Computed from this object's own state — a structural condition is unmet. This is not a fixed disclaimer: the conditions below are each testable, and a build in which they are met renders READY.
This page published a pooled odds ratio and had no card on the index at all. Two of its three trials register only continuous primaries.
In this section: Registration and administrative information (PROSPERO field set) · What this object claims to contain
| Review title | Incretin therapies in heart failure with preserved ejection fraction |
|---|---|
| Review question (PICO) | This page published a pooled odds ratio and had no card on the index at all. Two of its three trials register only continuous primaries. |
| Population | Adults with HFpEF (LVEF >=45%) and obesity (BMI >=30), NYHA II-IV |
| Intervention | Incretin-based therapy (semaglutide 2.4 mg, tirzepatide 15 mg) |
| Comparator | Placebo |
| Background / rationale | Not recorded — this object holds no background field, and an introduction generated without one would be argument that no source in this review supports |
| Eligibility criteria | not recorded on the page this object was built from |
| Information sources | 0 source layers, listed on the Extraction tab |
| Search strategy | The executed strings, datetimes, filters and hit counts are on the Search tab |
| Study selection process | Not recorded — no screening record was recoverable from the page this object was extracted from, so who screened and how cannot be stated |
| Number of screeners | Not recorded — this object records no screener count |
| Data extraction items | Extracted per trial on this object: registry id; primary publication; arms; per-arm counts and the published effect. Fields not listed here were not extracted. |
| Outcomes and prioritisation | Registered as continuous in two trials -- and pooled here as an event count; Change in KCCQ clinical summary score |
| Risk of bias method | Not recorded — no per-domain RoB-2 assessment exists yet |
| Synthesis methods | Pooled under random-effects with the REML random-effects, inverse-variance (tau-squared estimated as zero) estimator on the log scale. Whether they were prespecified is not recorded on this object. |
| Subgroup analyses | Not recorded — this object records no subgroup-analysis field at all, so whether any were prespecified is UNKNOWN -- which is not the same as none having been |
| Meta-bias assessment | Not recorded — no funnel, Egger or Peters value is held on this object, so no meta-bias assessment is claimed |
| Certainty assessment | Not recorded — no GRADE record was recoverable from the page this object was extracted from |
| Funding | Not recorded — no funding statement is recorded for this review |
| Competing interests | Not recorded — no declaration is recorded |
| Built | 2026-08-18T00:00Z |
| Schema | 2-authored-from-source |
not recorded on the page this object was built from
In this section: Extracted values, and where each came from · References · Included studies · Regulatory documents · Methods and guidance · Sources · Registered as continuous in two trials -- and pooled here as an event count · Registered as continuous in two trials -- and pooled here as an event count · What was measured · Pooled result · Contributing trials · Endpoint definitions, read from the registry · Named on this review, contributing nothing to its pool · Change in KCCQ clinical summary score · Change in KCCQ clinical summary score · What was measured · Pooled result · What this pool holds constant · Contributing trials · Endpoint definitions, read from the registry · Named on this review, contributing nothing to its pool · The methods rule governing this decision
One row per extracted value. Every row carries the value, the verbatim sentence it was read from, a resolvable link to the source, and whether the number was read or derived. Where any of those is absent the row says so rather than omitting the value.
| Trial / outcome | Value as extracted | Verbatim source sentence | Source links | Read or derived |
|---|---|---|---|---|
| STEP-HFpEF DM primary | no effect value held for this outcome | no source sentence recorded — this value cannot be checked against a quoted line here | NCT04916470 · source | DERIVED by us from not carried forward -- the estimate is withdrawn The counts above are what the page held for this trial. The endpoint definition beside them is verbatim from the registry, and is what the withdrawal turns on. |
| SUMMIT primary | no effect value held for this outcome | no source sentence recorded — this value cannot be checked against a quoted line here | NCT04847557 · source | DERIVED by us from not carried forward -- the estimate is withdrawn The counts above are what the page held for this trial. The endpoint definition beside them is verbatim from the registry, and is what the withdrawal turns on. |
| SUMMIT kccq_css_change | MD 6.9 (95% CI 3.3 to 10.6) | no source sentence recorded — this value cannot be checked against a quoted line here | NCT04847557 · source | DERIVED by us from the trial's own ClinicalTrials.gov results record READ, not derived: this is the registry's own between-arm difference for the registered primary outcome. |
| Layer | Source | How it was obtained |
|---|
| Comparator | placebo |
|---|---|
| Estimand | first event — random-effects |
| Effect scale | reported on the log scale; nothing is pooled on this outcome |
What this does not establish: not that incretin therapies do not help in heart failure with preserved ejection fraction. STEP-HFpEF, STEP-HFpEF DM and SUMMIT all report benefit on the outcomes they were designed around -- symptom burden measured by KCCQ, and body weight -- and SUMMIT additionally reports its heart-failure event composite. The trials here measure symptom scores and weight, and nothing on this page estimated those. What is confirmed: summit's stored counts ARE its registered event co-primary, and all three identities are correct. The replacement question is available and parked: all three trials share a KCCQ clinical-summary-score primary, so a mean difference synthesis of symptom benefit is squarely defensible and needs no estimand judgement -- only each trial's least-squares mean difference read from its results record. That is the second such case in this programme, after BOCOCIZUMAB_LIPID.
Registered as continuous in two trials -- and pooled here as an event count. k = 2.
A published synthesis of this question included 4 trials where this pool carries 2. Source: Musa & Musa, BMC Cardiovascular Disorders 2026, PMID 41906074, a prospectively registered systematic review (CRD420251237462).
Which trials are ABSENT is not established. The published abstract names none of its included studies and no included-study table was read, so this is a difference in stated counts and nothing more. It is not a claim that 2 specific trials were missed, and nobody here has identified them.
