Data collection and study selection:
Our search retrieved 655 records from PubMed, Scopus, Web of Science, and Cochrane library. There were 75 duplicates. After title and abstract screening, we eliminated 557 records. Afterward, we screened 23 full texts for eligibility. Finally, nine records were included in our study: four published clinical trials and five registered protocols from Clinicaltrials.gov, and only eight studies were included in the meta-analysis (Figure. 1).
The total sample size for this meta-analysis was 1258 participants. 262 persons received a placebo and 886 participants received Cotadutide. There were no concomitant treatment modalities except in two studies. In NCT03235050, participants in all study groups received metformin tablets and a separate group was treated with liraglutide to compare it with Cotadutide and placebo. Moreover, during the treatment period of the study NCT03444584, participants were on metformin and dapagliflozin as well. Table 1 elucidates the full summary of the included studies. The baseline characteristics of the participants are illustrated in Table 2.
Quality assessment results:
The risk of bias summary is illustrated in figures 2 and 3. Regarding the Randomization process bias, all the studies were of low risk in terms of the randomization process except for NCT03550378, which was judged as some concerns because there was inadequate information about the allocation concealment, randomization, and baseline balance.
Regarding the intended interventions bias, most of the included trials had a low risk of bias in terms of deviations from the intended interventions except for NCT03645421 and NCT03745937, which were judged as some concerns. This is because there was no information about the statistical analysis used to estimate the effect of assignments in both of them despite blinding the personnel.
Regarding the missing outcome data bias, most of the included trials had a low risk of bias in terms of the missing outcome data due to applying the intention to treat analysis. We judged NCT03645421 and P.D. Ambery et al. as high risk of bias because the authors applied as-treated analysis (14).
Regarding the measurement outcome bias, we judged the risk of bias in the measurement of the outcome as low risk of bias in most of the studies due to blinding of all outcome assessors and using appropriate methods in measuring the outcomes. We judged NCT03645421 and NCT03550378 as some concerns due to the lack of information about blinding the outcome assessor.
For the selection of the reported results bias, the risk of bias due to the selection of the reported results ranged between low and some concerns. We judged all the registered protocols as some concerns because there is no published data yet to compare it with the protocols. The published studies (6,14–16) were of low risk as all outcomes mentioned in the results were present in the protocols.
For other sourced of bias, we judged almost all the studies as high risk in terms of other potential sources of bias as most of them are registered protocols without any published papers yet. Parker et al (15) stated the lack of statistical power to draw inferences between cohorts and the absence of validated questionnaires as a limitation in their study and so we judged it as having a high risk of bias. Accordingly, Ambery et al (6) had a relatively small population size which we considered as high-risk potential. Only (14,16) showed no other potential sources of bias.
Efficacy endpoints:
Percentage decrease in body weight: The pooled effect estimate of five studies favored Cotadutide 300 mcg over placebo (MD = 3.31, p > 0.00001). Pooled data were homogenous under a fixed effect model (p = 0.80, I² = 0%); Figure. 4.
Decrease in glycated hemoglobin (HbA1c): The pooled effect estimate of five studies showed that Cotadutide is significantly better than placebo (MD = 0.68, 95% CI [0.58, 0.79], p > 0.00001). Pooled data were homogenous under a fixed effect model (p = 0.05, I² = 55%); Figure. 5.
Percentage decrease in glucose area under the plasma concentration curve (AUC [0-4h]): The pooled effect estimate of six studies favored Cotadutide 300 mcg over placebo (MD = 30.15, p > 0.00001). Pooled data were heterogeneous under a random effect model and the heterogeneity was best resolved by leaving out NCT03596177 (p = 0.08, I² = 48%); Figure. 6.
Decrease from baseline in fasting plasma glucose over time (mg/dl): The pooled effect estimate of four studies favored Cotadutide over the placebo (MD = 31.31, p > 0.00001). Pooled data were heterogeneous under a random effect model and the heterogeneity was best resolved by leaving out NCT03645421 (p < 0.17, I² = 40%); Figure. 7.
Anti-drug antibodies (ADA): Nahra et al reported a statistically significant increase in the number of participants with ADA in the Cotadutide group over the placebo (155 out of 256 in the Cotadutide group and three out of 256 in the placebo group); NCT03444584, NCT03550378, and NCT03596177 reported non-significant results on the number of participants having ADA. NCT03444584 reported 2 out of 24 and 1 out of 24 in the Cotadutide group and the placebo group, respectively. In NCT03550378, two out of 21 participants in the Cotadutide group had ADA in comparison to the 20 persons on placebo in which none of them developed ADA. In NCT03596177, three out of 14 and zero out of seven participants experienced ADA in the Cotadutide group and the placebo group, respectively.
Safety endpoints:
Treatment-emergent adverse events (TEAEs): The pooled effect estimate of six studies showed a statistically significant increased risk of TEAEs in the Cotadutide group compared to placebo (RR = 1.40, p = 0.0007). Pooled data were homogenous under a fixed-effect model (p = 0.23, I² = 26%); Figure. 8.
Treatment-emergent serious adverse events (TESAEs): We didn’t do a meta-analysis for this outcome because none of the participants suffered any TESAEs in three out of six studies in both the Cotadutide and the placebo group. Only three studies reported some participants having TESAEs and they are relatively very low. In terms of TESAEs, NCT03550378 reported two out of 21 persons in the Cotadutide group and two out of 20 in the placebo group. On the other hand, NCT03596177 reported two persons out of 18 and zero out of seven in the Cotadutide group and the placebo group, respectively. In the MAD portion of the study by P. Ambery et al, they reported one out of seven participants having TESAEs in the Cotadutide 200 mcg group and none out of 19 participants in the placebo group (6).