Psoriasis looks like a skin disease and is not one. The visible part — sharply demarcated red plaques with silvery scale, characteristically on elbows, knees, scalp and lower back — comes from keratinocytes proliferating far faster than normal because immune signalling tells them to. But the same immune process affects joints, and travels with cardiometabolic and psychological disease, which is why calling psoriasis cosmetic is both wrong and consequential [1] [2] [3].
It is also the clearest example in medicine of a disease whose mechanism was worked out and then, deliberately, drugged. Two decades ago treatment meant tar, topical steroids, ultraviolet light and drugs borrowed from oncology and transplant medicine, and "response" meant a 75 percent improvement in a minority of patients. Now, in head-to-head trials, three-quarters of patients reach 90 percent clearance [4]. The centerpiece is about how large that shift is and how it was measured.
Start here: the axis
The pathogenesis is now understood as a self-sustaining loop [5] [6].
Dendritic cells produce interleukin-23. IL-23 sustains and expands Th17 cells — a T-helper lineage whose differentiation programme is directed by the nuclear receptor RORγt [7] [8] [9]. Th17 cells produce interleukin-17, which acts on keratinocytes, driving them to proliferate, to produce antimicrobial peptides and chemokines, and to recruit more inflammatory cells — which sustains the dendritic-cell activation that started it.
Each step of that chain was established experimentally. IL-23 was shown to drive a pathogenic T-cell population causing autoimmune inflammation [10], and to be the critical cytokine rather than IL-12 [11]; IL-17 and Th17 biology was characterised [12]; and in the standard mouse model, imiquimod-induced psoriasis-like skin inflammation was shown to be mediated specifically via the IL-23/IL-17 axis [13].
Every drug in the modern armamentarium blocks one step of that loop, which is what makes psoriasis a demonstration case rather than merely a treatment success.
It is not confined to skin. Up to a third of people with psoriasis develop psoriatic arthritis, an inflammatory arthritis with its own course and outcome [14], and psoriasis is associated with cardiovascular and metabolic disease and with depression [3] [15].
Pillar 1: measurement and diagnosis
The diagnosis is visual
Plaque psoriasis is diagnosed by looking: well-demarcated, erythematous, scaly plaques in characteristic distribution. Biopsy is occasionally needed to settle a differential, not routinely. Other forms — guttate psoriasis following streptococcal infection, pustular and erythrodermic variants, and nail and scalp disease — are recognised by pattern [16].
Severity, and the three numbers that matter
PASI — the Psoriasis Area and Severity Index — scores erythema, thickness and scaling across four body regions, weighted by the area involved. It is the endpoint of essentially every trial in this review [6], and it is used as a relative measure: PASI 75, PASI 90 and PASI 100 mean 75, 90 and 100 percent improvement from a patient's own baseline.
That relative framing is worth pausing on, because the bar moved as the drugs improved. The trials of the older agents report PASI 75 as their primary endpoint [17] [18]; the trials of the newer ones report PASI 90 [4] [19], because reporting PASI 75 for a modern biologic would compress most of the distribution into a single number. Comparing "response rates" across drug generations without noticing which threshold was used is therefore a reliable way to be misled.
Body surface area gives a simpler severity measure, and the Dermatology Life Quality Index captures what PASI cannot — that a small amount of psoriasis in a visible or sensitive site can be more disabling than extensive disease elsewhere.
What to screen for
Two things beyond the skin. Psoriatic arthritis, because it is common, because joint damage is irreversible, and because it changes drug choice — several agents treat both, and not all do [14] [20]. And cardiometabolic risk, because the association is consistent and the risk is modifiable [3].
Centerpiece: a simple simulatable model of what targeted therapy delivers
Comparing drug classes across separate trials is a well-known way to get the wrong answer: populations, eras and endpoint definitions all differ. So everything here comes from one randomised trial that ran an anti-IL-23 antibody, an anti-TNF antibody and placebo alongside each other and reported the same outcomes for all three at three timepoints out to a year [4].
At week 16, PASI 90 was achieved by 73.3 percent on the anti-IL-23 antibody, 49.7 percent on the anti-TNF antibody, and 2.9 percent on placebo [4].
Two checks, on relationships the arithmetic was not handed. The trial reports two outcome measures, and for all six arm-timepoint pairs the less stringent one (investigator assessment of clear or minimal skin) exceeds the more stringent one (PASI 90) — as it must. And the week-16 ordering — anti-IL-23 above anti-TNF above placebo — holds independently on both measures.
The result the arithmetic produces is not the one you would guess. The advantage of the newer class over the older does not fade as both settle into maintenance; it widens. The gap goes from 23.6 percentage points at week 16 to 27.2 at week 24 to 28.4 at week 48, because the two arms move in opposite directions: the anti-IL-23 response rises from 73.3 to 76.3 percent while the anti-TNF response drifts from 49.7 down to 47.9. The number needed to treat for one additional PASI 90 response accordingly falls from 4.2 to 3.5 over the year.
That is the teaching point. Blocking a cytokine that sits at the top of the causal chain, rather than a broadly acting inflammatory mediator, produces both deeper clearance and better durability — and the durability difference is invisible at the twelve-to-sixteen-week timepoint at which drugs are usually licensed and compared.
The placebo column matters too, and is unusual. At 2.9 percent PASI 90 [4], the placebo response in psoriasis is small — quite unlike the large placebo effects seen in migraine prevention or depression. That is a consequence of the outcome being an observed, scored physical sign rather than a reported symptom, and it means effect sizes in this field can be read more directly than in most of medicine.
