From Clinical Trial Evidence to Clinical Guidelines: Perspectives on REWIND from Clinicians in the Gulf and Levant Region Open Access

Over a decade ago, the European Medicines Agency (EMA) and the US Food and Drug Administration (FDA) made it mandatory for all new glucose-lowering therapies to be the subject of a cardiovascular outcomes trial (CVOT) in patients with type 2 diabetes (T2D) [1]. This regulation arose due to concerns regarding the increased risk of adverse cardiovascular (CV) outcomes in association with some glucose-lowering therapies [1]. At that time, the idea that physicians could provide both glycaemic control and CV protection with one medication was novel. However, these mandatory CVOTs have demonstrated that some glucose-lowering therapies reduce CV risk [2] and have led to new indications for some drugs [1].

In the light of the scientific evidence from multiple CVOTs involving administration of glucagon-like peptide-1 receptor agonists (GLP-1 RAs), the American Diabetes Association (ADA) and ADA/European Association for the Study of Diabetes (EASD) guidelines currently recommend these agents as the first injectable therapy, in preference to insulin, in patients with T2D [3, 4]. A number of guidelines also advise that GLP-1 RAs (or SGLT2 inhibitors) with proven CV benefit should be considered for first-line therapy in patients with T2D who have or are at high risk of atherosclerotic cardiovascular disease (ASCVD), regardless of HbA1c [3‒8]. It should be noted that numerous clinical guidelines categorize patients with T2D who have not had a prior CV event as being at high CV risk, and those with T2D and ASCVD as being at very high or extreme CV risk, depending on their clinical history [6, 9].

One of the most important distinctions among the GLP-1 RA-based CVOTs, and one that has implications for real-world practice, is the spectrum of CV risk within the patient population at the time of randomization. Of the eight GLP-1 RA-based CVOTs published to date (ELIXA [lixisenatide], LEADER [liraglutide], SUSTAIN-6 and PIONEER-6 [semaglutide], EXSCEL [exenatide once weekly], Harmony Outcomes [albiglutide – no longer available for clinical use], REWIND [dulaglutide], and AMPLITUDE-O [efpeglenatide]) [10‒17], the proportion of patients with a history of cardiovascular disease (CVD) at baseline ranged from 31% (REWIND) to 100% (ELIXA and Harmony Outcomes) (Table 1). REWIND is the only GLP-1 RA-based CVOT in which <70% of the trial population had CVD at baseline (Table 1). Our aim was to use a blend of published data and author opinion to explore the importance and potential clinical implications of this finding, and of the REWIND trial results, for people with T2D in the Gulf and Levant region.

An understanding of the extent to which GLP-1 RA CVOT data relate to real-world populations is important for physicians in every jurisdiction. The eight published CVOTs involving GLP-1 RAs [10‒12, 14‒19] have notable and clinically relevant differences in baseline patient characteristics and design (Table 1). The most distinctive CVOT is REWIND. Compared with the other seven GLP-1 RA-based CVOTs, REWIND had a higher percentage of female participants (46% vs. 31–39%), a lower percentage of patients with established CVD (31% vs. 73–100%), a lower mean HbA1c at baseline (7.3 vs. 7.7–8.9%) and a longer follow-up duration (5.4 vs. 1.3–3.8 years) [16, 17, 21]. REWIND was also one of only two trials in which superiority of the active treatment versus placebo was specified as the sole primary objective [21]. The other CVOT that was designed to test prospectively for superiority was AMPLITUDE-O [17].

