COA and PRO Instruments in FDA GLP-1 Obesity and CV Programs

COA and PRO instruments in the FDA record for GLP-1 obesity and cardiovascular programs

Clinical outcome assessments (COAs) and patient-reported outcomes (PROs) occupy a specific, well-bounded role in the FDA files for GLP-1 (and dual GIP/GLP-1) products used in weight management and cardiovascular risk reduction. Across the semaglutide (Wegovy), liraglutide (Saxenda), and tirzepatide (Zepbound) programs, the same short list of instruments recurs, physical-functioning measures dominate the confirmatory hierarchy, and FDA reviewers repeatedly draw a line between statistical significance and demonstrated clinical meaningfulness. This overview walks through what actually appears in the approved labeling and in the clinical and statistical reviews, product by product, and then draws out the cross-cutting regulatory themes.

The recurring instrument set

Four PRO instruments carry most of the weight-management COA burden, supplemented by symptom, appetite, and psychiatric-safety measures:

SF-36 / SF-36v2 Physical Functioning subscale (36-item Short Form Health Survey, version 2, acute recall) — a generic health-related quality-of-life measure; the physical functioning domain is the piece FDA and sponsors elevate into the efficacy hierarchy ¹⁶⁹¹¹⁹¹³⁹.
IWQOL-Lite-CT (Impact of Weight on Quality of Life-Lite for Clinical Trials version, v3.0) — the obesity-specific, weight-related quality-of-life instrument, with a Physical Function composite that is the analytic focus ¹¹⁹⁹⁴¹³⁹.
EQ-5D-5L (EuroQol five dimensions, five levels), including the index score and the VAS — a generic preference-based health-status measure used as a supportive or exploratory endpoint ⁹²¹⁰²¹³⁹.
PHQ-9 and C-SSRS — the 9-item depression questionnaire and the Columbia-Suicide Severity Rating Scale, used consistently as psychiatric-safety instruments (and, in some programs, an eligibility screen) rather than efficacy endpoints ²¹⁰¹⁶⁴⁶⁸⁸⁷.

Program-specific additions include the Weight Related Signs and Symptoms Measure (WRSSM), a Visual Analogue Scale for appetite and the Control of Eating Questionnaire (semaglutide), TRIm-Weight and the DTSQs (liraglutide), and, for the obstructive sleep apnea indication, the PROMIS sleep short forms and a family of Patient Global Impression scales (tirzepatide).

Semaglutide (Wegovy): STEP obesity program and the SELECT cardiovascular trial

Wegovy carries the most developed PRO story of the class.

Physical functioning as the confirmatory PRO. In the original chronic weight management program (the STEP trials, FDA trial numbers 4373/4374/4375/4376), both the SF-36v2 Physical Functioning subscale and the IWQOL-Lite-CT Physical Function score sat in the confirmatory secondary endpoint hierarchy ¹⁶⁹¹⁶⁰¹⁰¹. In STEP 1 (Trial 4373), SF-36 physical functioning improved by a mean of +2.21 with semaglutide 2.4 mg versus +0.41 for placebo (estimated treatment difference +1.80; 95% CI +1.16 to +2.45; p<0.0001), and IWQOL-Lite-CT Physical Function improved by +14.68 versus +5.24 (ETD +9.44; 95% CI +7.46 to +11.42; p<0.0001) ¹⁷². STEP 2 (Trial 4374) showed the same direction with smaller separations: SF-36 physical functioning ETD +1.52 (p=0.0061) and IWQOL-Lite-CT Physical Function ETD +4.83 (p=0.0018) ¹⁷⁰.

FDA's clinical-significance caveat. The statistical reviewer flagged that although the between-group PRO differences were statistically significant, their clinical significance was uncertain, because raw item- and domain-level score changes were minimal ⁹⁴⁹⁶. That reservation is the central regulatory tension in this space and recurs across all three products.