Two honest limits. This is one trial comparing two specific molecules; agents differ within both classes, and the anti-IL-17 drugs — not represented here — achieve comparable clearance by blocking the next step down the same axis [18] [21] [22]. And PASI measures skin. It does not measure joints, and the drug that clears skin best is not automatically the right choice for someone whose dominant problem is arthritis.
Pillar 2: treatment
Topicals and light, which still treat most patients
Most psoriasis is limited, and most patients are treated with topical corticosteroids and vitamin D analogues, often combined. This is unglamorous, effective for limited disease, and constrained by adherence — applying an ointment to plaques twice daily indefinitely is a real burden.
Phototherapy — narrowband UVB, or psoralen plus UVA — remains useful for extensive disease, limited chiefly by the need to attend a clinic repeatedly.
The conventional systemics
Methotrexate, ciclosporin and acitretin were the systemic options for decades. All work; all were adopted from elsewhere; all carry organ toxicity requiring monitoring — hepatic for methotrexate, renal and hypertensive for ciclosporin, teratogenic and lipid effects for the retinoids, which were introduced for severe psoriasis in the 1970s [23]. They remain relevant where biologics are unavailable or unaffordable, which is most of the world.
The biologics, in the order the axis was drugged
Anti-TNF came first, borrowed from rheumatology. Etanercept [17], infliximab [24] and adalimumab [25] all improved psoriasis substantially, and etanercept also improved fatigue and depression scores [15] — an early signal that psoriasis is more than skin.
Anti-IL-12/23 was the first agent aimed at the axis itself. Ustekinumab targets the p40 subunit shared by IL-12 and IL-23, and in phase 3 outperformed placebo [26] [27] and, head-to-head, etanercept [28].
Anti-IL-17 blocks the effector cytokine. Secukinumab, an anti-IL-17A antibody, achieved PASI 75 in 81.6 percent of patients against 4.5 percent on placebo, and outperformed etanercept in a within-trial comparison [18]. Ixekizumab [29] [21] [30] and brodalumab, which blocks the IL-17 receptor rather than the cytokine [31] [22], followed.
Anti-IL-23 blocks the p19 subunit specific to IL-23, sparing IL-12 — a deliberate increase in selectivity over ustekinumab. Guselkumab [4] and risankizumab [19] are the results, and they are the agents in the centerpiece.
Oral targeted agents
Apremilast inhibits phosphodiesterase 4, raising intracellular cAMP and damping inflammatory cytokine production [32] [33]. It is less effective than the biologics and is oral, which for many patients is the trade they want.
Deucravacitinib inhibits TYK2, the kinase transducing IL-23 and type I interferon signalling — an oral drug acting on the same axis the antibodies target, one step inside the cell. Phase 2 data exist in psoriatic arthritis [34].
Psoriatic arthritis changes the calculus
Several agents treat both skin and joints, and the evidence is separate: etanercept [35], ustekinumab [36], secukinumab [37] and apremilast [38] [39] [40] all have psoriatic-arthritis trial data, and treatment recommendations are issued separately by rheumatology and dermatology-rheumatology groups [41] [42] [20]. Choosing a drug for someone with both means asking which problem dominates.
Safety, stated proportionally
These are immunosuppressive drugs and the risks are real but, for the IL-17 and IL-23 agents, notably modest — in the trial above, adverse event rates were comparable between guselkumab, adalimumab and placebo [4]. Anti-TNF carries the tuberculosis reactivation risk familiar from rheumatology [43]. The IL-23 axis is not immunologically idle either: people with genetic defects in IL-23-dependent interferon-γ immunity are susceptible to tuberculosis [44], which is a reason for vigilance rather than alarm, since therapeutic blockade is partial and reversible. Anti-IL-17 agents can unmask or worsen inflammatory bowel disease, which is the clearest instance of the same axis behaving differently in different tissues.
Pillar 3: what is unresolved
Why the same axis behaves differently in skin and gut. Blocking IL-17 clears psoriasis and can worsen inflammatory bowel disease, while blocking IL-23 helps both [45] [46]. That divergence is the most interesting unsolved problem in this field and bears directly on how the axis should be targeted in patients with both.
Durability and drug survival. The centerpiece shows the classes diverging over one year [4]. Whether that continues, and what determines loss of response, is not settled.
Who needs which drug. With several classes achieving high clearance, selection is largely by comorbidity, route and cost rather than by any predictor of individual response — the same gap that runs through rheumatoid arthritis and inflammatory bowel disease.
Whether clearing skin changes the comorbidities. Psoriasis is associated with cardiovascular disease [3]; whether suppressing the inflammation reduces cardiovascular events, rather than merely the plaques, is the question that would establish psoriasis as a systemic disease in the strongest sense.
Access. The clearance rates above are achieved with drugs that are expensive and injected. Most people with psoriasis worldwide are treated with topicals and, at best, methotrexate.
Dig deeper in lmmol
Psoriasis, rheumatoid arthritis and inflammatory bowel disease share more than a category: they share the anti-TNF agents, the IL-23 pathway, and the problem of choosing among mechanisms without a predictive test. The instructive divergence is that ustekinumab treats both psoriasis and Crohn's disease by blocking IL-23 [46] [26], while anti-IL-17 agents clear psoriasis and can worsen bowel disease — the same cytokine axis with opposite tissue consequences [45]. Multiple sclerosis completes the immune-mediated set with a fourth pattern, and thyroid disease offers the contrast of autoimmunity managed by replacing the hormone rather than suppressing the immune system. The cardiometabolic associations connect to heart failure, obesity and fatty liver disease [3]; the psychological burden, measurable and responsive to treatment [15], connects to depression. The full collection is at health.