Compared with the other CVOTs, REWIND’s design features allow better assessment of the potential long-term effects of treatment on glycaemia, bodyweight, CV risk, and safety as well as – and most importantly – evaluation of both primary and secondary CV prevention in a population that resembles real-world practice more closely than the other trials [16, 22]. This latter point is supported by Boye et al. [23], who compared the characteristics of the patients enrolled in GLP-1 RA-based CVOTs with real-world patient populations in the USA. These authors concluded that 13%, 13%, 16%, and 43% of the US adult population with T2D would have been eligible for each of four GLP-1 RA-based CVOTs (LEADER, SUSTAIN-6, EXSCEL, and REWIND), respectively. This conclusion is echoed by the authors of this paper, all of whom estimate that only a minority of patients with T2D in their countries are represented by the populations of the first seven GLP-1 RA-based CVOTs (ELIXA, LEADER, SUSTAIN-6, PIONEER-6, EXSCEL, Harmony Outcomes, and AMPLITUDE-O) and who agree that the REWIND cohort most closely resembles their local T2D patient populations in baseline characteristics and CVD prevalence. Our informal estimates, drawn from real-world experience, of the prevalence of ASCVD in patients with T2D in our countries (Kuwait, ∼40%; Lebanon, 20–25%; Pakistan, ∼25%; Qatar, 20–30%; and UAE, 5–30%) contrast with CVD prevalences of 73–100% in the non-REWIND CVOT cohorts and align fairly closely with that of the global population of patients with T2D (32%) [24]. Importantly, disparities in CVD prevalence are often based less on region than on differences between primary and referral level care, with the upper end of the ranges quoted above generally relating to patients in secondary or tertiary care. These estimates are supported at present by a limited number of publications [25‒27]. For example, using data from the UAE, Al Hammadi et al. [25] and the Imperial College London Diabetes Centre [27] reported, respectively, that 18% and 10% of patients with diabetes had CVD, while Baagar et al. [26] reported coronary artery disease in 18% of Qatari patients with T2D. Therefore, there is a need for further accurate data.

The approach to management of T2D within the Gulf and Levant region varies by country and region. However, diabetologists generally follow the major guidelines from the USA (ADA and American Association of Clinical Endocrinologists/American College of Endocrinology) [4, 28] and Europe (ADA/EASD) [3, 5], as well as local guidelines, where available (e.g., those published by the Emirates Diabetes Society [29]). In contrast, cardiologists tend to utilize the American College of Cardiology/American Heart Association (ACC/AHA) or European Society of Cardiology/EASD guidelines [6‒8].

One organization – the ADA – has adopted the revolutionary approach of generating a “living” guideline [4], which is continuously updated online, allowing rapid incorporation and dissemination of RCT results to the diabetology community. A common theme of all current guidelines is that physicians should adopt a patient-centred approach to treatment and that, when choosing the pharmacological regimen, they should consider patient preference, as well as comorbidities, CV risk, bodyweight, the risk of side effects, including hypoglycaemia, and cost [3‒5, 29].

Given that ASCVD and heart failure are important causes of diabetes-related morbidity and mortality [4], reducing the risk of these conditions is essential for patient management in the Gulf, the Levant region and worldwide, and all current guidelines are clear that a patient’s CV history should have a major bearing on the specific glucose-lowering treatment implemented [3‒5, 29]. The presence of T2D automatically increases the risk of CVD, but multiple CV risk factors, target organ damage, or established ASCVD elevate the risk level to “very high” [29]. It is therefore important to identify and manage all ASCVD risk enhancers, including overweight/obesity, smoking, dyslipidaemia, renal disease, and hypertension (Fig. 1) [4, 30]. A number of risk calculators for heart failure are available for patients with diabetes [31]. Of these, the Building, Relating, Assessing, and Validating Outcomes (BRAVO) risk engine and Risk Equations for Complications Of type 2 Diabetes (RECODe) model are particularly useful as they allow simultaneous estimation of the risk of heart failure and other macrovascular and microvascular diabetes complications [31].

Risk calculators such as the ACC/AHA ASCVD risk calculator [32] can be programmed into electronic health records (e.g., Salama in the United Arab Emirates [UAE]), making them easy to use and automating the recording of results. However, existing American and European risk scores [32, 33] often greatly underestimate CV risk in the Gulf and Levant countries, mainly because many patients in this region develop T2D and CV risk factors at a younger age than is acknowledged in risk calculators developed in the west [29, 34]. For this reason, the Emirates Diabetes Society recommends screening for diabetes from 30 years of age [29]. However, the high prevalence of T2D in adolescents in some parts of the Gulf and Levant region suggests that screening should be instituted even earlier than this. The difficulties inherent in choosing any of the existing risk calculators for clinical decision-making with regard to primary CVD in the UAE are illustrated by the findings of a recent study which reported poor agreement (overall concordance coefficient, 0.50) among five externally validated risk assessment tools for CVD when applied to a large UAE dataset [35]. Although the patient population to which the UAE data relate differs from our patients in that they had no history of T2D or CVD, the results remain relevant. Generation of regional risk calculators for both CVD and heart failure is essential.

