Psoriasis is a difficult condition to live with and to study. It is chronic and relapsing, it flares and settles for reasons only partly understood, and it affects far more than the skin: itching, pain, visible lesions, sleep, and mood. More than 100 million people worldwide live with it. Standard care — topical medicines, phototherapy, systemic drugs, and biologic agents — helps many people a great deal, and leaves real gaps: long-term topical steroids can thin the skin, systemic drugs need safety monitoring, and biologics are not suitable for everyone.
That is the context for a new synthesis of the trial evidence on acupuncture-related therapies for psoriasis. Published on 2 November 2026 in Healthcare, it is an updated systematic review and network meta-analysis from researchers at Zhejiang Chinese Medical University and the Jiaxing Hospital of Traditional Chinese Medicine. It does not announce a winner. Its most useful contribution is showing why a winner cannot honestly be announced yet.
What the researchers did — and why the method matters
The team searched nine Chinese and international databases through 30 March 2026 and kept only randomized controlled trials in people with psoriasis vulgaris, the common plaque form. Thirty-three trials, involving 2,711 participants, made it in. The protocol was registered in advance (PROSPERO CRD420261435406), risk of bias was assessed with the revised Cochrane tool (RoB 2), and certainty of evidence for the main outcome was graded with CINeMA, the standard framework for network meta-analysis.
The method matters more than usual here because the previous network meta-analysis in this area, published in 2017 from searches completed in March 2016, grouped a wide range of interventions into broad buckets labelled simply "acupuncture" or "medication." That made comparison easy and interpretation hard: a reader could not tell whether any benefit came from manual needling, from fire needling, or from a combination that also included herbal medicine or phototherapy.
This review refuses that shortcut. It defines fifteen treatment nodes according to what was actually administered: manual acupuncture (MA), fire-needle acupuncture (FN), moxibustion (MOX), and microwave acupuncture (MWA) as single techniques, plus combinations such as FN plus Chinese herbal medicine (FN + CHM), MA plus Western medicine (MA + WM), and FN plus ultraviolet radiation (FN + UR). Western medicine (WM) itself was grouped in the main analysis to include topical, systemic, and combined treatments — a pragmatic choice the authors test, and partly undo, in sensitivity analyses.
The primary outcome was the Psoriasis Area and Severity Index (PASI) after treatment: a standard score combining how much skin is affected with redness, thickness, and scaling. Secondary outcomes included total response rate, quality of life (DLQI), itching scores, inflammatory markers, and adverse events.
What they found: different regimens ahead on different outcomes
Against Western medicine alone, four regimens showed statistically significant PASI reductions: MA + WM (standardized mean difference, SMD −2.44), MA + CHM (−1.79), FN alone (−1.77), and FN + CHM (−1.74). On the surface, that looks like a clear result. The rest of the paper explains why it is not.
First, no single regimen was consistently superior. For total response rate, MA + WM, FN, FN + CHM, and MWA + CHM were significantly favorable. For quality of life, it was FN, FN + CHM, MA + CHM, and moxibustion. For itching, MWA + CHM showed the largest estimate (SMD −3.35), followed by FN, FN + CHM, and MA + CHM. For the inflammatory marker TNF-alpha, FN, MA + CHM, and MA + WM were lower than Western medicine; for vascular endothelial growth factor (VEGF), FN + CHM and MWA + CHM were lower.
That pattern — different "winners" for skin scores, itching, quality of life, and blood markers — is itself informative. Psoriasis is not one symptom. But it also means there is no simple league table to carry away.
Second, the authors grade the PASI evidence as very low certainty. All four significant PASI comparisons had prediction intervals crossing the prespecified equivalence range. A prediction interval asks not just "what is the average effect across these trials?" but "what effect might a future, similar trial plausibly find?" When it crosses equivalence, a future trial could plausibly find no meaningful benefit at all. Heterogeneity was substantial (I² = 93.83%): the trials disagreed a great deal, with different disease durations, treatment durations, needling protocols, and comparator definitions feeding into one pooled number.
The ranking trap, illustrated twice
Network meta-analyses produce rankings, usually as SUCRA values (0–100%) summarizing how likely a treatment is to rank near the top. This paper contains two clean demonstrations of why they should not be quoted alone.
The regimen with the highest SUCRA value for PASI, FN + UR at 83.4%, did not differ significantly from Western medicine at all (SMD −2.90, 95% confidence interval −6.70 to 0.91). The top-ranked regimen for total response rate, MOX + UR at 87.7%, had a risk ratio of 7.03 with a confidence interval stretching from 0.37 to 133.80 — an estimate so imprecise it is compatible with anything from substantial harm to enormous benefit. A ranking built on such uncertainty is a ranking of noise, not of knowledge.
