Chinese polyherbals show preliminary benefits in diabetic osteoporosis

A network meta-analysis of 34 randomized controlled trials involving 3,720 people with type 2 diabetes and osteoporosis found that adding one of six commercial Chinese polyherbal preparations to conventional treatment was associated with improvements in several bone-density measurements, biochemical markers and fasting blood glucose. Point estimates for adverse events were lower with five combinations than with conventional treatment alone, although no individual safety comparison reached statistical significance. The trials were conducted in China, and the authors cautioned that low or very low certainty in many comparisons makes the apparent advantages of particular products exploratory rather than proof of superiority.

Researchers searched eight English- and Chinese-language databases, along with conference proceedings and reference lists, for studies published through November 2025. Eligible trials enrolled patients with type 2 diabetes and osteoporosis confirmed by dual-energy X-ray absorptiometry and compared an oral polyherbal preparation plus conventional biomedical care against the same background care without the herbal product. Six preparations were evaluated: Xianling Gubao, Jintiange, Tangmaikang, Gushukang, Liuwei Dihuang Wan and Guli.

Background regimens included calcium and vitamin D supplementation, active vitamin D analogues, bisphosphonates such as alendronate or zoledronic acid, and glucose-lowering medicines, but the specific agents, doses, frequencies and durations varied substantially among studies. Two reviewers independently screened the studies, extracted data and assessed bias using Cochrane Risk of Bias 2.0, while the certainty of network estimates was graded with the Confidence in Network Meta-Analysis framework. Analyses used a frequentist model with maximum-likelihood estimation.

Continuous outcomes were expressed as mean differences and dichotomous outcomes as risk ratios or odds ratios, each with 95% confidence intervals. Relative rankings were summarized through the surface under the cumulative ranking curve, or SUCRA, which ranges from 0 to 1. The investigators interpreted SUCRA values alongside effect estimates, confidence intervals and certainty ratings rather than treating the rankings as definitive guides to treatment.

Most pairwise comparisons had low or moderate certainty, and no outcome as a whole was consistently rated high certainty, although individual findings such as the phosphorus estimate involving Xianling Gubao reached that level. Bone mineral density, the principal surrogate outcome available in the eligible trials, generally favored adding a polyherbal preparation to conventional medicine. In 22 trials with 2,521 participants, five of six regimens increased lumbar-spine density relative to conventional care alone.

Guli had the highest exploratory ranking, with a mean difference of 0.18 and a 95% confidence interval of 0.09 to 0.27, followed by Xianling Gubao at 0.11, with a 95% confidence interval of 0.08 to 0.15. Evidence for the lumbar-spine estimates ranged from very low to moderate certainty. In 21 trials involving 2,296 people, all five assessed combinations increased femoral-neck density, but only Gushukang plus conventional care did so significantly, yielding a mean difference of 0.14 and a 95% confidence interval of 0.01 to 0.27.

No significant difference separated the polyherbal regimens from one another at that site. Results for other skeletal locations followed a similar pattern, with point estimates favoring combination therapy but few significant differences among the preparations themselves. Seven trials with 832 participants assessed the femoral trochanter, where Jintiange ranked first with a mean difference of 0.10 and a 95% confidence interval of 0.05 to 0.16.

Five trials with 621 participants evaluated Ward’s triangle, with Xianling Gubao producing a mean difference of 0.09 and a 95% confidence interval of 0.05 to 0.13; Jintiange generated an estimate of 0.09, with a 95% confidence interval of 0.06 to 0.11. Although the point estimates were favorable, pairwise differences among the tested combinations were not significant. Evidence for these regional density outcomes was rated very low to low certainty.

The included trials did not report incident fractures, trabecular bone score or other bone-microstructure measures suitable for quantitative synthesis, so the analysis could not determine whether the observed density increases corresponded to fewer fractures. The biochemical results were mixed and outcome-dependent. Five of six combinations significantly increased osteocalcin, which the researchers interpreted as a pharmacodynamic indicator of bone formation or turnover rather than direct evidence of fracture prevention.