For patients with T2D in the Gulf and Levant region, the therapeutic regimen is typically based on lifestyle modification and management of all risk factors, with the aim of preventing a primary or secondary CV event and/or exacerbation of heart failure [3‒5, 29]. GLP-1 RAs with proven CV benefit are generally recommended for prevention of major adverse CVD events (MACE; primary or secondary), whereas patients with established heart failure and reduced ejection fraction typically receive a SGLT2 inhibitor with proven benefit to reduce the risk of worsening of heart failure, hospitalization for heart failure, and CV death [3‒5, 29]. In the context of this discussion, it should be noted that both the 2022 ADA/EASD guidelines, and the 2023 ADA guidelines, recommend the use of GLP-1 RAs in combination therapy in two main situations: (i) addition of a GLP-1 RA to insulin-containing regimens to improve the efficacy and durability of insulin [3, 4]; and (ii) in combination with an SGLT2 inhibitor (where both drugs have demonstrated CV benefit) in patients with ASCVD or multiple ASCVD risk factors because using these drugs in combination may additively reduce the risk of adverse CV and renal events [3, 4].

These therapeutic approaches reflect US and European guidelines [3‒5]. However, management of T2D may be influenced by cultural differences among regions. These differences affect local levels of some important CV risk factors, which are particularly high in the Gulf and Levant region. For example, in 2016, the prevalence of obesity in adults was >30% in the majority of countries in the Gulf and the Levant compared with 20–23% in western Europe [36]. This disparity is primarily related to the changes in culture, lifestyle, and diet that have accompanied economic development in the Gulf and Levant region [37, 38]. The prevalence of smoking is also high and rising [39]; in 2010–2011, the BREATHE study reported a 31% prevalence of cigarette and water pipe use in North Africa and the Middle East, with a high of 54% in Lebanon [40]. In addition, poor control of dyslipidaemia has been documented in Lebanon and Jordan [41, 42], Pakistan (second National Diabetes Survey of Pakistan) [43] and Dubai [44], and low levels of glycaemic control have been reported in Lebanon (IDMPS study [45]) and Dubai [45]. These findings may reflect, in part, suboptimal management of CV risk factors and use of guideline recommendations [45‒47].

Cultural differences between the west and some parts of the Gulf and Levant region may also confound treatment. For example, a diabetogenic environment promoted by myths, misconceptions, and cultural norms [37] is likely to counteract physician-led attempts to manage the patient’s disease. Many of the myths relate to the use of traditional therapies [37], most commonly herbal remedies [48]. These are widely used, often without consulting or informing the patient’s physician [48, 49]. In addition to the obvious safety concerns [48], it is our perception that the prevalence of myths and misconceptions may reduce adherence to all aspects of treatment, including lifestyle change, medication, laboratory tests, and follow-up physician visits. Also, unlike other jurisdictions, in the Gulf and Levant region patients can consult multiple physicians who may differ in their educational background, speciality, medical opinion, preferred guidelines, and perceived importance of comorbidities. These disparities may dilute key messages. In addition, consumption of medical misinformation may be more likely in a relatively young patient population that is more reliant on social media than their older counterparts. A recent systematic review found that, not only did social media users have a high probability of finding misleading videos relating to diabetes, but that such videos were more popular than those containing evidence-based health information [50].

