Impact of statin use on major adverse cardiovascular events in patients with normal stress myocardial perfusion imaging and elevated coronary artery calcium score

Abstract

Introduction:

Normal stress myocardial perfusion imaging (MPI) carries a favourable prognosis. Conversely, elevated coronary artery calcium (CAC) is associated with increased major adverse cardiovascular events (MACE). There is limited information on the prognosis and management of patients with elevated CAC and normal MPI. We aimed to assess the outcomes of patients with elevated CAC and normal MPI in relation to post-MPI statin use.

Methods:

A retrospective review of normal MPI with CAC score >300 was performed between 1 March 2016 and 31 January 2017 in a Singapore tertiary hospital. Patients with known atherosclerotic cardiovascular disease or left ventricular ejection fraction <50% on MPI were excluded. Patient demographics, prescriptions and MACE (cardiac death, nonfatal myocardial infarction and/or ischaemic stroke) at 24 months after MPI were traced using electronic records. Binary logistic regression was used to evaluate for independent predictors of MACE.

Results:

We included 311 patients (median age 71 years, 56.3% male), of whom 65.0% were on moderate to high-intensity statins (MHIS) after MPI. MACE was significantly lower in the post-MPI MHIS group (3.5% vs. 9.2%, P = 0.035). On univariate binary logistic regression, post-MPI MHIS use was the only significant predictor for MACE (odds ratio [OR] 0.355 [95% confidence interval (CI) 0.131–0.962], P = 0.042), even after multivariate adjustment (adjusted OR 0.363, 95% confidence interval 0.134–0.984, P = 0.046).

Conclusion:

Post-MPI MHIS use is associated with lower MACE and is an independent negative predictor for 24-month MACE among patients with normal MPI and CAC >300.

INTRODUCTION

Stress single-photon-emission computed tomography (SPECT) myocardial perfusion imaging (MPI) is useful for evaluation of suspected coronary artery disease (CAD).[1] Normal stress SPECT MPI confers excellent prognosis with low rates of major adverse cardiovascular events (MACE) within a 'warranty period' of approximately two years.[2] Similarly, coronary artery calcium (CAC) scoring has also been shown to correlate with MACE.[3] Guidelines recommend using CAC to guide decisions for statin use[4] and are supported by recent studies showing improved outcomes when statins are initiated in patients with elevated CAC.[5] Concurrent CAC scoring with SPECT MPI has been advocated for various reasons, including improved interpretation of borderline SPECT MPI abnormalities.[6] However, the optimal management of patients with elevated CAC and normal stress SPECT MPI is unknown. We aimed to assess the outcomes of patients with elevated CAC and normal stress SPECT MPI in relation to statin use.

METHODS

We retrospectively studied all patients with normal stress-rest SPECT MPI performed for evaluation of suspected CAD from 1 March 2016 to 31 January 2017 in a Singapore tertiary hospital. These patients also had a concurrently measured total Agatston CAC score >300. Cases with normal stress-only MPI were also included. Patients with known atherosclerotic cardiovascular disease were excluded. We also excluded patients with left ventricular ejection fraction (LVEF) of <50% on gated SPECT MPI. Patient demographics, prescriptions and clinical progress were traced from electronic medical records. We defined moderate to high-intensity statins (MHIS) as use of at least rosuvastatin 5 mg, atorvastatin 10 mg, simvastatin 20 mg, lovastatin 40 mg or pravastatin 40 mg, in accordance with the American College of Cardiology/American Heart Association Task Force guidelines.[7] The medical records were also reviewed for any occurrence of MACE (cardiac death, non-fatal myocardial infarction and/or ischaemic stroke) within 24 months of the MPI. The study was approved by the hospital's institutional review board.