The estimate above is computed from the trials this object carries and is not withdrawn or altered by this note. What a reader is owed is the scope: a larger evidence base on this question exists in the literature, and whether this review's search should be widened is a decision that has not been made.
Authored from source. The outcome name inherited from the converter says 'Change in KCCQ-CSS' while the stored measure is an odds ratio over event counts -- the name and the measure describe different quantities.
What each trial COUNTED, in its own registry record's words, before anything here was pooled. A sentence saying what HAPPENED is not a sentence saying what was COUNTED, and only the second one licenses a pool. Read from the registry on 2026-08-18. Follow the registration link to disagree with any row.
| Trial | Rank in its own trial | Registered outcome measure | Description, verbatim | Analysis set / window |
|---|---|---|---|---|
| STEP-HFpEF DM NCT04916470 | PRIMARY -- primaryOutcomes[0] on this registration | Change in KCCQ (Kansas City Cardiomyopathy Questionnaire) clinical summary score | Score on a scale of 0 to 100 where the score 100 means the least burden for the participant. | Both registered primaries are continuous change scores. |
| SUMMIT NCT04847557 | PRIMARY -- primaryOutcomes[0] on this registration | Change From Baseline in the Kansas City Cardiomyopathy Questionnaire (KCCQ) Clinical Summary Score (CSS) | The KCCQ is a 23-item, participant self-administered questionnaire that assesses impacts of heart failure "over the past 2 weeks" on the following 7 domains: * Physical Limitation (6) * Symptom Stability (1) * Symptom Frequency (4) * Symptom Burden (3) * Self-Efficacy (2) * Quality of Life (3) * Social Limitation (4) Each of the 23 individual items are answered on Likert scales of varying lengths (5, 6, or 7-point scales). KCCQ-CSS includes the symptom and physical limitation domains of the KCCQ. Scores are obtained by averaging the associated individual items and transforming the score to a 0 to 100 range. Higher scores indicate better health status. Least Square (LS) mean was determined using ancova model with Baseline + HF Decompensation Within 12 Months of Screening + T2DM Status + Baseline BMI group (\<35, \>=35 kg/m2) + Treatment (Type III sum of squares) as variables . | The event co-primary is a first-occurrence composite. |
STEP-HFpEF dm is excluded from the replacement for a provenance reason, not a clinical one. Its ClinicalTrials.gov record carries no results section, so no registry-sourced KCCQ difference exists for it. Its published value exists in the literature and is deliberately not imported: every number on this page is read from a registry results record, and breaking that for one extra data point would cost the guarantee that makes the rest checkable.
| Trial | Why it contributes nothing | What the registry holds |
|---|---|---|
| STEP-HFpEF DM NCT04916470 | the registration carries no results section, so no registry-sourced value exists | nothing -- the registration has protocol data only |
| Comparator | placebo |
|---|---|
| Estimand | mean change from baseline — random-effects |
| Effect scale | pooled on the natural scale |
MD 7.43 (5.09 to 9.77), 95% interval
Change in KCCQ clinical summary score. random-effects, estimator REML random-effects, inverse-variance (tau-squared estimated as zero), k = 2.
τ² not stated · I² 0% · Q 0.1375 on 1 df
Favours: the intervention (higher is better)
Q = 0.1375 on 1 df, I-squared 0 per cent. At k=2 that means the two trials agree, not that the finding generalises.
This is a replacement, not a correction. The withdrawn estimate on this page was an odds ratio over event counts, and two of the three included trials register only continuous co-primaries. But all three register the same continuous outcome -- change in the KCCQ clinical summary score -- so a symptom-benefit analysis needs no estimand judgement at all. It is pooled at k=2, not k=3, and that is the honest number. STEP-HFpEF DM's ClinicalTrials.gov record carries no results section, so this project has no registry-sourced value for it. Its published KCCQ difference exists in the literature and is not used here, because every value on this page is read from a registry results record and importing one from elsewhere would break that guarantee for the sake of a third data point. The two values are the registry's own: step-HFpEF 7.8 (4.8 to 10.9) over 243 against 237, and SUMMIT 6.9 (3.3 to 10.6) over 301 against 313. Pooled: MD 7.43 (5.09 to 9.77) points on the KCCQ clinical summary score, k=2, 1,094 participants. Q = 0.14 on 1 degree of freedom, I-squared 0 per cent -- the two trials agree almost exactly.
What a reader should know. (1) k=2 and the interval shown is the wald one. At two studies a Hartung-Knapp interval uses a t-distribution on one degree of freedom. This project's interval is -7.74 TO 22.60 and it includes no effect. An earlier version of this sentence gave 1.80 to 13.06 and said the interval still excluded no effect; those figures are metafor's raw unfloored knha output, not this project's, and the conclusion drawn from them was wrong. The raw inflation factor here is 0.3708, so the unfloored adjustment narrows rather than widens -- the instability the house floor exists to prevent. Both are recorded. (2) heterogeneity is zero here, which at k=2 means the two trials agree, not that the finding is general. (3) the third trial is missing for a provenance reason, not a clinical one. STEP-HFpEF DM would very likely move this estimate very little; it is left out because its registry record has no results, and that is a rule about where numbers may come from rather than a judgement about the trial. (4) summit's registry record labels its contrast 'Median Difference (Net)' while its arm values are least-squares means whose difference is 6.83. The label and the quantity disagree in the registry. The reported value is used and the discrepancy is stated rather than resolved. (5) a seven-point KCCQ difference is a symptom benefit, not an event reduction. Five points is the usual threshold for a clinically meaningful within-patient change. Nothing here estimates an effect on death or hospitalisation -- SUMMIT registers a heart-failure event co-primary and this analysis does not use it.
