Itch, Pain and Quality-of-Life Claims in FDA Dermatology Drug Labeling

Itch, pain and quality-of-life claims in FDA dermatology labeling: what was granted, what was refused, and why

FDA review records for dermatology drugs and biologics show a consistent pattern. Single-symptom patient-reported outcomes (PROs) with a clear concept usually reach Section 14 of the label, provided they are prespecified, multiplicity-controlled and statistically significant. Worst-itch numeric rating scales (NRS) and symptom diaries with separate itch and pain items are the main examples. Broad quality-of-life (QoL) instruments are refused almost every time. The Dermatology Life Quality Index (DLQI), Children's DLQI (CDLQI) and disease-specific QoL scores all fall in this group. POEM, a multi-symptom frequency score rather than a QoL instrument, was also refused. Reviewers most often cite content validity, multi-barreled items and concepts that are too distal to interpret. Between these two groups sit partial or descriptive claims. These arise when the instrument is acceptable but the data are not: missing diary data, self-selected PRO subsets, or no defensible meaningful-change threshold.

The evidentiary standard FDA applies

FDA's PRO guidance says symptom or QoL claims can be supported when the evidence comes from a well-defined, reliable instrument used in an appropriately designed investigation. The standard of evidence is the same as for any other labeling claim ⁵⁴. Several conditions drive most of the dermatology decisions below:

• Claim-instrument fit. The instrument's concepts, conceptual framework, scoring and intended population must match the endpoint and the proposed claim ⁵⁷⁶².
• Content validity comes first. Qualitative evidence must show the items capture the concept that matters to patients. Other validity or reliability evidence cannot make up for inadequate content validity ⁵³⁶⁵.
• Trial design. The PRO should be a prespecified objective with the exact instrument version named in the protocol. Open-label trials are rarely adequate for PRO labeling claims ⁶³.
• Clinical meaningfulness. Statistical significance alone may not establish benefit. Disease-specific guidance also calls for enough patients with the symptom, at sufficient baseline severity, to show meaningful improvement ⁷⁶⁶⁶.
• Component claims. A claim about an individual domain of a multi-domain PRO generally requires that the domain was prespecified as a separate endpoint ⁷².

Summary of labeling outcomes

Itch claims FDA granted

The ≥4-point worst-itch NRS responder has become the default

Most granted itch claims use the same construct. The endpoint is the proportion of patients with a ≥4-point improvement on an 11-point worst-itch NRS, analyzed among patients with baseline scores of at least 4. The labeled responder rates for the atopic dermatitis products below are listed in Atopic dermatitis endpoints behind FDA approvals; this section covers how FDA decided each claim.