MPI was performed using technetium (Tc99m) tetrofosmin and a cadmium zinc telluride–based camera (Discovery NM530c, GE Healthcare, Haifa, Israel). All patients underwent stress testing with either exercise on the Bruce protocol, intravenous dipyridamole or intravenous dobutamine infusion. This may be done as part of a one- or two-day rest/stress protocol. In a one-day protocol, 8 mCi of Tc99m tetrofosmin was used for rest and 24 mCi for post-stress imaging. For logistical reasons, a two-day study using 20 mCi for each imaging was done in some patients. All patients with body weight >80 kg also underwent only two-day studies with 25–30 mCi of Tc99m tetrofosmin used for each imaging, depending on actual body weight. Pretest preparation, cardiac stress, image acquisition and processing were performed in accordance with standard published protocols.[8,9] Because attenuation correction was not available on this gamma camera system, post-stress prone imaging was routinely performed after supine imaging in patients without physical limitations to prone positioning to assist in identification of diaphragmatic attenuation artefacts. MPI data was interpreted by consultant cardiologists with advanced nuclear cardiology training and accreditation (e.g. Certification Board in Nuclear Cardiology) according to accepted guidelines.[8]

Included patients underwent a non-enhanced computed tomography (CT) study using a 256-slice dual-source CT scanner (Siemens Somatom Definition Flash, Siemens healthcare; Erlangen, Germany) on the stress MPI day. The images were acquired and Agatston CAC scoring performed according to established protocols.[10,11]

Normally distributed continuous variables were presented as mean ± standard deviation and were tested using t-test, whereas skewed data was presented as median with interquartile range (IQR) and compared using the Wilcoxon rank-sum test. Categorical variables were presented as frequencies and percentages and tested using the Chi-square test. Logistic regression was performed to evaluate for independent variables predicting MACE at 24 months. Multivariate analysis using the backward stepwise likelihood ratio method was performed for all variables identified with P < 0.1 on univariate analysis with adjustment for age and gender. All statistical analyses were performed using IBM SPSS Statistics version 16.0 (IBM Corp, Armonk, NY, USA); significance tests were two-sided at the 5% significance level.

RESULTS

There were 1,209 patients with normal stress MPI during the study period, of whom 326 had CAC >300. After the exclusion of patients with LVEF <50%, 311 patients were included for analysis, of whom 202 (65.0%) were on MHIS after MPI. The baseline characteristics of the 311 cases are presented in Table 1, stratified by post-MPI statin status. The overall median age was 71 years, and 56.3% were male. Cardiovascular risk factors such as hypertension, diabetes mellitus and smoking were present in 79.1%, 50.8% and 16.4%, respectively. In terms of baseline medications before MPI, 52.4% were on MHIS, and 38.3% were on at least one antiplatelet agent. Pre-MPI MHIS as well as pre- and post-MHIS antiplatelet use was significantly higher among patients on MHIS after MPI than among those not on MHIS after MPI (80.7% vs. 0%, 46.5% vs. 22.9%, 66.3% vs. 34.9%, respectively; P < 0.001 for all comparisons).

Table 1.

Baseline characteristics and outcomes of patients stratified by post-MPI statin status.

Characteristic	No. (%)/Mean±standard deviation			P
	All (n=311)	Non-MHIS (n=109)	MHIS (n=202)
	Age (yr)	71±8	71±9		71±8	0.724
Male gender	175 (56.3)	57 (52.3)	118 (58.4)	0.299
Diabetes mellitus	158 (50.8)	50 (45.9)	108 (53.5)	0.201
Hypertension	246 (79.1)	89 (81.7)	157 (77.7)	0.416
Smoking history	51 (16.4)	12 (11.0)	39 (19.3)	0.059
Baseline MHIS use	163 (52.4)	0 (0)	163 (80.7)	<0.001
Antiplatelet use
At baseline	119 (38.3)	25 (22.9)	94 (46.5)	<0.001
After MPI	172 (55.3)	38 (34.9)	134 (66.3)	<0.001
MPI protocol				0.874
Stress only	13 (4.2)	5 (4.6)	8 (4.0)
1 day stress/rest	213 (68.5)	76 (69.7)	137 (67.8)
2 days stress/rest	85 (27.3)	28 (25.7)	57 (28.2)
MPI stress agent				0.213
Exercise	40 (12.9)	16 (14.7)	24 (11.9)
Dipyridamole	257 (82.6)	91 (83.5)	166 (82.2)
Dobutamine	14 (4.5)	2 (1.8)	12 (5.9)
Agatston score*	685.0 (461.3-1,244.0)	681.0 (434.0-1,284.0)	690.5 (476.1-1,242.5)	0.104
Gated LVEF	69±9	69±9	69±9	0.652
Any MACE	17 (5.5)	10 (9.2)	7 (3.5)	0.035
Cardiac death	8 (2.6)	5 (4.6)	3 (1.5)	0.134
NFMI	4 (1.3)	2 (1.8)	2 (1.0)	0.614
TIA/CVA	5 (1.6)	3 (2.8)	2 (1.0)	0.348