Correction, and it reverses what this page said about its own sensitivity analysis. The caveat on this pool stated that the Hartung-Knapp interval 'gives 1.80 to 13.06 -- far wider, and arguably the honest interval at this k. It still excludes no effect.' the figures 1.80 TO 13.06 ARE metafor's raw unfloored knha output. This project's interval is -7.7429 TO 22.6028, and it includes no effect.
The raw variance-inflation factor on this pool is 0.3708 -- metafor's Hartung-Knapp standard error is barely a third of the random-effects one -- so the unfloored adjustment narrows the interval instead of widening it. That is the exact instability this project's floor exists to prevent, and it is why the raw figures looked reassuringly wide next to the Wald interval while being the product of a correction pointing the wrong way.
The point estimate and its wald interval are unchanged: MD 7.43 (5.09 to 9.77) points on the KCCQ clinical summary score, k = 2, 1,094 participants. What changes is a sentence about what the sensitivity analysis showed, which said the opposite of what it shows. A seven-point KCCQ difference remains a symptom benefit and this page still estimates nothing about death or hospitalisation.
| Dimension | Across the pooled cohorts | If it differs, why it is crossed anyway |
|---|---|---|
| effect measure | ESTABLISHED | both are the registry's own between-arm differences on the KCCQ clinical summary score |
| estimand | ESTABLISHED | all three included trials register the KCCQ clinical summary score; the two pooled here have registry results |
| Trial | Analysed treatment / control |
MD (95% CI), or events | Source of this cell |
|---|---|---|---|
| SUMMIT NCT04847557 | 301 / 313 | 6.9 (3.3 to 10.6) | : |
The replacement outcome. The withdrawn odds-ratio outcome is kept on this object so a reader can see what was retracted.
What each trial COUNTED, in its own registry record's words, before anything here was pooled. A sentence saying what HAPPENED is not a sentence saying what was COUNTED, and only the second one licenses a pool. Read from the registry on 2026-08-18. Follow the registration link to disagree with any row.
| Trial | Rank in its own trial | Registered outcome measure | Description, verbatim | Analysis set / window |
|---|---|---|---|---|
| SUMMIT NCT04847557 | PRIMARY | Change in Kansas City Cardiomyopathy Questionnaire Clinical Summary Score (KCCQ-CSS) | The registry reports this as the between-arm difference in change from baseline in the KCCQ clinical summary score. | As registered by each trial. STEP-HFpEF reports an 'Estimated Treatment Difference'; SUMMIT's registry record labels its contrast 'Median Difference (Net)' while its arm values are least-squares means (19.51 against 12.68, a difference of 6.83) -- the label and the quantity do not match in that record, and the value used here is the one the registry reports. |
STEP-HFpEF dm is excluded from the replacement for a provenance reason, not a clinical one. Its ClinicalTrials.gov record carries no results section, so no registry-sourced KCCQ difference exists for it. Its published value exists in the literature and is deliberately not imported: every number on this page is read from a registry results record, and breaking that for one extra data point would cost the guarantee that makes the rest checkable.
| Trial | Why it contributes nothing | What the registry holds |
|---|---|---|
| STEP-HFpEF DM NCT04916470 | the registration carries no results section, so no registry-sourced value exists | nothing -- the registration has protocol data only |
Decision: Cochrane Handbook 6.5.1 s10.10.4.4: REML is preferred to DerSimonian-Laird for the between-study variance, and is the RevMan default.
Cochrane Handbook sections: 10.10.4.4
followed
In this section: Registered as continuous in two trials -- and pooled here as an event count · Risk of bias · GOSH · Trial-sequential analysis · Meta-regression · Change in KCCQ clinical summary score · Forest plot · Visual abstract · Risk of bias · GOSH · Trial-sequential analysis · Meta-regression
⬇ SVG (vector) 1 KB
Only the vector format is offered for this figure, by choice: it is line art — discrete marks, glyphs and labels — which SVG reproduces exactly at any size, so a raster would add weight without adding fidelity. The build was able to rasterise; it was asked not to.
Not drawn, because there is nothing to draw it from.
⬇ SVG (vector) 1 KB ⬇ PNG 3500x540 81 KB
All raster formats were generated at build time from the same SVG as the graphic above (SHA-256, first 16 hex characters: 9091822e99aa8781), at 600 dpi for a 6-inch print width, on a white ground. They are not screenshots of the page and do not change with the theme you are reading in.
Submission formats written to the figures directory, not embedded here because an uncompressed 600 dpi TIFF is tens of megabytes: EPS (vector, journal's first choice for line art) — gosh.eps, 2 KB; TIFF (uncompressed, 600 dpi -- journal fallback) — gosh.tiff, 5,537 KB.
Deliberately not drawn. Every subset meta-analysis of 2 trials is 3 points; a cloud that small cannot show the multimodality GOSH exists to reveal, and a reader would take the picture as evidence of its absence.
⬇ SVG (vector) 1 KB ⬇ PNG 3500x650 87 KB
All raster formats were generated at build time from the same SVG as the graphic above (SHA-256, first 16 hex characters: 25ed9800fdf56614), at 600 dpi for a 6-inch print width, on a white ground. They are not screenshots of the page and do not change with the theme you are reading in.
Submission formats written to the figures directory, not embedded here because an uncompressed 600 dpi TIFF is tens of megabytes: EPS (vector, journal's first choice for line art) — tsa.eps, 2 KB; TIFF (uncompressed, 600 dpi -- journal fallback) — tsa.tiff, 6,665 KB.
TSA boundaries depend entirely on a target information size that must be pre-specified. This review's protocol registers none, so any boundary drawn here would be a parameter chosen after seeing the data -- which is the practice TSA exists to protect against.
⬇ SVG (vector) 1 KB ⬇ PNG 3500x540 87 KB
All raster formats were generated at build time from the same SVG as the graphic above (SHA-256, first 16 hex characters: e35ec93bbc9a1c10), at 600 dpi for a 6-inch print width, on a white ground. They are not screenshots of the page and do not change with the theme you are reading in.
