Biomarkers, CAC scoring offer advantages to predict CV risk in patients with diabetes
Biomarkers to predict CV risk
Michael Davidson, MD, FACC, FNLA, professor and director of preventive cardiology at the University of Chicago Pritzker School of Medicine, said there are advantages to risk prediction with biomarkers vs. CAC, especially in patients in diabetes, whose lifetime risk for CVD is high.
“Having coronary calcium score itself provides important risk prediction, which is hard to beat as a risk predictor, but on the other hand, it doesn’t give any useful information that could help modify therapy to address residual risk,” Davidson said.
Biomarkers, including non-HDL, apolipoprotein B, LDL particle number and lipoprotein(a), have been determined to be causal for CVD, which allows improved risk prediction aside from LDL-C and the ability to target therapy to reduce the risk for major adverse CV events, according to the presentation. Inflammation is another key factor in CV risk in patients with diabetes.
There have been numerous studies that used CAC to assess therapeutic approach. The CHICAGO trial showed no difference in CAC score between glimepiride and pioglitazone during an 18-month to 2-year period. In the Periscope Companion trial, both treatment arms did not show a difference in CAC scores. In the BELLES trial, women were assigned high-dose atorvastatin or pravastatin, and at 2 years, CAC scores did not change in the two groups.
LDL-C has often been the target for risk prediction, but more information has been found on other lipoproteins.
“To me, it’s been a significant, historic injustice that LDL-C gets all the credit or the blame or the targeting, but triglyceride-rich lipoproteins are at least as atherogenic … as LDL,” Davidson said. “It’s been the unfortunate issue that is now becoming more understood and how we can better reduce this cholesterol to modify risk.”
Recent epidemiological trials have shown that patients with low HDL have high triglycerides, which results in a very high LDL particle number. In these patients, risk follows LDL particle number, not LDL-C, and LDL particle number is often discordant from LDL-C in patients with diabetes, according to the presentation. About 41% of patients with diabetes and LDL-C less than 100 mg/dL have elevated LDL particles.
“The triglyceride-rich cholesterol vs. LDL line up well in regards to the evidence for targeting treatment except for evidence from large outcome trials, approach,” Davidson said.
“We have not yet seen the data from FOURIER, whether the Lp(a) population had greater benefit with a PCSK9 inhibitor that would help us understand whether targeting that may affect outcomes,” Davidson said. “We’re kind of waiting for that type of information, although clearly [it’s a] very important predictor for risk.”
Based on this evidence, organizations such as the National Lipid Association and the
Inflammation has been changing the landscape with evolving new data. Patients with low LDL and high C-reactive protein that were assigned rosuvastatin in the JUPITER trial had a substantial benefit compared with placebo. Other studies, including IMPROVE-IT and FOURIER, showed that patients benefited from a nonstatin lowering of LDL on top of statin therapy, Davidson said.
Genomic risk scores have also been a target for recent studies.
“What’s also exciting is genomic risk scores that even the highest burden of genetic risk derived the largest relative and absolute clinical benefit with statin therapy,” Davidson said. “These risk scores are not quite there yet, but there’s certainly within the range of what high [CAC] score can predict for the population. So you have a diabetic patient who presents with a high risk score, really why wait for them to have [CAC] when you can now intervene even more effectively early, as you can try to prevent from even converting from a zero score to a positive [CAC] score with this approach.”
Although CAC scoring can be a strong risk predictor, oftentimes the risk is determined once it is too high.
“The real question for now: What can I do to affect a patient’s outcome? And LDL-C itself is just one part of the story,” Davidson said. “There [are] many other causal factors for atherosclerosis in patients. Potentially biomarkers can provide the greatest value to target the appropriate patients for residual risk reduction.”
Imaging and CAC scoring
In a separate presentation, Matthew J. Budoff, MD, FACC, FAHA, professor of medicine, program director and director of cardiac CT at Harbor-UCLA Medical Center in Torrance, California, discussed how imaging modalities to determine CAC scores can benefit patients with diabetes.
“What I want to show is not how to identify for 5 years whether someone’s going to have an event for a survival rate of 95% or 90%, but show which patients with diabetes are going to have an event with much more robust stratification,” Budoff said.
Patients with high risk for CV events may be treated with PCSK9 inhibitors and other high-priced therapies, whereas those with low risk can be treated with statins and other therapies that are often used in patients with diabetes, according to the presentation.
The 2010 ACC/American Heart Association guidelines made a IIa recommendation to screen asymptomatic adults aged at least 40 years with diabetes by measuring CAC.
The Framingham and United Kingdom Prospective Diabetes Study scores are different than CAC scoring, as CV events mainly occur in patients with intermediate risk.
“That’s a problem because if you’re going to say, ‘I’m going to treat all intermediate and high-risk patients,’ then you’re treating 70% to 80% of the population, in diabetes maybe even more than that, to try to eliminate the risk, but a calcium score can help us in a much more distinct way,” Budoff said.
CRP may be a new target of therapy, but it only predicts 60% to 65% of all events vs. 70% to 90% with CAC scoring, according to the presentation.
The Rotterdam Heart study showed that the “addition of CRP did not improve C-statistic or reclassification.” The MESA study had similar results in patients with type 2 diabetes. Adding other factors such as family history and ankle-brachial index did not affect specificity as much as CAC.
“Four studies comparing the predictive abilities of [high-sensitivity] CRP with CAC have demonstrated that CAC remains an independent predictor of [CV] events in multivariable models, whereas CRP no longer retains a significant association with incident CHD,” according to the presentation.
Although CRP is a strong target of therapy, Budoff said it will not identify who needs to be treated. Once a patient is identified, CRP or biomarkers may be used how exactly to treat those at high risk.
The number to treat in various studies has been a topic of interest. In MESA, half of the cohort had a CAC score of zero, but 549 patients had to be treated to prevent one event, but in the 25% of patients with a high CAC score, the number needed to treat was reduced to 24 patients.
Treating patients with aspirin is still under debate, as the number needed to treat is higher than the number to harm in patients with a score of zero, but a score of one lowers this difference, according to the presentation.
“The simple way that I use [CAC] is not only to decide who to get aggressive with, with lipid lowering, but also who to add 81 [mg] of aspirin to because I think that we can identify those with atherosclerosis, and there’s no doubt that aspirin is good in secondary prevention,” Budoff said.
Patients with diabetes oftentimes take many medications to manage their condition, and adding more medications to prevent CV events may lead to compliance issues, he said.
“There is no stronger motivator that we’ve been able to find than showing them that they have plaque in their coronaries, and therefore, best get aggressive with their own health,” Budoff said. “We’ve not only been able to show that a [CAC] score motivates things like statin adherence and weight loss, but it improves aspirin utilization. It improves better diet. It improves long-term exercise, so it can invoke a powerful message to the patient that they have a problem, even though they feel well, and therefore need to change what they’re doing.” – by Darlene Dobkowski
Budoff MJ. Imaging is better in predicting CV risk in diabetes.
Davidson MH. Biomarkers are better in predicting CV risk in diabetes. Both presented at: Heart in Diabetes Medical Conference; July 14-16, 2017; Philadelphia.
Disclosures: Budoff reports serving as a consultant for GE. Davidson reports consulting for Amgen, Merck, Regeneron and Sanofi.