*Data presented as median (interquartile range). LVEF: left ventricular ejection fraction, MACE: major adverse cardiovascular events, MHIS: moderate-high intensity statin, MPI: myocardial perfusion imaging, NFMI: nonfatal myocardial infarction, TIA/CVA: transient ischaemic attack/cerebrovascular accident

The majority (82.6%) of patients underwent dipyridamole stress testing. The median Agatston score was 685.00 and mean gated SPECT LVEF was 69%. The median CAC was not significantly different in patients with or without MHIS before MPI (CAC with MHIS: 691.0 [IQR 474.8–1,294.0) vs. CAC without MHIS: 681.2 [IQR 458.4–1,177.0), P = 0.788). The MPI stress modality, CAC and LVEF were also not significantly different between patients with or without post-MPI MHIS.

MACE within 24 months of MPI was observed in 17 (5.5%) patients. There was significantly lower MACE among patients on MHIS after MPI (3.5% vs. 9.2%, P = 0.035). On univariate binary logistic regression, post-MPI MHIS use was the only predictor for MACE occurrence (odds ratio [OR] 0.355 [95% confidence interval (CI) 0.131–0.962], P = 0.042) and remained significant after adjustment for age, gender and pre-MPI MHIS use (adjusted OR 0.363 [95% CI 0.134–0.984], P = 0.046) [Table 2].

Table 2.

Binary logistic regression for predictors of MACE after MPI.

Variable	Univariate		Multivariate
	OR (95% CI)	P	OR (95% CI)	P
Age	0.998 (0.940-1.060)	0.961	-	-
Gender		0.433	-	-
Male	Ref
Female	0.048 (0.555-3.941)
Diabetes mellitus	1.409 (0.522-3.802)	0.498	-	-
Hypertension	2.045 (0.456-9.181)	0.350	-	-
Smoking history	1.098 (0.304-3.969)	0.886	-	-
Baseline statins		0.060	-	-
Non-MHIS	Ref
MHIS	0.359 (0.123-1.044)
Baseline antiplatelet	0.658 (0.226-1.917)	0.443	-	-
MPI protocol			-	-
1 day stress/rest	Ref
2 days stress/rest	1.047 (0.357-3.067)	0.933
Stress only	0.000 (0.000)	0.999
MPI stress agent			-	-
Exercise	Ref
Dipyridamole	1.095 (0.239-5.008)	0.907
Dobutamine	1.462 (0.122-17.482)	0.764
Agatston score	1.000 (1.000-1.001)	0.163	-	-
Gated LVEF	0.979 (0.927-1.035)	0.460	-	-
Post-MPI statin		0.042		0.046
Non-MHIS	Ref		Ref
MHIS	0.355 (0.131-0.962)		0.363 (0.134-0.984)*
Post-MPI antiplatelets	0.547 (0.203-1.477)	0.234	-	-

*Hosmer Lemeshow goodness-of-fit test (P=0.297), classification table (94.5%), area under receiver operating characteristic curve (62.6%). CI: confidence interval, LVEF: left ventricular ejection fraction, MACE: major adverse cardiovascular events, MHIS: moderate-high intensity statin, MPI: myocardial perfusion imaging, OR: odds ratio, Ref: reference group