Submission formats written to the figures directory, not embedded here because an uncompressed 600 dpi TIFF is tens of megabytes: EPS (vector, journal's first choice for line art) — bubble.eps, 2 KB; TIFF (uncompressed, 600 dpi -- journal fallback) — bubble.tiff, 5,537 KB.
The protocol pre-specifies no moderator, and at k = 2 a meta-regression would be fitted on 2 points. Not drawn rather than drawn with a caveat: a bubble plot invites reading a slope, and there is no slope here that any reader should read.
Drawn from the same stored estimates the table above lists. Box area is proportional to inverse-variance weight.
x-axis range
⬇ SVG (vector) 1 KB ⬇ PNG 4500x920 93 KB
All raster formats were generated at build time from the same SVG as the graphic above (SHA-256, first 16 hex characters: 7dee28b6e7788827), at 600 dpi for a 6-inch print width, on a white ground. They are not screenshots of the page and do not change with the theme you are reading in.
Submission formats written to the figures directory, not embedded here because an uncompressed 600 dpi TIFF is tens of megabytes: EPS (vector, journal's first choice for line art) — forest_fit.eps, 3 KB; TIFF (uncompressed, 600 dpi -- journal fallback) — forest_fit.tiff, 12,129 KB.
Changing the range moves the axis window only. The guides stay labelled with the null and the extremes of the plotted intervals, so no plotted value and no printed number differs between these views — and that is checked at build time, not asserted. Ranges not offered because they would crop the data: 0.5 to 2, 0.25 to 4, 0.7 to 1.3.
⬇ SVG (vector) 2 KB ⬇ PNG 4500x2100 232 KB
All raster formats were generated at build time from the same SVG as the graphic above (SHA-256, first 16 hex characters: d0c29f9f2fc83ae9), at 600 dpi for a 6-inch print width, on a white ground. They are not screenshots of the page and do not change with the theme you are reading in.
Submission formats written to the figures directory, not embedded here because an uncompressed 600 dpi TIFF is tens of megabytes: EPS (vector, journal's first choice for line art) — visual-abstract.eps, 4 KB; TIFF (uncompressed, 600 dpi -- journal fallback) — visual-abstract.tiff, 27,685 KB.
Projected from the canonical object, so it carries the same k, the same pooled estimate and the same interval as the paper and cannot drift from them. The interval is drawn to scale against the no-difference line, which it EXCLUDES: 5.09 to 9.77, with no difference at 0.
⬇ SVG (vector) 1 KB
Only the vector format is offered for this figure, by choice: it is line art — discrete marks, glyphs and labels — which SVG reproduces exactly at any size, so a raster would add weight without adding fidelity. The build was able to rasterise; it was asked not to.
Not drawn, because there is nothing to draw it from.
⬇ SVG (vector) 1 KB ⬇ PNG 3500x540 81 KB
All raster formats were generated at build time from the same SVG as the graphic above (SHA-256, first 16 hex characters: 9091822e99aa8781), at 600 dpi for a 6-inch print width, on a white ground. They are not screenshots of the page and do not change with the theme you are reading in.
Submission formats written to the figures directory, not embedded here because an uncompressed 600 dpi TIFF is tens of megabytes: EPS (vector, journal's first choice for line art) — gosh.eps, 2 KB; TIFF (uncompressed, 600 dpi -- journal fallback) — gosh.tiff, 5,537 KB.
Deliberately not drawn. Every subset meta-analysis of 2 trials is 3 points; a cloud that small cannot show the multimodality GOSH exists to reveal, and a reader would take the picture as evidence of its absence.
⬇ SVG (vector) 1 KB ⬇ PNG 3500x650 87 KB
All raster formats were generated at build time from the same SVG as the graphic above (SHA-256, first 16 hex characters: 25ed9800fdf56614), at 600 dpi for a 6-inch print width, on a white ground. They are not screenshots of the page and do not change with the theme you are reading in.
Submission formats written to the figures directory, not embedded here because an uncompressed 600 dpi TIFF is tens of megabytes: EPS (vector, journal's first choice for line art) — tsa.eps, 2 KB; TIFF (uncompressed, 600 dpi -- journal fallback) — tsa.tiff, 6,665 KB.
TSA boundaries depend entirely on a target information size that must be pre-specified. This review's protocol registers none, so any boundary drawn here would be a parameter chosen after seeing the data -- which is the practice TSA exists to protect against.
⬇ SVG (vector) 1 KB ⬇ PNG 3500x540 87 KB
All raster formats were generated at build time from the same SVG as the graphic above (SHA-256, first 16 hex characters: e35ec93bbc9a1c10), at 600 dpi for a 6-inch print width, on a white ground. They are not screenshots of the page and do not change with the theme you are reading in.
Submission formats written to the figures directory, not embedded here because an uncompressed 600 dpi TIFF is tens of megabytes: EPS (vector, journal's first choice for line art) — bubble.eps, 2 KB; TIFF (uncompressed, 600 dpi -- journal fallback) — bubble.tiff, 5,537 KB.
The protocol pre-specifies no moderator, and at k = 2 a meta-regression would be fitted on 2 points. Not drawn rather than drawn with a caveat: a bubble plot invites reading a slope, and there is no slope here that any reader should read.
In this section: Summary of findings · Reproducibility artifact · Other recorded qualifications
| Outcome | k | Pooled effect | Certainty |
|---|---|---|---|
| Registered as continuous in two trials -- and pooled here as an event count | 2 | not pooled | See comment12 |
| Change in KCCQ clinical summary score | 2 | MD 7.43 (5.09 to 9.77) | See comment32 |
| Canonical object | — |
|---|---|
| Registered at | — — |
| Schema | 2-authored-from-source |
| Object last updated | 2026-08-18T00:00Z |
| Source object SHA-256, first 16 hex characters | dfa3fb6e4b318378 |
| Page generated | 2026-08-28 19:54 UTC |
| Generator build | a441e4187. The standard this page was built to is NOT ESTABLISHED: this object carries no build_stamp, so the version above is the standard that exists now, not the one this page was built against. |
This page CANNOT be rebuilt from what is recorded here: no canonical object is recorded and no registered commit is recorded. Each figure on the Analysis tab downloads as an SVG carrying exactly the values shown.
