Coronary Outcomes After Screening for Asymptomatic Coronary Artery Disease in Patients with Type 2 Diabetes: The DIAD Study: A Randomized Controlled Trial

JAMA 2009; 201 (15) 1547-1555.

Background

Approximately 200 million people around the world have DM with CAD being a major health concern. CAD is the leading cause of mortality among this population and the concern is that the presenting symptom may be MI or sudden cardiac death. DM has been recognized as a “CAD equivalent” and currently efforts have been directed at reducing cardiovascular risk factors to prevent CAD. There has been significant interest in the role of screening for asymptomatic CAD in this population in an effort to improve cardiovascular outcomes especially with the wide availability of noninvasive imaging studies.

The DIAD study was designed to assess whether routine screening for CAD in asymptomatic individuals identifies patients with type 2 diabetes as being at high cardiac risk and whether it affects their cardiac outcomes.

Methods : RCT

Dates / Enrollment July 2000 and August 2002, follow up through September 2007
Participants
Figure 1 / 1123 patients at 14 US Centers randomly assigned to screening and control groups
--randomization through permuted block design
--inclusion criteria: 50 to 75y at enrollement, DM dx >30yo, no h/o ketoacidosis
--exclusion criteria: Angina pectoris or chest discomfort, stress test or coronary angiography within last 3 years, h/o MI, heart failure or coronoary revascularization, abnormal rest EKG (pathologic Q waves, >1mm ST depression, deep negative T waves, complete LBBB), any clinical indication for stress test, active bronchospasm which precluded use of adenosine, limited life expectancy due to cancer/ESRD/ESLD.
Screening Group / 561 randomized to stress MPI imaging with adenosine Tc-99m sestamibi
--Cardiologists at each site read stress MPI images and communicated result with participants and physicians
--Studies were also interpreted by independent expert panel
--Nonperfusion abnormaltieis also assessed: ischemic EKG changes, transient LV dilation, baseline LV dysfunction
--Further therapy/intervention per patient’s physician
Control Group / 562 randomized to receive usual care
--Includes clinically indicated cardiac perfusion studies
1o Endpoints / Nonfatal MI
Cardiac Death
2o Endpoints / Unstable Angina
Heart Failure
Stroke
Coronary Revascularization
Determination of Outcome / Questionnaires q6mo to participants including health status, medications, cardiac events, additional stress imaging, coronary angiography and revascularization
Analysis / Intention to Screen comparison of cardiovascular death or nonfatal MI
- Study designed with power 80% based on expected 5-10% event rate over 5y


Results

Baseline Characteristics: Similar

Follow up Mean 4.6y

Results found in Table 2, also shown in Figure 2a

No Screening / Screening
Primary Events
--Non fatal MI
--SCD / 17/562 = 0.0302 = 3.0%
--10/562 = 0.0178 = 1.8%
--7/562 = 0.0125 = 1.3% / 15/561 = 0.0267 = 2.7%
--7/561 = 0.0125 = 1.3%
--8/561 = 0.0143 = 1.4%
Secondary Events
--Unstable Angina
--Heart Failure
--Stroke / 14/562 = 0.0249 = 2.5%
--3/562 = 0.005 = 0.5%
--7/562 = 0.0125 = 1.3%
--5/562 = 0.009 = 0.9% / 21/561 = 0.0374 = 3.74%
--4/561 = 0.007 = 0.7%
--7/561 = 0.0124 = 1.2 %
--10/561 = 0.0178 = 1.8%
All Cause Mortality / 15/562 = 2.7% / 18/561 = 3.2%

All results shown are not statistically significant.

--Revascularizations including PCTA, CABG did not differ between the 2 groups (5.5 vs 7.8%)

Table 3: Events according to Findings on Imaging.

--78% had normal imaging, 10% small perfusion defect and 6% had moderate or large defect

--In the moderate to large defect group: 7 of 33 had revascularization (21%) – 1 pt had both

--Showed that moderate to large perfusion defect had 12% event rate (4 of 33).

Table 4: Follow up/Studies and Medication use

--Overall similar number of angiography in both groups, though more in screening group at <120d (13%)

--Screening group had 21% nonprotocol stress vs 30% no screening (670 vs 170 studies)

--Pharmocologic management did not differ between the 2 groups

Table 5: Post hoc analysis for factors assoc with primary events

Validity

Are the studies valid?

·  Was the assignment of patients to screening randomised? Yes

·  Was the randomization list concealed? Yes

·  Was follow-up of patients sufficiently long and complete? Yes

·  Were all patients analysed in the groups to which they were randomised? Yes

·  Were patients and clinicians kept “blind” to screening? No

·  Were the groups treated equally, apart from the experimental treatment? Yes, appears so though this was determined by their physician

·  Were the groups similar at the start of the trial? Yes

Discussion

·  This was largely a negative study which did not show any difference in outcomes in screening for asymptomatic cardiovascular disease among diabetics.

·  It was underpowered to detect differences at the observed event rate

·  Repeat imaging reported in prior publication

·  Are there any important conclusions to be drawn from this study? Yes

·  This study brings up an important point –lower risk of significant CV disease among diabetics than previously thought as shown by this study and others:

·  DIAD: 10 year risk estimates: 5.4 – 6.0%

·  ACCORD (Effects of IntensiveGlucose Lowering in Type 2 Diabetes)(NEJM 2008; 358 (24) 2545-2560.) 10,251 patients with HbA1c 8.1.

Standard Group CV Death and MI: 1.91% per year = 19.1% 10 year risk

Intensive Group CV Death and MI: 2.01% per year = 20.1% 10 year risk

·  ADVANCE (Intensive Blood Glucose Control and Vascular Outcomes in Patients with Type 2 Diabetes) (NEJM 2008.) 11,140 patients with HbA1c 6.5-7.3.

Intensive group CV Death and MI: 7.2% over 5 years = 14.4% 10 year risk

Standard Group CV Death and MI in standard group 8.0% = 16.0% 10 year risk

Reevaluating our interventions:

Recently published in JAMA 2008; 300 (18). 2134 - 2141

·  Low Dose Aspirin for Primary Prevention of Atherosclerotic Events in Patients with Type 2 Diabetes

16 center, 2539 patients, 4.4y f/u. Randomized to ASA or no ASA. No difference between 2 groups in atherosclerotic events: ASA group 5.4% vs no ASA 6.7% (not statistically different)