The SELECT cardiovascular readout. In the 2024 SELECT review supporting the cardiovascular risk-reduction indication, the PRO panel shifted to generic and symptom measures. The EQ-5D-5L (index and VAS) appeared among additional supportive secondary endpoints (baseline EQ-5D-VAS 77.15 in both arms) ⁹²¹⁰², and the Weight Related Signs and Symptoms Measure (WRSSM) was an exploratory endpoint ¹⁰². Neither generated a headline efficacy claim; SELECT rested on the confirmed MACE endpoint, not on PRO data.

Appetite and eating behavior. In mechanism-oriented Trial 4455, a Visual Analogue Scale for appetite was a secondary endpoint and the Control of Eating Questionnaire an exploratory one; the review noted semaglutide 2.4 mg reduced appetite and energy intake but cautioned the exploratory results might not support labeling ¹⁰⁷.

Psychiatric safety. PHQ-9 and C-SSRS were used for psychiatric safety surveillance. PHQ-9 also functioned as an eligibility screen (randomization required a PHQ-9 score below 15) ¹⁶⁴. Reported C-SSRS/PHQ-9 signals were low and broadly balanced against placebo across the phase 3 trials ²¹⁰⁹⁵⁹³.

Liraglutide (Saxenda): the obesity PRO template

Saxenda's file predates the IWQOL-Lite-CT era and uses the earlier obesity instrument plus generic and treatment-satisfaction measures.

IWQOL-Lite (original version) — obesity-specific quality of life across physical function, self-esteem, sexual life, public distress, and work domains; a PRO endpoint whose improvements were reported as driven mainly by physical function ³⁶³⁸. In Trial 1922, liraglutide 3.0 mg improved the physical function domain (+4.92, p=0.0006) and total score (+2.75, p=0.0136) versus placebo ⁸³.
SF-36 — generic health status; in Trial 1839 the sponsor reported significant gains in overall physical and mental health domains, with domain-level estimated differences favoring liraglutide (for example physical functioning +1.57, bodily pain +1.88, general health +1.87) ³⁶³⁷.
TRIm-Weight — a treatment-impact/satisfaction measure (daily life, weight management, treatment burden, side-effect experience, psychological health); in Trial 1839 the weight-management score difference was +18.10 ³⁸³⁷.
DTSQs (Diabetes Treatment Satisfaction Questionnaire, status version) — in Trial 1922, liraglutide 3.0 mg improved the total score by +1.44 (p=0.0066) ⁸³.
Sleep-apnea instruments — in Trial 3970 (a sleep apnea study), the Functional Outcomes of Sleep Questionnaire (FOSQ) and the Epworth Sleepiness Scale (ESS) were patient-reported outcomes; most FOSQ domains were not statistically significant, and ESS improved similarly in both arms (-2.52 liraglutide vs -2.33 placebo) ⁶⁶⁶⁵.
PHQ-9 and C-SSRS — psychiatric-safety instruments; PHQ-9 totals and question-9 (self-harm) positivity were comparable between Saxenda and placebo over the multi-year program ⁷⁵⁶⁹⁷⁰.

A recurring FDA observation in this file is that the PRO endpoints in Trial 1839 were not pre-specified in the endpoint hierarchy and were not adjusted for multiplicity, limiting the strength of any labeling claim ⁸².

Tirzepatide (Zepbound): SURMOUNT obesity program and the OSA indication

The tirzepatide review is the most methodologically detailed on COA measurement.

SF-36v2 Physical Functioning as the lead COA. In SURMOUNT-1 and SURMOUNT-2, the SF-36v2 Physical Functioning domain was the efficacy COA carried forward; the IWQOL-Lite-CT Physical Function composite and EQ-5D-5L were treated as exploratory ¹³⁹. FDA reported that Week-72 change was statistically significant in SURMOUNT-1 and nominally significant in SURMOUNT-2, but that treatment-placebo differences were small and baseline scores sat near the general-population mean, producing ceiling effects that left little room to improve ¹³⁹²⁰⁰.

A responder / impaired-subgroup strategy. Because of the ceiling problem, FDA and the sponsor focused on the subgroup with impaired physical function at baseline. In that subgroup, Week-72 SF-36v2 Physical Functioning change favored tirzepatide: SURMOUNT-1 treatment difference +3.1 (95% CI 1.2, 5.0); SURMOUNT-2 differences +2.8 and +3.0 for the 10 mg and 15 mg doses ²⁰¹.