Statements this review records about the limits of what it establishes, which are not shown elsewhere on this page. They are reproduced from the object verbatim and are not summarised.
| Components as the registry states them | Two continuous co-primaries: 'Change in KCCQ clinical summary score' and 'Change in body weight'. Neither is an event count. The 7 and 18 stored on this page correspond to no registered primary outcome of this trial. |
|---|---|
| Components as the registry states them | a continuous KCCQ primary and 'First Occurrence of the Composite Endpoint of Heart Failure (HF) Outcomes'. The 29 and 52 stored here DO correspond to that event co-primary. |
| Components as the registry states them | KCCQ clinical summary score, change from baseline. A continuous symptom measure, higher is better. |
| Analysis set as the registry states it | As registered by each trial. STEP-HFpEF reports an 'Estimated Treatment Difference'; SUMMIT's registry record labels its contrast 'Median Difference (Net)' while its arm values are least-squares means (19.51 against 12.68, a difference of 6.83) -- the label and the quantity do not match in that record, and the value used here is the one the registry reports. |
| A a published synthesis carried more trials than this pool | A published synthesis of this question included 4 trials where this pool carries 2. Source: Musa & Musa, BMC Cardiovascular Disorders 2026, PMID 41906074, a prospectively registered systematic review (CRD420251237462). |
| B whether the extra trials can be named | Which trials are ABSENT is not established. The published abstract names none of its included studies and no included-study table was read, so this is a difference in stated counts and nothing more. It is not a claim that 2 specific trials were missed, and nobody here has identified them. |
| C what this means for the estimate above | The estimate above is computed from the trials this object carries and is not withdrawn or altered by this note. What a reader is owed is the scope: a larger evidence base on this question exists in the literature, and whether this review's search should be widened is a decision that has not been made. |
| Caveats | What a reader should know. (1) k=2 and the interval shown is the wald one. At two studies a Hartung-Knapp interval uses a t-distribution on one degree of freedom. This project's interval is -7.74 TO 22.60 and it includes no effect. An earlier version of this sentence gave 1.80 to 13.06 and said the interval still excluded no effect; those figures are metafor's raw unfloored knha output, not this project's, and the conclusion drawn from them was wrong. The raw inflation factor here is 0.3708, so the unfloored adjustment narrows rather than widens -- the instability the house floor exists to prevent. Both are recorded. (2) heterogeneity is zero here, which at k=2 means the two trials agree, not that the finding is general. (3) the third trial is missing for a provenance reason, not a clinical one. STEP-HFpEF DM would very likely move this estimate very little; it is left out because its registry record has no results, and that is a rule about where numbers may come from rather than a judgement about the trial. (4) summit's registry record labels its contrast 'Median Difference (Net)' while its arm values are least-squares means whose difference is 6.83. The label and the quantity disagree in the registry. The reported value is used and the discrepancy is stated rather than resolved. (5) a seven-point KCCQ difference is a symptom benefit, not an event reduction. Five points is the usual threshold for a clinically meaningful within-patient change. Nothing here estimates an effect on death or hospitalisation -- SUMMIT registers a heart-failure event co-primary and this analysis does not use it. |
| Why the earlier refusal does not carry | This figure was refused earlier, on the ground that every number in this object was registry-sourced and breaking that for one data point costs the guarantee that makes the rest checkable. That reasoning was about an unsourced import. The source is now identified, quotable and tiered, and the mixed provenance is declared per trial rather than hidden. The refusal was right about an unsourced number and does not extend to a sourced one. |
| What this does not establish | Not that the two trials are free of bias -- no RoB-2 assessment is recorded. Not that tirzepatide and semaglutide are interchangeable: this pools two different incretin agents against placebo on one instrument, which is a decision about the question being asked and is visible on the face of the topic. Not that the abstract figure equals the full-text table figure. |
| Why the test is the title and not the disclosure | Disclosure does not make an incoherent pool coherent. Four cardiology topics closed on exactly this point -- not whether the combination was disclosed, but whether the question the page asks is the question the pool answers. Had this been titled for one agent, declaring the second would not have fixed it; the title would have had to change, as azilsartan's did. |
| What the pool does not establish about the class | It does not establish that the class is homogeneous. I-squared is 0% and tau-squared is zero, and AT k=2 that is weak evidence of agreement rather than strong evidence of it: with two studies a heterogeneity statistic cannot distinguish genuine similarity from insufficient data to detect a difference. The two point estimates (6.83 and 7.30) are close, which is consistent with a class effect and does not demonstrate one. A reader should take this as a class-level estimate of two agents, not as evidence that incretins behave alike on this instrument. |
| What it does do for our identity check | The registration identity was previously checked in one direction only -- from the registry to the trial -- because the nejm paper could not be opened. This paper prints, in its own abstract, that STEP-HFpEF DM is registered as NCT04916470. That is an independent published source tying the trial name to the registration, which strengthens the identity materially. It is still not the check that was wanted: it does not establish which NCT is printed on the nejm paper the 7.3-point figure was quoted . That remains unverified. |
| What this comparison does not establish | Not that our pooled KCCQ estimate is right or wrong -- their paper does not report the quantity we pool. Not that semaglutide and tirzepatide behave alike on heart-failure events; their analysis contains no tirzepatide. Not that a symptom-score benefit implies an event benefit, or the reverse: that inference is not licensed by either synthesis and neither paper claims it. |