• Dupixent, atopic dermatitis. The label presents ≥4-point Peak Pruritus NRS responder rates for adults ²⁷², adolescents ²⁷⁰ and children ²⁷⁷. FDA had advised earlier that a ≥4-point reduction was appropriate and a ≥3-point reduction could be supportive ²⁶⁶. At BLA review, the COA reviewer found the Peak Pruritus NRS to be a well-defined and reliable measure of pruritus intensity, and found that cumulative distribution analyses supported the ≥4-point threshold ³. This followed an earlier FDA position that did not concur the pruritus NRS could support an itch claim. At that stage FDA cited unclear item wording (average vs worst itch), a 12-hour protocol recall that did not match the 24-hour instrument, and poorly described IVRS/IWRS administration comparability ¹⁰⁴.
• Dupixent, prurigo nodularis. Itch response was defined as a ≥4-point WI-NRS reduction. Week 24 results were 60.0% vs 18.4% in PRIME and 57.7% vs 19.5% in PRIME2. Composite WI-NRS plus IGA PN-S 0/1 results are also labeled ⁵⁴¹.
• Opzelura. The label reports Week 8 ≥4-point Itch NRS responder rates in patients with baseline ≥4 ³⁶⁶. Reviewers found the Itch NRS valid and reliable for worst itch. They also found that a 4-point change had been documented as meaningful to patients and that the treatment effect was clinically meaningful ³⁵⁵. Before the pivotal trials, FDA had said the endpoint could support labeling if it was multiplicity-controlled and enough subjects had baseline scores of at least 4 ³⁵⁴.
• Adbry. FDA confirmed that significant results for the prespecified pruritus endpoint could appear in labeling, but DLQI and SCORAD results could not ⁴⁷². The label reports ≥4-point Worst Daily Pruritus NRS responders at Week 16 for ECZTRA 1, 2 and 3 (ECZTRA 3 with TCS) ⁴⁸⁴.
• Ebglyss. The label reports Week 16 ≥4-point Pruritus NRS responder rates for ADvocate 1 and 2 ⁵¹⁵. During labeling review, FDA found that the applicant had changed the weekly NRS derivation before unblinding. The original rule required at least 4 of 7 daily ratings; the revised rule allowed as few as 1 of 7. FDA asked for a sensitivity analysis using the original rule so that the label would not be misleading ⁴¹⁶.
• Taltz. FDA initially rejected a ≥3-point responder definition. It cited uncertain clinical meaningfulness, no baseline-itch enrollment criteria, and the concern that the same change may mean different things at different baselines ³⁰². FDA also asked for responder analyses at each time point and replication in two trials ³⁰¹. The final COA review found the Itch NRS valid and reliable enough to support an itch-improvement claim. The reviewer still noted uncertainty about the 4-point threshold because it had been derived from a clinician-reported anchor, while observing robust cumulative distribution results ¹⁶⁸¹⁰⁵. The label includes "Itch NRS (≥4 improvement) in subjects with baseline Itch NRS ≥4" ⁶⁰⁹.
• Zoryve (psoriasis). FDA considered a single-item NRS acceptable for worst pruritus and a ≥4-point threshold reasonable. It rejected a Week 2 key secondary timepoint because labeling-intended secondary endpoints should be assessed at the primary timepoint (Week 8) ⁹³. Week 8 WI-NRS response was 66.7% vs 25.7% and 68.6% vs 33.3% ⁵³⁵. The Week 2 result was significant in only one of the two trials ¹¹⁵.
• Wynzora. Reviewers stated that pruritus NRS results would be included in labeling ¹²⁸. The label reports ≥4-point improvement at Week 4 in 60.3% vs 21.4% on vehicle ²⁰⁴.
• Rhapsido (chronic spontaneous urticaria). FDA recommended separate itch (ISS7) and hives (HSS7) co-primary endpoints so that an efficacy finding could not be driven by one component of UAS7 ⁴⁹⁸. The label reports statistically significant ISS7 improvement at Week 12 in both REMIX trials ⁵⁰⁵.

Itch claims FDA limited or refused

• Cibinqo. The label reports PP-NRS4 responder rates at Week 2. It does not report Week 12 PP-NRS percentages. Week 12 is a descriptive statement that a higher proportion of subjects achieved improvement in itching. FDA had objected to a Week 12 PP-NRS-only claim because the instrument was collected only at clinic visits and about 20% to 36% of Week 12 data were missing in the treatment arms. Reviewers accepted the descriptive Week 12 wording after daily PSAAD item 1 (itch) showed a consistent direction of effect ¹⁷¹⁴²⁰.
• Eucrisa. FDA said the sponsor had not defined the enrolled population by baseline pruritus severity and had not provided a PRO dossier. Without one, reliability, validity and ability to detect meaningful change were not established, and the pruritus assessment had "limited regulatory utility" ³¹. The endpoint was changed to exploratory and was not intended for labeling. The result was also not replicated across the two Phase 3 trials, and the 0.55-day difference was not clinically meaningful ⁴²¹⁴²⁵.
• Ultravate (halobetasol). For an itch scale intended for labeling, FDA required limited, clinically relevant, multiplicity-adjusted secondary endpoints. It also asked that the scale's response categories be non-comparative ⁴³⁰.

