Transitivity in the incretin network — search strategy
Sources, search strings, screening flow and extraction procedure, reproduced so that the review can be repeated.
§2Search strategy
The search was executed on 14 June 2024 and re-run at each review. It is reproduced here in full so that it can be repeated. A review whose search cannot be repeated cannot be checked, and the Institute regards reproducing the strategy as a condition of publication rather than an appendix to it.
§2.1Sources searched
Table 2. Sources searched, with the coverage period of each.
| Source | Coverage | Records |
|---|---|---|
| MEDLINE (Ovid) | Inception to the search date | 489 |
| Embase (Ovid) | Inception to the search date | 169 |
| Cochrane Central Register of Controlled Trials | Current issue at the search date | 157 |
| ClinicalTrials.gov | All records, including those without posted results | 212 |
| WHO International Clinical Trials Registry Platform | All records | 402 |
| EU Clinical Trials Register | All records | 416 |
| Regulatory assessment reports (FDA, EMA, MHRA, PMDA) | Published review documents | 341 |
| Sponsor clinical study reports where obtainable | Requested; obtained only in part | 573 |
| Record counts are per source before de-duplication and sum to more than the de-duplicated total in §2.3. | ||
§2.2Search strategy
The strategy below is the MEDLINE (Ovid) version. Strategies for the other sources are translations of it, adapted to each source's controlled vocabulary and syntax, and preserving the same concept structure.
2 semaglutide.mp. OR tirzepatide.mp. OR liraglutide.mp. OR retatrutide.mp. OR cagrisema (cagrilintide with semaglutide).mp. [and further compound terms]
3 "NN9535".ti,ab,kf. OR "NNC0113-0217".ti,ab,kf. OR "LY3298176".ti,ab,kf. OR "tirzepatide (INN)".ti,ab,kf. OR "NN2211".ti,ab,kf. OR "liraglutide (INN)".ti,ab,kf. OR "LY3437943".ti,ab,kf. OR "retatrutide (INN)".ti,ab,kf. OR "CagriSema".ti,ab,kf. OR "cagrilintide 2.4 mg / semaglutide 2.4 mg".ti,ab,kf.
4 1 OR 2 OR 3
5 exp Obesity and overweight in adults/
6 obesity.ti,ab,kf.
7 5 OR 6
8 4 AND 7
9 randomized controlled trial.pt. OR controlled clinical trial.pt. OR randomi#ed.ti,ab. OR placebo.ti,ab. OR clinical trials as topic.sh. OR randomly.ti,ab. OR trial.ti.
10 exp animals/ NOT humans.sh.
11 8 AND 9 NOT 10
12 remove duplicates from the preceding set
No language restriction was applied. No date restriction was applied. Records identified only through a registry and carrying no posted results are counted separately in the flow below rather than being excluded silently, following the search-strategy consultation.
§2.3Screening flow
Table 3. Screening flow from records identified to studies included.
| Stage | Records |
|---|---|
| Records identified from database searching | 1,719 |
| Records removed as duplicates | 539 |
| Records screened on title and abstract | 1,180 |
| Records excluded at title and abstract | 1,124 |
| Full-text reports assessed for eligibility | 56 |
| Full-text reports excluded, with reasons | 32 |
| Studies included in the qualitative synthesis | 24 |
| Studies included in the quantitative synthesis | 0 |
| Two reviewers screened independently at both stages. Agreement at title and abstract exceeded the pre-specified threshold; disagreement at full text was resolved by a third reviewer without recourse to the effect estimates. | |
§2.4Data extraction and certainty assessment
- Extraction was performed independently in duplicate onto a piloted form. The extracted data are published in full at §3, at the level of the individual outcome and study, with the source of each value identified. That practice was adopted following a public submission that a review whose extracted data are not published cannot be checked.
- Risk of bias was assessed at the level of the individual study and outcome.
- Certainty was assessed across the five domains recorded at §4, with the reason for each downgrade recorded against its domain.
- Clinical study reports were requested for every contributing trial. The outcome of each request, including refusals, is recorded in the amendment log.