In REWIND, the primary composite outcome (first occurrence of non-fatal myocardial infarction, non-fatal stroke, or death from CV causes [including unknown causes]) was observed in 594 (12.0%) participants in the group receiving dulaglutide (incidence rate: 2.4 per 100 person-years) and in 663 (13.4%) participants in the placebo group (2.7 per 100 person-years, with a hazard ratio [HR] of 0.88 and 95% confidence interval [CI] of 0.79–0.99; p = 0.026) [16]. This is similar to the HR of 0.86 (95% CI, 0.80–0.93) for MACE reported in a meta-analysis of data from the eight GLP-1 RA-based, placebo-controlled CVOTs [51]. However, it should be noted that the HR of 0.88 in REWIND was achieved in a population that was at relatively low CV risk compared with the other seven GLP-1 RA-based CVOTs (incidence of MACE in the placebo groups: 2.7 vs. 3.7–6.3 events per 100 patient-years; Table 2) [20]. REWIND’s recruitment of a patient cohort in which the majority of patients had CV risk factors but no prior CVD allowed meaningful exploration of the effects of dulaglutide in patients with and without prior CVD, and subsequent demonstration in a prior and post hoc analyses that dulaglutide significantly reduces the risk of both primary and secondary CV events in patients with T2D [16, 22]. In REWIND, the HR for MACE was similar in participants with previous CVD and those without (p_interaction = 0.97) [16]. Not surprisingly, given the difference in absolute risk of a CVD event between those with and without a prior CVD event, over the course of the study the number needed to treat (NNT) differed between the entire REWIND cohort (21% of whom had a prior event; NNT = 60 for entire cohort) and the subgroup of patients with a prior event at baseline (NNT = 18) [16]. For comparison, the 5‐year NNT for moderate intensity statin therapy in primary prevention is 75 for patients with a 10-year coronary heart disease risk of 7.5–9.9%, and 62 for those with a 10.0–19.9% risk [52].

The REWIND results appear to have had an unprecedented effect on practice guidelines in this therapeutic area. The 2019 update to the 2018 ADA/EASD guidelines for the management of hyperglycaemia in T2D incorporates the following recommendation: “We now also suggest that to reduce risk of MACE, GLP-1 receptor agonists can […] be considered in patients with T2D without established CVD with indicators of high risk, specifically, patients aged 55 years or older with coronary, carotid or lower extremity artery stenosis >50%, left ventricular hypertrophy, an estimated glomerular filtration rate <60 mL/min/1.73 m² or albuminuria” [5]. The recommendation that GLP-1 RAs should be considered for the primary prevention of CVD is accompanied by the statement: “To date, the level of evidence to support the use of GLP-1 receptor agonists for primary prevention is strongest for dulaglutide but lacking for other GLP-1 receptor agonists”. It is important to recognize that the ADA/EASD guidelines update further states that it is unclear whether the lack of evidence for a beneficial effect of other GLP-1 RAs on primary prevention of MACE outcomes in CVOTs is due to study details or lack of efficacy per se [5], as the numbers of primary prevention patients in these CVOTs were not sufficient to determine the presence or absence of a protective effect.

The use of GLP-1 RAs in the primary prevention of CV events in patients with T2D is also supported by the 2020 Emirates Diabetes Society guidelines, which recommend that a GLP-1 RA (or a SGLT2 inhibitor) with proven CV benefit should be used in the primary prevention of CV events in patients who are at very high CV risk, regardless of HbA1c [29]. Dulaglutide is the only antidiabetic agent for which the FDA has granted a primary CV prevention label indication [53].

Table 3 shows the HRs for MACE for the active treatment versus placebo groups in patients with or without CVD at baseline in the five CVOTs in which primary prevention of MACE could be determined (LEADER, SUSTAIN-6, PIONEER-6, EXSCEL, REWIND, and AMPLITUDE-O). The p values for interaction analyses based on CVD status at baseline (Table 3) show that a significant interaction was detected only in LEADER, in which liraglutide significantly reduced the risk of MACE in patients with CVD at baseline (HR, 0.83; 95% CI, 0.74–0.93) but not in those without (HR, 1.20; 95% CI, 0.86–1.67) [54]. However, when considering the other CVOTs – recognizing that caution is needed when making between-trial comparisons and acknowledging the absence of significant documented interactions – we know that substantially more data are needed to estimate interactions than main effects [61] and that these CVOTs were powered to estimate the effect of the intervention on the primary outcome, not interactions. With this in mind, the subgroup analyses shown in Table 3 suggest that other GLP-1 RAs may also have differing effects on MACE prevention, dependent on the patient’s risk profile. For example, both efpeglenatide (AMPLITUDE-O) and semaglutide (SUSTAIN-6) demonstrated similar patterns to liraglutide in LEADER (AMPLITUDE-O: CVD at baseline: HR, 0.71; 95% CI, 0.57–0.90; no CVD at baseline: HR, 1.71; 95% CI, 0.48–6.07; SUSTAIN-6: CVD at baseline: HR, 0.72; 95% CI, 0.55–0.93; no CVD at baseline: HR, 1.00; 95% CI, 0.41–2.46) [12, 17, 20]. In contrast, in the REWIND trial, the HR for MACE in patients receiving dulaglutide versus placebo was 0.87 (95% CI, 0.74–1.02) in patients with or without CVD at baseline [16].