Jintiange had the highest rank, with a mean difference of 1.80 and a 95% confidence interval of 1.21 to 2.39, while Guli ranked second at 0.99, with a 95% confidence interval of 0.16 to 1.83. Jintiange was also estimated to outperform Tangmaikang, Liuwei Dihuang Wan and Xianling Gubao for this marker, although one of those confidence intervals extended to zero. For alkaline phosphatase, Tangmaikang had the largest estimated reduction, but the result was imprecise and not statistically significant.

Xianling Gubao ranked highest for serum calcium and phosphorus; its phosphorus estimate had a 95% confidence interval of 0.17 to 0.29 and high certainty, while its calcium result and other phosphorus comparisons were supported by lower-certainty evidence. Liuwei Dihuang Wan ranked first, producing an estimated mean difference of −1.03 with a 95% confidence interval of −1.70 to −0.35. Jintiange followed at −0.65, with a 95% confidence interval of −1.48 to 0.17, and Gushukang at −0.38, with a 95% confidence interval of −1.88 to 1.12.

No significant pairwise differences distinguished the combinations, and certainty for this outcome was low to very low. The analysis also ranked all six regimens above conventional care alone for “total effective rate,” a composite response measure defined differently in each study. Gushukang ranked highest, with a risk ratio of 1.68 and a 95% confidence interval of 1.22 to 2.32.

Because the definition combined changes in bone density, pain or visual analogue scale scores, traditional Chinese medicine syndrome scores or symptoms depending on the trial, the measure is not equivalent to a standardized clinical endpoint. Safety findings did not establish that any preparation was safer than another. Five combinations had point estimates suggesting fewer adverse events than conventional care alone, but none produced a statistically significant reduction.

Guli had the highest safety ranking, with a SUCRA value of 81.1%, followed by Tangmaikang at 71.0% and Xianling Gubao at 58.2%. The corresponding evidence was rated very low to low certainty, and no significant pairwise safety differences were detected. The authors noted that reactions reported in the herbal groups were mostly mild gastrointestinal discomfort that resolved with symptomatic treatment or dose reduction.

The analysis therefore did not demonstrate a general safety advantage for the preparations, even though their pooled point estimates for adverse events were generally lower than those for conventional care alone. The findings remain provisional because of substantial clinical and methodological limitations. Every included study was conducted in China and published in Chinese, limiting certainty that the results apply to other ethnic populations, healthcare systems or clinical settings.

Nine of the 34 trials were rated at high risk of bias, while the other 25 had some concerns; allocation concealment and blinding were poorly described, and adherence to treatment was uncertain across all studies. A sensitivity analysis excluding the high-risk trials was not performed because removing them would have broken important links in the evidence network and produced unstable estimates. Treatment duration ranged from one month to one year, preparation doses and courses varied, and baseline bone-density data were incompletely reported.

Most importantly, no trial supplied the fracture or bone-quality outcomes needed to determine whether the observed density gains translate into fewer broken bones. The analysis instead identifies gaps for future trials with standardized background care, longer follow-up and patient-important skeletal endpoints.

Subject of Research: Comparative efficacy and safety of commercial Chinese polyherbal preparations used with conventional treatment for diabetic osteoporosis.

Article Title: The efficacy and safety of combining different commercial Chinese polyherbal preparations with conventional Western drugs in the treatment of diabetic osteoporosis: a network meta-analysis

Article References: Ni, X., Guan, Y., Zhi, S., & Wang, X. (2026). The efficacy and safety of combining different commercial Chinese polyherbal preparations with conventional Western drugs in the treatment of diabetic osteoporosis: a network meta-analysis. Frontiers in Pharmacology, 17, Article 1880450. https://doi.org/10.3389/fphar.2026.1880450

Image Credits: AI Generated

DOI: 10.3389/fphar.2026.1880450

Keywords: diabetic osteoporosis, network meta-analysis, polyherbal preparations, randomized controlled trials, bone mineral density, bone turnover markers, fasting blood glucose, adverse events, Chinese medicine, efficacy, safety, combining