Pain claims

Skin pain has reached labels in psoriasis, usually as one item in a multi-symptom diary. In hidradenitis suppurativa (HS), where the pain endpoints failed statistically, it reached a label only as a descriptive statement: the Bimzelx label notes a worst skin (lesion) pain reduction at Week 16 without response rates ⁶²⁸.

• Siliq. The label reports that more patients achieved a PSI rating of 0 or 1 on every item, including itch and pain ³⁴⁸. FDA had asked for more support for pain because the original PRO dossier was internally inconsistent. The sponsor ran further qualitative interviews showing that patients reported and understood lesion pain, and the reviewer considered the concern resolved ¹¹⁰. The COA review found the PSI fit for purpose for a claim covering itching and pain ³⁴⁴.
• Cosentyx (psoriasis). The COA review found content validity, validity and reliability for the itch, pain and scaling diary items ²⁴⁴⁹¹. The clinical reviewer was more critical. Only about 40% of participants contributed diary data, there were no baseline symptom-severity entry criteria, and the PRO endpoints had been added by protocol amendment. The reviewer also judged the results not clinically meaningful as independent therapeutic targets ⁴⁸⁹. The label carries a limited statement: improvements in itch, pain and scaling were observed among the 39% of subjects who took part in PRO assessments ^495496497.
• Skyrizi. The COA review found the Psoriasis Symptom Scale (PSS) fit for purpose for a claim covering pain, itching, redness and burning ²⁸. The label states that about 30% of risankizumab patients reached PSS 0, compared with 1% on placebo ⁵⁵⁷.
• Tremfya. FDA found qualitative evidence that patients could distinguish pain from burning and stinging ¹¹⁴. It noted substantial quantitative overlap among these concepts and a 25.5% floor effect for pain ⁴⁴³⁴⁴⁰. The PSSD symptom domain was found fit for purpose ⁵⁸⁴. The label states that greater improvements in itch, pain, stinging, burning and skin tightness were observed at Week 16, and that more subjects than on adalimumab were symptom-free at Week 24 ⁶¹⁷.
• Bimzelx (psoriasis). FDA could not reliably set within-patient meaningful-change thresholds for the PSD items, for two reasons. The PGAP anchor had substantial Week 16 missing data and a mismatched recall period. DLQI item 1 was unsuitable as an anchor because it combines itch, pain and stinging in one question ⁵⁶⁷¹⁴⁶. Because the trials showed significant separation from placebo, including more patients reaching a score of 0 on all three symptoms, FDA agreed to a descriptive claim for itch, pain and scaling ¹¹¹¹⁴⁶. The label lists PSD itching, pain and scaling as Week 16 secondary endpoints ⁵⁹⁷.
• Bimzelx and Cosentyx (HS). The reviewer said the HSSDD worst skin pain item could in principle support labeling. However, pain was not significant in HS0003. In HS0004, failure of an earlier endpoint in the testing hierarchy meant significance could not be claimed for pain ²⁵. The reviewer also suggested that any pain wording be qualified as "lesion pain" ¹²³. The final Bimzelx label follows that approach: section 14.5 describes a reduction in worst skin (lesion) pain at Week 16 versus placebo, with no response rates ⁶²⁸. For Cosentyx, neither dose achieved significance on skin pain NRS30 in FDA's per-trial analysis, and the US HS label presents HiSCR50 only ⁶²⁵⁶²⁶. The EU SmPC, which pools the two trials, reports a statistically significant NRS30 response for the Q2W dose (36.6% vs 23.0% with placebo) ⁶²⁷.