The results of the GLP-1 RA-based CVOTs – and REWIND in particular – have provided robust scientific evidence to assist healthcare practitioners in making informed decisions in clinical practice. When presented with a patient with T2D who has multiple CV risk factors or established CVD, physicians now have the support of both clinical trial evidence and guideline recommendations when considering whether to initiate GLP-1 RA therapy. Differences in trial design/patient cohort between REWIND and the other GLP-1 RA-based CVOTs should inform the type of real-world patient who may best benefit from each drug. Dulaglutide is currently the only GLP-1 RA to demonstrate protection from primary CV events in patients with T2D when compared to placebo.

In addition to dulaglutide’s effects on glycaemic control and CV event prevention [16], other factors may be relevant to patient acceptance and, as a consequence, clinical outcome. These include the therapeutic regimen (once weekly [dulaglutide, exenatide (Bydureon), semaglutide (Ozempic)] versus once [liraglutide, lixisenatide, semaglutide (Rybelsus)] or twice [exenatide (Byetta)] daily) and route of administration (subcutaneous [dulaglutide, exenatide, liraglutide, lixisenatide, semaglutide (Ozempic)] versus oral (semaglutide [Rybelsus]) [55, 56, 60]. For those GLP-1 RAs that are injected, ease of use of the injection device may be an important factor in patient acceptance, and differences in patient preference have been documented. For example, in the PREFER study, in which asked patients were asked whether they preferred the dulaglutide device or the semaglutide device, there was a substantial and significant preference for the dulaglutide option (84 vs. 12%) [57]. Weight loss may also affect patient acceptance. This appears to be a class effect, with some variation among GLP-1 RAs in the extent of the weight reduction achieved [60].

The focus of current guidelines on prevention of primary and secondary CV events is clear [5, 29]. However, in our clinical experience, implementation of this cardiocentric approach to the management of T2D generally remains poor in the Gulf and Levant region, with many physicians (and patients) remaining focussed on a glucocentric approach. Although we are aware that there have been recent improvements in identifying and treating CV risk factors in the region [44], primary care physicians in particular need encouragement to apply current guideline recommendations [47, 58]. There is also a need for increased community awareness of CV risk factors, and education about the importance and effectiveness of early diagnosis and treatment of these risk factors to reduce the incidence of CV events.

It is important to recognize that GLP-1 RA therapy is frequently indicated in patients whose glycaemic control is good. For example, the 2020 update to the ADA/EASD guidelines recommends that GLP-1 RA (or SGLT2 inhibitor) therapy should be considered for patients who are at high CV risk or who have established ASCVD, heart failure, or chronic kidney disease (CKD), independent of their HbA1c level or glycaemic target [5]. This recommendation is supported by the Emirates Diabetes Society guidelines [29] and reflects the REWIND inclusion criteria (no lower limit for HbA1c at baseline) and the finding that dulaglutide’s efficacy in MACE prevention was equivalent in those with HbA1c ≥ 7.2% and <7.2% at baseline [5, 16]. Additional indications for the use of a GLP-1 RA include weight management (minimizing weight gain or promoting weight loss) and minimizing the risk of hypoglycaemia [5].

