Predicting the Risk of CVD in Diabetes Patients

Data analysis suggests the need to create CVD prediction models specifically focused on the risk factors that affect patients with type 2 diabetes. 

Cardiovascular disease is one of the major causes of increased morbidity and mortality in today’s population. Patients who have diabetes are at increased risk of developing cardiovascular diseases. Cardiovascular diseases (CVD) is a broad term that includes several heart conditions such as ischemic heart disease, stroke, and vascular diseases. Blood vessels of different sizes are adversely affected by diabetes, leading to conditions such as atherosclerosis, which affects the coronary artery, leads to ischemia, and results in heart attack, stroke, and several other comorbidities. 

 A study conducted in the United Kingdom (UK) sought to compare the performance of risk prediction scores for cardiovascular diseases such as coronary heart disease and stroke. The study is based on 22 cardiovascular risk prediction scores pulled from UK literature review data of a total of 168,871 people living with type 2 diabetes, irrespective of their cardiovascular outcome. The trials eligibility includes age 18 years or greater without any cardiovascular disease outcomes. Patients were divided into two categories:16,887 used as an independent training sample size, and 151,984 used as a comparison model. Statistical analysis was conducted using Harrell’s C statistic, and calibration was beneficial for evaluating the models based on discrimination. CVD history, sex, age, and statin use at the time of diagnosis all played a role in subgroup performance. The discriminative capacity of the two models was formally tested, using the test data by comparing the differences in C statistics. The result showed that nine out of the 22 identified cardiovascular risk prediction models were derived from type 2 diabetes, including non-diabetes patients and type 2 diabetes. The thirteen scores were used to create general population samples. CVD was predicted by ten rules, cardiovascular heart disease (CHD) by seven, stroke by three, and heart failure by two. All risk ratings considered traditional CVD risk factors like age, gender, blood pressure, and smoking status, including the 20 risk rating in the lipid information. In addition, the presence or absence of type 2 diabetes was frequently included as a predictor in the scores that included a proportion of people with diabetes. Still, risk factors that are particular to diabetes patients, such as diabetes duration and glycemic status, were not included.  

According to the study findings, the ten years median follow-up of type 2 diabetes patients from the time of their diagnoses showed a total of 22.7% of patients who suffered from cardiovascular disease, atrial fibrillation, or heart failure event. About 17.19% of the total had events of cardiovascular diseases, 12.22% had congestive heart failure, 8.19% had atrial fibrillation, 5.6% had heart failure, and 3.98% had a stroke. Furthermore, medical history and prescription data could be easily obtained before diagnosing type 2 diabetes. Measurable risk factors such as blood pressure were collected utilizing a 12-month window before and one week following diagnosis. Validation studies can be carried out in primary care records, where such risk cores are used in practice. The authors recognize that such data may have missing data and coding mistakes. Still, the more artificial circumstances of a research cohort study or clinical trial reflect their real value better. Furthermore, real-world clinical data allows for an entire population approach, whereas cohorts and trials have entrance constraints. 

Many cardiovascular prediction tools continue to evolve as the risk of developing CVD increases for patients with type 2 diabetes. However, several tools have not been validated for the type 2 diabetes patient population. This study found that CVD risk scores obtained from the general population performed worse in type 2 diabetes patients. The CVD risk scores produced from people with diabetes did not perform better in general, making it difficult to predict the effectiveness of the scores. The scores performed similarly for a broader definition of CVD, including HF and AF, both of which are more common in people with diabetes. 

 Practice Pearls: 

• Patients with diabetes are at increased risk of developing cardiovascular diseases, and general cardiovascular risk prediction assessment performs below this population’s expectations. 
• Predictive performance was improved when considering diabetes patients with cardiovascular diseases such as atrial fibrillation and heart failure. 
• The study focuses on the need to create a CVD prediction model that considers the risk factors and outcomes relevant for patients who suffer from diabetes. 

Dziopa K, Asselbergs FW, Gratton J, Chaturvedi N, Schmidt AF. Cardiovascular risk prediction in type 2 diabetes: A comparison of 22 risk scores in primary care settings – Diabetologia. SpringerLink. https://link.springer.com/article/10.1007/s00125-021-05640-y. Published January 15, 2022.  

Betsy Adegborioye, PharmD Candidate 2022, South College School of Pharmacy.