| A the sensitivity interval this page quoted was not this project s | Correction, and it reverses what this page said about its own sensitivity analysis. The caveat on this pool stated that the Hartung-Knapp interval 'gives 1.80 to 13.06 -- far wider, and arguably the honest interval at this k. It still excludes no effect.' the figures 1.80 TO 13.06 ARE metafor's raw unfloored knha output. This project's interval is -7.7429 TO 22.6028, and it includes no effect. |
| B why the raw interval is the wrong one here specifically | The raw variance-inflation factor on this pool is 0.3708 -- metafor's Hartung-Knapp standard error is barely a third of the random-effects one -- so the unfloored adjustment narrows the interval instead of widening it. That is the exact instability this project's floor exists to prevent, and it is why the raw figures looked reassuringly wide next to the Wald interval while being the product of a correction pointing the wrong way. |
| Scope | primary: first heart-failure hospitalisation. Secondary: KCCQ clinical summary score, six-minute walk distance, weight, all-cause mortality. |
| The finding of this comparison | They pooled where we declined, which is the opposite of class 76. This topic was named in advance as a predicted class-76 instance on the grounds that it pools two different drugs -- semaglutide and tirzepatide. The published synthesis pools that same class mix across more trials, in a prospectively registered review, and pools the first-heart-failure-hospitalisation outcome that this object declares primary and leaves unpooled. The prediction was wrong and is recorded as wrong. Whether this object should pool its primary is a content decision and is not made here. |
| What this review does not publish | What this review does not publish. This review also examined one outcome for which no pooled estimate is published: Registered as continuous in two trials -- and pooled here as an event count (k=2) -- two of the three trials on this page register only continuous primary outcomes, and this P. Each of those outcomes is recorded on this review with the reason it was not pooled, and none of them contributes to the estimate reported above (7.43 at k=2). A reader meeting the pooled result should not take it as covering every outcome, or every trial, this review examined. This sentence exists because a second, independent reading -- by a different model family, given the object and the abstract and none of our conclusions -- found that a reader would reasonably conclude the review pools all relevant trials, and that the object does not support that. The withdrawals were already recorded beside the outcomes they concern; they were invisible from the abstract, which is where the impression forms. |
In this section: Paper · Title · Abstract · Introduction · Methods — eligibility criteria · Methods — outcomes sought at every registered rank · Methods — synthesis · Results · Comparison with published syntheses · Limitations · Software availability · Note on registration · References · Figures · Trial characteristics · Not reported in this record
Each statement below cites the record it came from; those sources are listed at the end of every section. Where the record holds nothing, the section says so by name rather than being left out, so a procedure that was not carried out can be told from one that simply was never mentioned.
Incretin therapies in heart failure with preserved ejection fraction1
This object contradicts itself about which semaglutide trial contributed. `inputs.trials` names NCT04916470; the pooled outcome `KCCQ css change` carries NCT04788511. These are different trials. Read from ClinicalTrials.gov API v2 on 2026-08-21: NCT04788511 is STEP-HFpEF, registered condition 'Obesity'; NCT04916470 is STEP HFpEF DM, registered condition 'Heart Failure With Preserved Ejection Fraction (HFpEF) and Diabetes Mellitus, Type 2'. One enrols people with diabetes and the other excludes them. Which of the two supplied the pooled value is not decidable from what this object holds, and it is not guessed here. It is recorded so that a reader is not shown a trial list that disagrees with the trials behind the number, and so the next pass resolves it against the source the value was read from rather than by picking one.2
titleREGISTRATION_DISCREPANCY_2026_08_21Question. This page published a pooled odds ratio and had no card on the index at all. Two of its three trials register only continuous primaries.1
Methods. estimates were pooled under a random-effects model with the REML random-effects, inverse-variance (tau-squared estimated as zero) estimator.2
Results. Change in KCCQ clinical summary score: mean difference 7.43 (5.09 to 9.77) across 2 trials; 2 of these pooled outcome(s) carry a stated qualification -- a referral or a recorded finding -- given in full in Results and not reproduced here; and no statistical heterogeneity was detected (I-squared 0%), though heterogeneity cannot be reliably assessed with two trials.3
questionresults.by_outcomeresults.by_outcomeBackground. This review asks: This page published a pooled odds ratio and had no card on the index at all. Two of its three trials register only continuous primaries. It pools 2 randomised trials, against placebo. The outcomes pooled are Registered as continuous in two trials -- and pooled here as an event count; Change in KCCQ clinical summary score.1
manuscript.introductionnot recorded on the page this object was built from1
screening.eligibilityEvery registered rank -- primary, secondary and other -- was read for 2 of 2 contributing trials before any decision about which outcomes could be combined; each records the time it was read.1
inputs.trials[].all_ranks_read_utcFor change in KCCQ clinical summary score, a random-effects model was fitted with the REML random-effects, inverse-variance (tau-squared estimated as zero) estimator.1
results.by_outcome.kccq_css_change.model, results.by_outcome.kccq_css_change.estimator_usedThe pooled mean difference for change in KCCQ clinical summary score was 7.43 (5.09 to 9.77) across 2 trials, against placebo.1
