The effect of preventative cardiovascular therapies on coronary artery disease in people with and without type 2 diabetes: a propensity-matched score study

Patients with diabetes are known to be at an increased risk for developing coronary artery disease (CAD) compared to patients with normoglycaemia. A number of studies over the past decades have shown that the management of cardiometabolic risk factors including hyperglycaemia, dyslipidemia, and hypertension through intensive multifactorial interventions reduce the development and progression of diabetes related macrovascular complications [1‐5]. Despite this, cardiovascular (CV) disease continues to be the leading cause of death for adult patients with type 2 diabetes mellitus (T2DM) and represents a substantial economic burden to both the patient and population [6, 7]. Whilst there have been a large number of clinical studies documenting the relatively poorer CV outcomes for patients with diabetes, there have been few studies that have examined the impact of diabetes on coronary artery anatomy. In particular, the extent of CAD involvement for patients with and without diabetes and the impact of traditional CV disease preventative therapies on disease within the coronary arteries, as assessed by angiography, has not been well characterised over the past 10 years. We have therefore evaluated the severity of CAD based on coronary angiogram for patients with and without T2DM admitted to a tertiary hospital in Australia over the seven-year period between 2013 and 2019. We hypothesis that people with diabetes will have a greater burden of atherosclerosis within their coronary arteries compared to those without diabetes, even following adjustment for the use of traditional CV protective medications.

Baseline patient characteristics of the study population are shown in Additional file 1: Table S1. Of first angiograms 1572 (27.2%) had T2DM. Patients with T2DM were significantly older, less likely to be born in Australia, less likely to speak English and had greater numbers of co-morbid conditions than those without T2DM. There were no significant differences in the indication for coronary angiography in patients with and without T2DM: chest pain, 22% versus 22%; history of Acute Coronary Event (ACS), 9% versus 10%; breathlessness, 8% versus 9%; positive non-invasive test, 36% versus 33% and other, 26% versus 27%, respectively.

We found that T2DM patients have a greater burden of CAD as assessed by angiography compared to those without diabetes. Interestingly we also found that T2DM patients had more extensive CAD as assessed by angiograms performed in 2018 compared to 2013, but that possibly this trend was attenuated in 2019. The above findings need to be put into the context of a possible change in the indication or thresholds for performing coronary angiograms in people with T2DM over the observational period of the study. However, possibly the most important finding from our study is that despite a greater use of CV protective medications by patients with T2DM, the extent of disease within the coronary arteries, as assessed by angiography, was still greater in patients with T2DM compared to those without T2DM. However, after adjustment for medication use, we found that there was no longer any difference in the extent of CAD between patients with and without T2DM. These findings highlight the exaggerated risk that patients with T2DM are still at for developing CAD. However, they also suggest that with the greater application of traditional CV disease preventative therapies, it may be possible to reduce the extent of the CAD burden in patients with diabetes to that seen in patients without T2DM.

Of the 367 patients with T2DM and medication details available, 78% (n = 287) had evidence of angiographic disease. This rate was less in patients without T2DM, of which 70% (n = 276) had evidence of angiographic CAD. In our unadjusted ordinal logistic regression model, prior to accounting for the use of preventative medications, patients with T2DM were significantly more likely (by 32%) to have more extensive CAD than patients without T2DM (OR: 1.32 95% CI: 1.01, 1.74). Following adjustment for the use of statins, RAS inhibitors and anti-platelet drugs, this significant difference in extent of CAD no longer persisted between patients with and without T2DM (OR: 1.14 95% CI: 0.85, 1.53).

When comparing the use of traditional CV disease preventative therapies between patients with and without T2DM; statins, RAS inhibitors and anti-platelet drugs were significantly more likely to be taken by patients with T2DM. The increase in medication use for patients with T2DM compared to those without T2DM was 23% for statins, 14% for RAS inhibiting agents and 7% for anti-platelet agents. All of these differences for the uptake of CV preventative medication in patients with T2DM compared to those without T2DM were statistically significant. Whilst these results are encouraging, our results also emphasise that more aggressive uptake and use of CV protective medications, possibly earlier after a diagnosis of T2DM is made, is still required to reduce the gap between the heightened risk for more severe CAD in patients with T2DM compared to those without T2DM. Unfortunately, we lacked information on achieved lipid or blood pressure measurements to gauge whether targeting the above medications to achieve certain biochemical or clinical parameters influenced the results we report.

It is well established that risk factor modification through the optimisation of glycated haemoglobin levels and blood pressure, treatment of microalbuminuria, dietary intervention, exercise and smoking cessation are able to reduce the risk in the development and progression of vascular complications [4, 12]. Our findings appear to be broadly consistent with those of a recent large epidemiological study which has also suggested that it is possible for patients with T2DM to eliminate their excess risk for acute myocardial infarction compared to the general population if they can achieve five risk-factor variables (glycated haemoglobin, low-density lipoprotein cholesterol, systolic blood pressure, albuminuria and smoking status) within target ranges/recommendations [5].