PGIS as the anchor. The Patient Global Impression of Status/Severity for physical activity (PGIS) served as the anchor measure to define meaningful change and the impaired-function subgroup, rather than as an efficacy endpoint itself. Responsiveness analyses tied PGIS improvement to medium-to-large SF-36v2 gains (effect sizes 0.77 to 1.23) ¹⁹⁴¹⁹¹¹⁹³. FDA noted EQ-5D-5L, as a generic preference-based measure, lacked content validity for estimating clinical benefit here ¹³⁹.

The obstructive sleep apnea indication. For the OSA program, tirzepatide used the PROMIS Sleep-Related Impairment short form (PROMIS-SF-SRI 8a) and the PROMIS Sleep Disturbance short form (PROMIS-SF-SD 8b), anchored by a set of Patient Global Impression of Severity scales for OSA (PGIS-OSA Sleepiness, PGIS-OSA Fatigue, and PGIS/PGIS-OSA Sleep Quality). FDA worked through anchor-based meaningful-change thresholds, accepting a 1-category improvement on the PGIS-OSA anchors and concluding the PROMIS-SF-SRI 8a treatment effect supported clinically meaningful improvement, while declining the sponsor's single proposed threshold for the sleep-disturbance form ³⁵.

The cardiovascular-outcome contrast: Ozempic and semaglutide in type 2 diabetes

Cardiovascular outcome programs in the class are notable for how little named PRO content they contain. In the Ozempic (semaglutide, type 2 diabetes) reviews, the efficacy endpoints were glycemic (HbA1c change and targets), body weight, and, for cardiovascular benefit, time to first adjudicated MACE ¹⁸⁵. Where the review mentioned a "diary or PRO questionnaire" or listed "PRO" among secondary objectives, no specific instrument (no SF-36, EQ-5D, or similar) was named in the documents reviewed ⁶⁴¹⁴². The practical lesson is that cardiovascular-outcome claims in this class have been built on hard adjudicated events, with PRO instruments playing at most a supportive role (as in Wegovy's SELECT, where EQ-5D-5L and WRSSM were supportive/exploratory only) ¹⁰².

Cross-cutting regulatory themes

Physical functioning is the anchor construct. Across semaglutide, liraglutide, and tirzepatide, the SF-36 (or SF-36v2) Physical Functioning subscale and the IWQOL-Lite-CT / IWQOL-Lite physical-function domains are the PROs that reach the confirmatory or lead-COA position; other domains and generic indices stay supportive or exploratory ¹⁶⁹¹¹⁹¹³⁹³⁶.
Statistical significance versus clinical meaningfulness. FDA reviewers repeatedly accepted that PRO differences were statistically significant while questioning their clinical significance, citing minimal raw score change and ceiling effects at near-normal baseline function ⁹⁴⁹⁶¹³⁹²⁰⁰.
Anchor-based interpretation is now expected. The tirzepatide reviews show FDA driving toward anchor-based meaningful-change analyses (PGIS anchors, impaired-function subgroups, responsiveness effect sizes) rather than accepting mean between-group differences at face value ²⁰¹¹⁹⁴¹⁹¹³⁵.
Multiplicity and pre-specification matter. The liraglutide file's caution that PRO endpoints were not pre-specified or multiplicity-controlled in a pivotal trial is a reminder that PRO labeling claims live or die on hierarchy placement ⁸².
Psychiatric-safety instruments are standard, not efficacy tools. PHQ-9 and C-SSRS appear across the obesity programs as safety surveillance (and occasionally eligibility screening), not as endpoints supporting benefit claims ²¹⁰¹⁶⁴⁶⁸⁷⁵.

A reader building a PRO strategy for a new incretin-based weight or cardiometabolic program would reasonably want to go deeper on the exact endpoint hierarchies and multiplicity plans per trial, the IWQOL-Lite-CT psychometric and content-validity dossiers FDA relied on, and how the OSA anchor-based thresholds were finally resolved. Those are all natural follow-up questions to put to Rhizome directly.