Noted on this pool. Correction, and it reverses what this page said about its own sensitivity analysis. The caveat on this pool stated that the Hartung-Knapp interval 'gives 1.80 to 13.06 -- far wider, and arguably the honest interval at this k. It still excludes no effect.' the figures 1.80 TO 13.06 ARE metafor's raw unfloored knha output. This project's interval is -7.7429 TO 22.6028, and it includes no effect. The raw variance-inflation factor on this pool is 0.3708 -- metafor's Hartung-Knapp standard error is barely a third of the random-effects one -- so the unfloored adjustment narrows the interval instead of widening it. That is the exact instability this project's floor exists to prevent, and it is why the raw figures looked reassuringly wide next to the Wald interval while being the product of a correction pointing the wrong way. The point estimate and its wald interval are unchanged: MD 7.43 (5.09 to 9.77) points on the KCCQ clinical summary score, k = 2, 1,094 participants. What changes is a sentence about what the sensitivity analysis showed, which said the opposite of what it shows. A seven-point KCCQ difference remains a symptom benefit and this page still estimates nothing about death or hospitalisation.2
The trials' results were closely consistent with one another: I-squared, the share of the variation between them that is more than chance alone would produce, was 0%; the heterogeneity test statistic Q was 0.138, on degrees of freedom 1. Q = 0.1375 on 1 df, I-squared 0 per cent. At k=2 that means the two trials agree, not that the finding generalises.3
The grounds for pooling are recorded as: This is a replacement, not a correction. The withdrawn estimate on this page was an odds ratio over event counts, and two of the three included trials register only continuous co-primaries. But all three register the same continuous outcome -- change in the KCCQ clinical summary score -- so a symptom-benefit analysis needs no estimand judgement at all. It is pooled at k=2, not k=3, and that is the honest number. STEP-HFpEF DM's ClinicalTrials.gov record carries no results section, so this project has no registry-sourced value for it. Its published KCCQ difference exists in the literature and is not used here, because every value on this page is read from a registry results record and importing one from elsewhere would break that guarantee for the sake of a third data point. The two values are the registry's own: step-HFpEF 7.8 (4.8 to 10.9) over 243 against 237, and summit 6.9 (3.3 to 10.6) over 301 against 313. Pooled: MD 7.43 (5.09 to 9.77) points on the KCCQ clinical summary score, k=2, 1,094 participants. Q = 0.14 on 1 degree of freedom, I-squared 0 per cent -- the two trials agree almost exactly. Where a later pool on this page rests on the same grounds, they are not restated.4
Registered as continuous in two trials -- and pooled here as an event count. These 2 trials are not pooled, and the reason is stated rather than the outcome being quietly omitted: Two of the three trials on this page register only continuous primary outcomes, and this page pooled event counts. Verbatim: STEP-HFpEF and STEP-HFpEF DM each register two co-primaries, 'Change in Kansas City Cardiomyopathy Questionnaire Clinical Summary Score' and 'Change in Body Weight'. Both are continuous change scores. Neither is an event count and neither can produce a 2x2 table. The counts stored for STEP-HFpEF DM, 7 of 310 against 18 of 306, correspond to no registered primary outcome of that trial. Summit is different and is the only row here with a defensible source: it registers a continuous KCCQ primary and 'First Occurrence of the Composite Endpoint of Heart Failure (HF) Outcomes', and its stored 29 of 364 against 52 of 367 do correspond to that event co-primary. So the pool joins one trial's registered event composite to another trial's unregistered counts. Separately: the outcome name and the measure describe different quantities. The outcome is named 'Change in Kansas City Cardiomyopathy Questionnaire Clinical Summary Score' and the stored measure is an odds ratio. A change score has no odds. Separately again, and it is why this topic was invisible: this page's card was invisible to our own tooling. The card existed and published 'HR 0.41 (0.22 to 0.79), k=3' -- a measure, a value and a trial count that all differ from the object's OR 0.4846 (0.3178 to 0.7389) at k=2. It was reported as having no card because two scripts matched card links with the pattern [A-Z0-9_]+\.html, which cannot match the lowercase 'p' in HFpEF. A page invisible to the tool that counts pages is worse than an uncounted one: it is counted as a different thing, and this card escaped card-alignment checking entirely for as long as those tools have run. And, recorded separately on the withdrawal itself: two of the three trials on this page register only continuous primary outcomes, and this page pooled event counts. Verbatim: STEP-HFpEF and STEP-HFpEF DM each register two co-primaries, 'Change in Kansas City Cardiomyopathy Questionnaire Clinical Summary Score' and 'Change in Body Weight'. Both are continuous change scores. Neither is an event count and neither can produce a 2x2 table. The counts stored for STEP-HFpEF DM, 7 of 310 against 18 of 306, correspond to no registered primary outcome of that trial. Summit is different and is the only row here with a defensible source: it registers a continuous KCCQ primary and 'First Occurrence of the Composite Endpoint of Heart Failure (HF) Outcomes', and its stored 29 of 364 against 52 of 367 do correspond to that event co-primary. So the pool joins one trial's registered event composite to another trial's unregistered counts. Separately: the outcome name and the measure describe different quantities. The outcome is named 'Change in Kansas City Cardiomyopathy Questionnaire Clinical Summary Score' and the stored measure is an odds ratio. A change score has no odds. Separately again, and it is why this topic was invisible: this page had no index card at all. It is linked from the index table and appears in no card, so it published a pooled odds ratio of 0.4846 (0.3178 to 0.7389) that the corpus's own card-alignment checking could never reach -- there was no card to compare against the page.5
Noted on this pool. A published synthesis of this question included 4 trials where this pool carries 2. Source: Musa & Musa, BMC Cardiovascular Disorders 2026, PMID 41906074, a prospectively registered systematic review (CRD420251237462). Which trials are absent is not established. The published abstract names none of its included studies and no included-study table was read, so this is a difference in stated counts and nothing more. It is not a claim that 2 specific trials were missed, and nobody here has identified them. The estimate above is computed from the trials this object carries and is not withdrawn or altered by this note. What a reader is owed is the scope: a larger evidence base on this question exists in the literature, and whether this review's search should be widened is a decision that has not been made.6