A number of trials, both pharmacological and epidemiological, have suggested that the use of statins and blood pressure (BP) lowering medications are most likely responsible for the change in risk for CV events that had been observed in recent times [13‐17]. The publication of these large influential studies in the 1990’s and early 2000’s and the further confirmation of the CV benefits of the above approach in later studies has led to a drastic increase in the uptake of statins and anti-hypertensive agents, in particular RAS inhibitors, for both high risk vascular patients with and without T2DM [18‐20]. However, over time, some concern about side-effects has been reported regarding the use of statins [21]. If patients had been ceasing statins due to this concern, we would have expected to see the effect equally across both patient groups with and without diabetes. Given the reported side effects associated with statins, the wider use of newer approaches to treat dyslipidaemia such as the PCSK-9 inhibitors and possibly, ethyl icosapent, may result in a further decrease in the CAD burden in both patient groups with and without diabetes [22].

Targets for optimising BP lowering therapy in patients with and without diabetes still remain to be optimally defined and indeed, there has been a relaxing of BP targets for patients with diabetes over the last 10 years [23]. However, RAS inhibiting agents are often prescribed for the cardio-renal effects in addition to their use as BP lowering agents. Unfortunately, the design of our study did not allow us to follow any temporal trends in number of patients with and without T2DM taking these agents. Furthermore, the value of aspirin therapy as a primary CV preventative agent, when offset by potential adverse events, has also been called into question in recent times. Whether this issue has differentially affected the use of aspirin therapy in patients with and without diabetes remains to be examined.

Our analysis of patients admitted for their first coronary angiogram at a single tertiary hospital shows that the extent of CAD being detected on coronary angiogram was greater over time in patients with T2DM. This was in contrast to patients without T2DM where no significant effect on time on CAD extent was demonstrated. These findings appear to be contrary to reports of reduced rates for myocardial infarctions over recent years for patients with and without diabetes, with relative rate reductions appearing to be being larger in patients with T2DM [24]. It is therefore important to emphasise that the results we report are not based on clinical events or outcomes but rather on findings from a coronary angiogram. It is possible that the initial increase in severity of CAD burden seen in our study for patients with T2DM may represent improved CAD screening over recent years rather than suboptimal application of preventative therapies to slow disease progression within the coronary arteries. A lower threshold for performing a coronary angiogram may also exist for patients with T2DM given the appreciated higher risk of CAD a diagnosis of T2DM confers and a higher appreciation of asymptomatic CAD in patients with T2DM [25]. Possibly, wider use of medications such as SGLT-2 inhibitors and GLP-1 receptor agonists which have been shown to have CV protective effects will result in improvements in CAD burden following 2019 for patients with diabetes, however a longer duration of follow-up will be required to confirm whether the attenuation of severity in CAD is a true finding [26‐29].

There were a number of limitations with this study. As we utilised hospital discharge coding data, information was not available on metabolic and blood pressure control or duration of diabetes. Therefore, we are unable to account for the effect that these preventative medications may have had on patient’s risk profile. This is a single-site study, reporting on the extent of CAD in patients attending a public tertiary referral and university teaching hospital with a state-wide catchment. However, the outcome of extent of CAD may be influenced by individual differences in hospital characteristics. There may also be a small number of patients that may have incorrectly been classified as either having had or not had T2DM [30]. However, the prevalence of diabetes that we found in our study of 28% is consistent with other local studies [31]. Another limitation to consider is that this study relied on individual cardiologists reporting on the indication for angiogram and the extent of CAD without a prospective set of criteria for grading the extent of CAD; however, this is balanced by the large number of experienced cardiologists reporting results. As patients with and without T2DM were assessed in an identical fashion, it is unlikely that the extent of CAD would have been recorded disproportionately for the two groups of patients. We report on the first recorded angiogram at our hospital, it is therefore possible that a patient may have received an angiogram or coronary invasive procedure previously at a different hospital. In this case, these patients may be considered to be high risk to warrant a second angiogram and this may occur more frequently in patients with T2DM. If this was the case, these patients would almost certainly be classified as high risk vascular patients and hence been more likely to be further investigated with another angiogram at our hospital. It is presumed that the above scenario occurred more frequently in patients with diabetes given the known exaggerated risk for CAD that this group is at.

These limitations are counterbalanced by the strengths of the study which include a large number of coronary angiograms from both patients with and without diabetes over a seven year period. As it was an observational study, propensity scores were used in the regressions in order to include as many coronary angiogram results from patients with diabetes as possible allowing for a representative analysis.

Our findings show that the extent of disease within the coronary arteries of patients with T2DM still remains more severe compared to those without T2DM. However, it is possible that this excess burden of disease can be attenuated to the levels seen in patients without T2DM by the more aggressive uptake of traditional CV protective medications. These results support those of a recent large epidemiological study that suggested that aggressive risk factor modification in patients with T2DM can eliminate the excess risk for myocardial infarction compared with that of the general population [5].