2 years of treatment, serum acetic acid levels decreased significantly more in the high-dose L reuteri group than in the placebo group.No other group-to-group differences in the change in short-chain fatty acids were observed (eTable 9 in Supplement 3).
A predefined subgroup analysis was conducted evaluating the relative change in tibia total vBMD in the ITT population for subgroups defined by years since menopause (<median years, Նmedian years), body mass index (BMI) (<25, Ն25 [calculated as weight in kilograms divided by height in meters squared]), and International Physical Activity Questionnaire (<median metabolic equivalent of task [MET]/week,Նmedian MET/week).There were no significant differences in change in tibia total vBMD during the 2-year treatment between any of the L reuteri treatment groups and placebo for any subgroup (Figure 2; eTable 10 in Supplement 3).However, a significant interaction between BMI group and treatment group (P = .04)was observed for change at year 2 but not at year 1 (P = .28).A significant correlation between BMI and relative percent change in tibia total vBMD was observed both in the high-dose L reuteri (r = 0.38; P < .001)and in the low-dose L reuteri (r = 0.29; P = .01)groups but not in the placebo group (r = 0.15; P = .22)at 2 years (eFigure in Supplement 3).
Exploratory but not predefined analyses of treatment effect on percent change in tibia vBMD according to subgroups of CTX, protein intake, calcium intake, and low total hip BMD T score were also performed and did not reveal any interactions between these variables and treatment effect (Figure 2; eTable 10 in Supplement 3).Another exploratory analysis to identify differences in baseline traits between responders and nonresponders (defined as percent change in total tibia vBMD at 2 years of more than −1.0) was performed on all women in the high-and low-dose groups (eTable 11 in Supplement 3).Higher weight, BMI, total body fat percent, and 10-year fracture probability (FRAX) for major osteoporotic fracture (with and without BMD) were associated with a more positive treatment response.

Discussion
In this RCT, supplementation with L reuteri for 24 months resulted in no significant differences in the primary outcome, the change in tibia total vBMD, or in any of the secondary outcomes compared with placebo.The lack of treatment effect on bone parameters was evident both in the ITT and PP populations, demonstrating that protocol violations did not mask any treatment effect of the study product.Similarly, multiple imputation was used to perform analyses of efficacy and group-to-group comparisons with all participants in the ITT population, and this analysis showed no significant differences in the primary outcome variable between either L reuteri dose vs the placebo group.
The 3 randomization groups were in general well matched across multiple parameters, encompassing anthropometric measures, medical history, gastrointestinal symptoms, 10-year fracture probabilities, physical activity, and bone characteristics, including BMD at the spine and hip and the primary outcome variable, vBMD at the tibia.However, an imbalance in calcium and protein intake was observed between the high-dose group and placebo group, with the latter having greater intakes of both of these nutrients.This may have affected the longitudinal changes in bone parameters since it is well established that increased calcium intake as well as protein intake both result in increased BMD. 29,30Arguing against this hypothesis, additional exploratory analysis of treatment effect according to calcium and protein intake was performed and did not reveal any

Figure 1 .
Figure 1.Participant Flowchart for Limosilactobacillus reuteri vs Placebo Groups 752 Women screened by telephone

Table 2 .
Analysis of the Relative Change in Bone Characteristics and Bone Turnover Markers of the Intention-to-Treat Population a

Table 3 .
Summary of AEs of the Safety Population a JAMA Network Open.2024;7(6):e2415455.doi:10.1001/jamanetworkopen.2024.15455(Reprinted) June 12, 2024 8/14 Downloaded from jamanetwork.comby guest on 07/01/2024 Figure 2. Subgroup Analyses of the Primary Efficacy Variable in the Intention-to-Treat PopulationPercent change in tibia ultradistal (standard site) total volumetric BMD (primary efficacy variable) is shown for subgroups of different baseline characteristics.Mixed models for repeated measures are applied with percent difference as the outcome variable; visit, treatment group, and interaction visit × treatment group as main fixed effects; and baseline value as covariate.An unstructured covariance pattern is used for correlated data repeated over time.BMD indicates bone mineral density; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); CTX, C-terminal telopeptide cross-links of collagen type I; IPAQ, International Physical Activity Questionnaire; and MET, metabolic equivalent of task.36.JanssonPA, Curiac D, Lazou Ahrén I, et al.Probiotic treatment using a mix of three Lactobacillus strains for lumbar spine bone loss in postmenopausal women: a randomised, double-blind, placebo-controlled, multicentre trial.Lancet Rheumatol.2019;1(3):e154-e162. doi:10.1016/S2665-9913(19)30068-237. Malmir H, Ejtahed HS, Soroush AR, et al.Probiotics as a new regulator for bone health: a systematic review and meta-analysis.Evid Based Complement Alternat Med.2021;2021:3582989.doi:10.1155/2021/358298938.Kim S, Jazwinski SM.The gut microbiota and healthy aging: a mini-review.Gerontology.2018;64(6):513-520.doi:10.1159/000490615Extended Methods -Blood Biochemistry Analysis and Statistical Methods eReference eTable 1. Patient Disposition and Data Sets Analyzed (ITT Population) eTable 2. Medical History (ITT Population) eTable 3. Prior Medications (ITT Population) eTable 4. Change in the Tibia Total Volumetric BMD According to Treatment Using Multiple Imputation in the ITT Population eTable 5. Main and Sensitivity Analyses of the Percent Change in Bone Characteristics and Bone Turnover Markers in the PP Population eTable 6. Compliance (Safety Population) eTable 7. Adverse Events by System Organ Class and per Treatment in the Safety Population eTable 8. Analyses of the GSRS in the Safety Population eTable 9. Exploratory Analyses of the Change in Serum Short-Chain Fatty Acids in the ITT Population eTable 10.Subgroup Analyses of the Primary Efficacy Variable (ITT Population) eTable 11.Logistic Regression for Responders Based on the Responders Defined as Percent Change in Total Tibia vBMD at Two Years>-1.0(All High-and Low-Dose Participants Included) eFigure.Pearson Correlation Between BMI and % Change in Tibia Total vBMD by Treatment Group Over Time (ITT Population)