The full model output for every pooled outcome -- the call, the estimator, the heterogeneity statistics and the back-transformed interval -- is stored verbatim on this review object at `results.by_outcome.<outcome>.r_output.verbatim`. It is not reproduced as an Extended data section in this article; an earlier version of this sentence said it was, and no such section exists.7
outcomes[id=kccq_css_change].name, results.by_outcome.kccq_css_change.pooled, outcomes[id=kccq_css_change].comparatorresults.by_outcome.kccq_css_change.POOL_FINDINGS_2026_08_21results.by_outcome.kccq_css_change.heterogeneity.i2, results.by_outcome.kccq_css_change.heterogeneity.q, results.by_outcome.kccq_css_change.heterogeneity.df, results.by_outcome.kccq_css_change.heterogeneity_statusresults.by_outcome.kccq_css_change.poolable_reasonoutcomes[id=primary].name, results.by_outcome.primary.poolable_reason, results.by_outcome.primary.pooled.withdrawn_reasonresults.by_outcome.primary.POOL_FINDINGS_2026_08_21results.by_outcomePublished syntheses were identified as follows. PubMed E-utilities, executed 2026-08-20. Query, counts and per-record disposition in ssot/incretin-hfpef-review/appraisal/the screen of published syntheses.json.1
published_comparison._how_identifiedEligible trials that do not contribute: STEP-HFpEF dm is excluded from the replacement for a provenance reason, not a clinical one. Its ClinicalTrials.gov record carries no results section, so no registry-sourced KCCQ difference exists for it. Its published value exists in the literature and is deliberately not imported: every number on this page is read from a registry results record, and breaking that for one extra data point would cost the guarantee that makes the rest checkable.1
This review does not claim that any per-trial count was checked against a results record.2
1 claim(s) previously made about this review were corrected and the corrections are retained on the object rather than overwritten.3
What this review does not give a clinician: the number of events in each arm, any absolute effect or number needed to treat and any measure of harm. A relative effect cannot be turned into the benefit or the risk facing one patient without them, so this review can support a judgement about what these trials found and not a judgement about what to do.4
eligible_but_not_contributing.noteverification_basis.what_is_not_claimedclaims_correctedresults.by_outcomeAnalyses were computed under R 4.6.0 with metafor 5.0.1.1
results.by_outcomeNo PROSPERO registration or protocol record is held on this object. That is not the same as knowing this review was not registered, and neither claim is made. It is recorded here as an absence so that a reader does not have to infer it from silence.1
bookkeeping_2026_08_21.whether_this_review_was_prospectively_registered| Trial | Registration | Registry | Read |
|---|---|---|---|
| NCT04788511 | NCT04788511 | ClinicalTrials.gov | |
| NCT04847557 | ClinicalTrials.gov | ||
| NCT04916470 | ClinicalTrials.gov |
| Citation | Identifier |
|---|---|
| GLP-1 and dual GIP/GLP-1 agonists in obese patients with HFpEF: a systematic review and meta-analysis of RCTs. BMC Cardiovascular Disorders. 2026. | PMID: 41906074 |
| Cited for | Reference |
|---|---|
| the model fitted, as R 4.6.0 with metafor 5.0.1 | Viechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat Softw. 2010;36(3):1-48. doi:10.18637/jss.v036.i03 |
| the reporting guideline this manuscript is written against | Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. doi:10.1136/bmj.n71 |
manuscript.referencesmanuscript.referencesmanuscript.referencesresults.by_outcome.kccq_css_change.per_trial, results.by_outcome.kccq_css_change.pooledresults.by_outcome.kccq_css_change.per_trial, results.by_outcome.kccq_css_change.pooledresults.by_outcome.primary.pooledresults.by_outcome.primary.pooled| Registration | Trial | Participants | Randomised | Primary outcome measured over | Design |
|---|---|---|---|---|---|
| NCT04847557 | SUMMIT | registered condition: Obesity; Heart Failure With Preserved Ejection Fraction (HFpEF) | not extracted | Baseline, Week 52; Baseline Up To 160 weeks | Tirzepatide (treatment, n=364) vs Placebo (control, n=367) [+2 further fields recorded on the object] |
| NCT04916470 | STEP-HFpEF DM | not recorded | not extracted | From baseline (week 0) to end of treatment (week 52) | Semaglutide 2.4 mg once weekly (OW) (treatment, n=310) vs Semaglutide placebo OW (control, n=306) [+2 further fields recorded on the object] |
inputs.trialsThe items below were not written because the record does not hold what they would be composed from. They are named here so that an absence is not mistaken for an omission.
| Article area | Item not reported | Reason |
|---|---|---|
| Abstract | The Background sentence of the abstract | Background is argument -- why this question matters -- and no arrangement of the quantities this object holds yields it. It is emitted only from an authored field and there is no composition path for it here |
| Abstract | The Conclusions sentence of the abstract | A conclusion is argument -- what a reader should now believe -- and it is emitted only from an authored field. Composing one from the estimate would be this renderer telling the reader what to conclude, which no field supports |
| Methods — search | A database search for primary trials, which was not carried out for this review; the included trials were identified from named registrations | No further reason is recorded. |
| Methods — synthesis | The model/estimator sentence for registered as continuous in two trials | And pooled here as an event count -- no pooled estimate was produced for this outcome, so no model was fitted and none is described |
| Methods — synthesis | The claim that records were screened in duplicate by two independent reviewers | No further reason is recorded. |
| Methods — synthesis | The claim that risk of bias was assessed with a named tool | No further reason is recorded. |
| Methods — synthesis | The claim that certainty of evidence was graded | No further reason is recorded. |
| Methods — synthesis | The claim that the review methods were prespecified before data collection | No further reason is recorded. |
| Note on registration | The registration note | No further reason is recorded. |
UNSTAMPED — unknown-version. This page carries no build_stamp, so the standard it was built to cannot be established. That is a statement about the record, not a claim that the page is below standard.
Every number on this page is projected from a single canonical object, and each measured cell names the source analysis it was read from. That is what is machine-checked. It does not establish that the underlying sources are right, only that this page faithfully reports them.