The most important conclusion that we can take from the results of the REWIND trial is that dulaglutide is the only GLP-1 RA with a CVOT that studied and demonstrated a CV benefit in patients without a history of CVD [17, 59]. REWIND’s results support physicians in the early use of dulaglutide in patients with near-target glycaemia and multiple CV risk factors but no prior CVD. Physicians should also note that REWIND provided proof of efficacy in a range of patient subgroups: the HR for MACE was unaffected not only by the patient’s prior history of CVD and baseline HbA1c level, but also by their age, sex, duration of diabetes, and body mass index at baseline [16]. However, regional and local socioeconomic, political and cultural issues will need to be overcome before patients in the Gulf and Levant region can realize the full benefit of a medication that can reduce the risk of primary CV events in patients with T2D.

REWIND’s results have provided new scientific evidence that has assisted in modifying pharmacological approaches to diabetology. However, many unanswered questions remain relating to the use of GLP-1 RAs in patients with T2D. For diabetologists in the Gulf and Levant region, the most compelling questions to explore in future CVOTs include the effects of GLP-1 RAs on the primary prevention of CV events in patients with T2D and CV risk factors that have not yet been examined in clinical trials (e.g., CKD), and on primary CV prevention in those with no CV risk factors other than T2D. The safety profile of GLP-1 RAs in patients with different baseline characteristics (e.g., those with CKD or peripheral vascular disease), and their effects on the incidence of retinopathy, should also be investigated, as should direct comparisons among different agents in this class. Such comparisons should include randomized clinical trials and real-world studies involving patient populations with and without CV disease at baseline, as well as those with stage 3 CKD. It would also be instructive if clinical trials included patients with baseline characteristics that more closely resemble those of the Gulf and Levant region. CVOTs should be supported by investigations into the effects of GLP-1 RAs on cardiac biomarkers and the mechanisms that underlie the CV protective effect of GLP-1 RAs.

The GLP-1 RA-based CVOTs conducted so far have led to a paradigm shift in T2D management, broadening the primary focus of treatment from control of hyperglycaemia to glycaemic control plus prevention of CV events. Their results have prompted the ADA/EASD and the Emirates Diabetes Society to recommend that GLP-1 RAs should be considered for both primary and secondary prevention of CVD in patients with T2D. It is likely that the results of REWIND – based as it was on extended follow-up of a population with a relatively low prevalence of prior CVD and powered for superiority over placebo – provided an evidence-based impetus for bringing about these changes. The views expressed in this paper have been brought together by scientific experts in order to explore the importance of REWIND’s findings for people with T2D who live in the Gulf and Levant region, and to highlight the potential implications of this trial on physicians’ treatment practices and patients’ lives. Dulaglutide represents a scientifically proven therapeutic option that, when compared with placebo, has been shown to control glycaemia, to potentially reduce bodyweight and to lower the risk of ASCVD – the leading cause of death in people with T2D – in patients with or without existing CVD. It therefore behooves practitioners who treat patients with T2D to become familiar with this drug and the evidence for its efficacy and safety and to determine where it may sit in their therapeutic armamentarium.

Talal Ashour, Sami Azar, Akram Echtay, Tarek Fiad, Ahmed Hassoun, Abdul Jabbar, Amin Jayyousi, and Hani Sabbour have no conflicts of interest to declare as authors participating in this publication. Muhammad Farooqi has been on the speaker’s bureau for Eli Lilly and Company CME programmes. Mohamed Hassanein has attended advisory board meetings and was a speaker for AstraZeneca, Boehringer Ingelheim, Eli Lilly and Company, Merck, MSD, Novartis, Novo Nordisk, Sanofi, and Servier. Kari Ranta, Fatih Tangi, and Ibrahim Turfanda are full-time employees and minor shareholders of Eli Lilly and Company.

Eli Lilly & Company Ltd provided financial support for the preparation of this article.

All named authors meet the International Committee of Medical Journal Editors (ICMJE) criteria for authorship for this article, take responsibility for the integrity of the work as a whole, and have given their approval for this version to be published. Kari Ranta and Ibrahim Turfanda contributed to the study conception and design. All authors contributed ideas, literature, and analysis throughout the preparation of the manuscript. All authors also commented on all versions of the manuscript and read and approved the final manuscript. Editorial assistance in the preparation of this article was provided by Janet Douglas and Karen Goa (Rx Communications, Mold, Wales). Support for this assistance was funded by Eli Lilly & Co., Dubai, United Arab Emirates. This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.