The sensitivity analyses make the same point from another direction. When the broad Western medicine node was split into topical, systemic, and combined treatments, the MA + WM comparison against systemic Western medicine alone was no longer significant (SMD −2.32, 95% confidence interval −4.98 to 0.35). Part of what looked like a treatment effect was an artefact of how the comparator was grouped. Treatment rankings, the authors write, "should be interpreted together with effect estimates and their uncertainty rather than as evidence of a clinically definitive best treatment."
Other limitations travel with the headline. Blinding is difficult in acupuncture trials, so subjective outcomes such as quality of life and itching are especially vulnerable to bias. Secondary-outcome certainty was not formally graded, because CINeMA was applied only to PASI. "Response" was defined differently across trials, so the pooled response rate is not a uniform endpoint. The pruritus and TNF-alpha networks showed global inconsistency, meaning direct and indirect comparisons did not fully agree. And every trial was conducted in China, with modest sample sizes and short treatment periods, so nothing here speaks to long-term benefit, recurrence, or safety over years.
Safety: sparse data, and one specific caution
The safety picture is thinner than the efficacy picture. The adverse-event network covered seventeen studies and 1,513 participants. Some regimens showed lower reported odds of adverse events than Western medicine, but events were sparse, several trial arms reported zero events, and definitions varied. The authors present no safety ranking: the network depended heavily on a single bridging comparison — one 2011 trial reporting 0 adverse events out of 30 participants with MA plus moxibustion versus 23 out of 30 with Western medicine — which pulled several estimates to extreme values. Under-ascertainment or selective reporting, they note, cannot be excluded.
There is also a psoriasis-specific caution that has nothing to do with statistics. Skin trauma can trigger new psoriatic lesions through the Koebner phenomenon, and a flare following acupuncture has been reported, although how often it happens remains uncertain. Needling is, by definition, controlled skin injury. That is not a reason for alarm, but it is a reason for the decision to be made with a clinician who knows the individual's disease, rather than on the strength of a pooled estimate.
What this does — and does not — change
The authors' conclusion is modest: acupuncture-related monotherapies and add-on regimens "may provide different benefits across clinically relevant outcomes," but no single treatment was consistently superior, and the findings "preclude firm treatment recommendations." They describe the results as more consistent with a possible adjunctive role — something evaluated alongside established psoriasis care for selected patients — than with replacement of that care.
That framing is the right way to read the paper. Three things are true at once:
First, the signal is real enough to take seriously. Thirty-three randomized trials is not a handful of case reports, the technique-specific node structure is a genuine improvement on the 2016 synthesis, and several regimens point in the same favorable direction.
Second, the signal is not strong enough to choose between regimens. Very low certainty for the primary outcome, substantial heterogeneity, prediction intervals crossing equivalence, sensitivity to comparator grouping, and no long-term data mean the current evidence cannot tell a patient or clinician which regimen, if any, is best.
Third, combination regimens cannot be read as acupuncture effects. MA + CHM, MA + WM, and FN + CHM contain herbal medicine or conventional drugs alongside needling, and their effects cannot be separated into components. A favorable result for FN + CHM is not evidence that fire needling alone works, and the paper never claims otherwise.
What would settle the question is also clear, and the authors list it: larger, multicenter trials with standardized interventions and outcome measures, longer follow-up, and more geographically diverse populations, so the next synthesis can move from "these regimens look promising on average" to "this regimen helps these patients, by this much, for this long." Until then, this review does something valuable: it updates a decade-old comparison, organizes it honestly by what was actually given, and refuses to turn a ranking into a recommendation. That restraint is exactly what makes the finding worth trusting as far as it goes — and no further.
Source: Researchers at Zhejiang Chinese Medical University and Jiaxing Hospital of Traditional Chinese Medicine. "Comparative Effectiveness and Safety of Acupuncture-Related Therapies for Psoriasis Vulgaris: An Updated Systematic Review and Network Meta-Analysis." Healthcare. 2026;14(21):3718. DOI: 10.3390/healthcare14213718. Published 2 November 2026; PROSPERO CRD420261435406.
If you are curious about the slow standing practice at the still end of this tradition, our guide to zhan zhuang standing meditation explains the basic posture and principles.
If you are interested in how the classical Chinese medical tradition thinks about health more broadly — its ideas of balance, prevention, and the body's patterns — our book Huangdi Neijing: A Modern Introduction offers a beginner-friendly guide to the foundational text of TCM.
This article is for educational purposes only and is not medical advice. If you have psoriasis or are considering any complementary therapy, please discuss it with your dermatologist or qualified healthcare professional, and do not stop prescribed treatment on the basis of a research summary.