Quality-of-life claims: near-uniform refusal

• DLQI. For Tremfya, FDA advised that DLQI is not a well-defined, reliable instrument because it does not measure a specific concept. It could be exploratory but could not support claims ¹⁵³. The Adbry, Skyrizi and Rhapsido reviews gave more detail: several items are irrelevant or insensitive to treatment effect in the target disease, the questions are multi-barreled, content validity is weak, and the concepts are complex and distal ^322549329498. In Skyrizi, risankizumab was statistically superior to ustekinumab on DLQI 0/1, but the multidisciplinary review concluded that "PSS, but not DLQI, was fit for purpose for labeling" ⁵⁴⁹.
• CDLQI. Pediatric psoriasis QoL results were not labeled. The Stelara review cited exploratory status without multiplicity adjustment ²²⁷. The Enbrel review cited the lack of validation ²³⁶. In acne, FDA told the Evoclin sponsor that DLQI/CDLQI would have no regulatory utility ²²⁹.
• POEM. FDA did not agree that POEM could support a Dupixent symptom claim. POEM was exploratory, and the link between symptom frequency and severity was not established. A summed score combining core symptoms with sleep disturbance was not interpretable; FDA considered sleep a disease impact to be scored separately. FDA also raised concerns about missing-item handling and limited language coverage ⁴¹⁹.
• Disease-specific QoL and sleep. For Bimzelx in HS, HiSQOL endpoints were exploratory with no formal hypothesis testing and were not labeled ¹²¹²⁵. For Opzelura, efficacy was not demonstrated on PROMIS sleep endpoints ³⁵⁹. Reviewers also found that the responder analysis included patients whose baseline scores made the 6-point improvement impossible to achieve ³⁵⁷.

Recurring themes in the reviews

1. Concept specificity drives acceptance. Single-concept items with 24-hour recall and a worst-severity anchor were accepted across programs ^35516893. Instruments that sum distinct concepts (POEM) or ask multi-barreled questions (DLQI) were rejected ⁴¹⁹⁵⁴⁹.
2. Baseline enrichment is expected. FDA repeatedly tied itch claims to analyses restricted to patients with baseline NRS ≥4 ³⁵⁴⁴⁷². Eucrisa and Taltz were both criticized for lacking baseline itch criteria ³¹³⁰².
3. Responder thresholds need patient-anchored support. FDA pushed back on ≥3-point cutoffs and questioned clinician-reported anchors for a patient-experienced symptom ³⁰²¹⁰⁵. Where no threshold could be set, the fallback was a descriptive claim or a "score of 0" analysis ¹⁴⁶⁵⁵⁷.
4. Data completeness can limit an otherwise valid instrument. Missing clinic-visit data kept Cibinqo's Week 12 itch result descriptive, while Week 2 PP-NRS4 rates were labeled. Self-selected PRO subsets (Cosentyx) and changes to diary derivation rules (Ebglyss) raised the same issue ^171489416.
5. Multiplicity and timepoint alignment are prerequisites. Exploratory endpoints cannot support claims ⁴¹⁹¹²¹. Labeling-intended secondary endpoints must be assessed at the primary timepoint and included in the testing hierarchy ⁹³²⁵.
6. Qualitative evidence resolves concept disputes. In Siliq and Tremfya, further patient interviews were enough to keep pain as a distinct labeled concept despite quantitative overlap or floor effects ^110114440.

Implications for development programs

Sponsors seeking itch or skin-pain labeling in dermatology should prespecify a single-item worst-severity NRS with 24-hour recall. They should enrich or stratify for baseline severity (typically NRS ≥4), place a ≥4-point responder endpoint at the primary timepoint within the multiplicity hierarchy, and collect it daily with a derivation rule agreed with FDA in advance. A DLQI, CDLQI or POEM endpoint can serve as a construct-validity anchor or support publications. The review record shows that these instruments are unlikely to support a Section 14 claim without substantial new content-validity evidence ^322549329. Composite symptom diaries can reach the label as descriptive statements, but the precise wording depends on whether FDA accepts a meaningful-change threshold ^146617557.