Indirect renal-response signals emerge for Chinese patent medicines in lupus nephritis
A 48-trial multiple-treatment meta-analysis has put Chinese patent medicines back into the discussion for lupus nephritis, but only on carefully limited terms. Across 3,726 adults, combining a Chinese patent medicine with standard drug therapy generally outperformed standard therapy alone for renal response and disease activity, yet every comparison between individual patent medicines rested on an entirely star-shaped network. In plain clinical terms, no trial directly compared one patent medicine against another, so the apparent hierarchy is built from indirect inference rather than head-to-head evidence.
The distinction matters at the bedside. When every node in a network is anchored to the same control arm, rankings can look persuasive while still being vulnerable to small-study effects, selective reporting, and imprecision from single-center trials.
What the network found for complete and partial renal remission
The most favorable signal for complete renal remission came from Yishenyangyinheji plus standard drug therapy. In the pooled analysis, this combination reached a relative risk of 2.37, with a 95% confidence interval of 1.60 to 3.53, a SUCRA of 89.80%, an absolute risk difference of 0.356, and a number needed to treat of 2.80. For clinicians, those numbers suggest a strong apparent effect, but the certainty is not as robust as the ranking looks on first pass.
The partial renal remission signal was led by Huobahuagen Tablets plus standard drug therapy, with a relative risk of 1.60, a 95% confidence interval of 1.02 to 2.51, a SUCRA of 96.00%, an absolute risk difference of 0.210, and a number needed to treat of 4.80. A one-sentence caution sits right beside that result: the partial remission advantage did not hold up cleanly after trim-and-fill adjustment for small-study effects, with the adjusted relative risk falling to 0.956 and the confidence interval spanning no effect. The ranking also shifted when the node model changed, which is exactly the kind of instability that can separate a promising signal from a practice-changing one.
Disease-activity improvement favored Total Glucosides of White Paeony
For systemic activity, Total Glucosides of White Paeony plus standard drug therapy ranked first for lowering the SLEDAI-2K score on indirect comparison. The standardized mean difference was −3.89, with a 95% confidence interval from −4.93 to −2.84, and a SUCRA of 99.20%. That is a large numerical shift in disease activity, though it still comes from the same indirect network structure that limits the renal-response rankings.
Biologically, the attraction of adjuvant herbal-derived therapy in lupus nephritis is easy to understand. LN is driven by immune-complex deposition and downstream inflammatory injury within the kidney, and therapies that suppress systemic activity can plausibly reduce proteinuria, improve urinary sediment, and stabilize creatinine trajectories. The problem is not plausibility. The problem is whether the published trial record is strong enough to distinguish genuine immunomodulation from trial-level noise, reporting bias, or differences in background therapy.
Serum creatinine and proteinuria showed only modest movement
A smaller signal appeared for serum creatinine and 24-hour urinary protein. Langchuangpingkeli and Tripterygium glycoside tablets each showed a marginal trend toward better creatinine reduction and lower 24-hour total protein, but neither reached statistical significance. For a kidney disease where serum creatinine can lag behind active glomerular injury, weak movement on this endpoint is not trivial. It suggests that some of the observed benefit may be concentrated in remission definitions and proteinuria-based measures rather than in a clearly demonstrated preservation of filtration function over the studied time frame.
The evidence base was narrow, regional, and vulnerable to bias
The evidence network was built entirely from Chinese trials, most of them small and single-center. The review also makes clear that safety reporting was thin. Adverse events were mostly gastrointestinal when they were reported at all, but there was no standardized adverse-event framework and no pooled safety analysis. No trial systematically assessed gonadal or reproductive outcomes, a particularly important gap when Tripterygium-derived agents are part of the discussion.
This is where clinical caution becomes non-negotiable. Tripterygium-related products carry reproductive toxicity concerns, yet the available randomized record does not provide the kind of structured long-term safety data needed for routine recommendation. Even when an intervention seems to reduce proteinuria or improve remission rates, nephrologists still need to know whether the same drug harms fertility, liver function, infection risk, or treatment adherence over time. None of that was adequately answered here.
Why the rankings should be treated as hypothesis-generating
The meta-analysis itself is explicit about the limitations of the network. Because the comparison set was star-shaped, all between-patent-medicine rankings were purely indirect. GRADE certainty was low for the complete remission signal and very low for the partial remission signal. Those are not the grades that support broad implementation in routine care.
A different kind of strength would have been more convincing: multicenter randomization, blinded outcome adjudication, standardized renal-response definitions, and consistent reporting of adverse events and reproductive outcomes. Instead, the available trials mostly ask the same question in the same direction against the same control arm, which makes the overall picture look orderly even when the underlying evidence remains fragile.
What adult lupus nephritis care can and cannot take from this analysis
For now, the practical reading is restrained. Chinese patent medicines added to standard drug therapy may improve complete renal remission, partial renal remission, and lupus activity scores in adults with lupus nephritis, but the apparent ranking of specific products should not be mistaken for a validated treatment hierarchy. Yishenyangyinheji and Huobahuagen Tablets came out on top for renal-response endpoints, and Total Glucosides of White Paeony led the disease-activity score, yet those positions rest on indirect estimates from a limited evidence base.
The most defensible conclusion is also the most conservative one: these agents remain investigational adjuncts rather than established components of lupus nephritis care. The next meaningful step is not another ranking exercise built on small single-center trials. It is a set of adequately powered, multicenter studies with long-term renal endpoints, standardized safety monitoring, and explicit reproductive follow-up before any adjuvant Chinese patent medicine can be recommended with confidence.