DISCUSSION

Historical data shows that normal stress SPECT MPI is associated with a very low risk of cardiac death and myocardial infarction,[12] with similar results replicated in contemporary studies using cadmium zinc telluride–based gamma cameras.[2] Similarly, CAC scoring has also emerged as a useful prognostic marker for MACE, with CAC >300 conferring almost ten times higher rates of coronary events.[3] Recent data further shows the association between CAC and MACE, including stroke, even among young adults with relatively low cardiac risk.[13] Studies involving concurrent CAC scoring with MPI for detection of subclinical atherosclerosis have shown that CAC can be found in more than half of the patients, with 12% having CAC >300.[14] Although increasing CAC scores results in a stepwise increase in MACE among patients with normal SPECT,[15] the optimal management of these patients is unknown. In this study, we showed that 24-month MACE occurred in 5.5% of patients with normal stress MPI and CAC >300, which differs from previous observations that normal stress MPI carries <1% MACE risk. There could be several reasons for this. First, our patients are relatively older (median age 71 years), and thus, most had to undergo pharmacologic instead of exercise stress. Although a previous study showed that normal pharmacologic stress MPI carries a higher MACE rate than normal exercise MPI,[16] recent studies have showed <1% annual event rates even when adenosine stress was exclusively used.[2] The more likely reason for the higher MACE rates seen in our study is that we have selected a subgroup of patients with higher cardiovascular risk by including only cases with CAC >300. MPI detects perfusion inhomogeneity from flow-limiting stenoses and may not fully identify cases at risk of acute coronary syndromes that are classically attributed to rupture of vulnerable thin-cap fibroatheroma. Furthermore, adequate myocardial perfusion may be maintained through well-developed collaterals, despite extensive coronary disease. CAC, by contrast, is a surrogate for overall atherosclerotic plaque burden and may be a better predictor for MACE. Another possible reason for this observation is related to MPI's reduced sensitivity for multivessel disease. Also described as “balanced ischaemia“, MPI may fail to detect cases with poor perfusion involving all myocardial walls because it only measures relative and not absolute tracer uptake. Therefore, we cannot exclude the possibility that some cases with elevated CAC and “normal” myocardial perfusion may have underlying multivessel disease and hence higher MACE during follow-up.

Given that these patients with elevated CAC are at higher risk despite a reassuring MPI result, we studied the impact of post-MPI statin use on their prognosis. Our data showed that patients on MHIS after MPI have a lower 24-month MACE than those without MHIS. Post-MPI MHIS use also emerged as the only significant predictor of 24-month MACE. The benefits of intensive statin therapy in patients with stable coronary disease are already well known.[17] This is likely related to multiple statin actions such as reduced arterial inflammation, decreased plaque volume and overall increase in atherosclerotic plaque stability.[18] The fact that MACE is only related to the post-MPI (and not the pre-MPI) MHIS status in our study is an important observation, although MACE reduction owing to preventive therapy being started after knowledge of scan results is not a new finding. In a post-hoc analysis of the Scottish Computed Tomography of the Heart (SCOT-HEART) trial, the reduction in MACE in the CT angiography group was attributed to initiation of preventive therapies in cases with nonobstructive plaque because rates of invasive coronary angiography did not differ between groups.[19] Our results underscore the importance of starting MHIS in response to elevated CAC, even when the stress MPI is normal. These findings add to the wealth of data behind the use of CAC in guiding statin use, consistent with the most recent guidelines recommending CAC measurement when a decision about statin therapy is uncertain and initiation of statins when CAC is at least 100.[7]

Antiplatelet agents are commonly used for primary prevention. In our study, neither baseline nor post-MPI antiplatelet use was a significant variable affecting MACE. These results echo the findings of recent primary prevention trials demonstrating the lack of benefit with antiplatelets[20,21] and support the latest guidelines in recommending against routine use of aspirin.[22]

There are several limitations in our study. First, the retrospective nature of the research is highly dependent on the accuracy of electronic medical records and may be prone to referral bias. We were also not able to verify patients' compliance to the prescribed statins. Second, the results of this single centre study may not necessarily be generalisable to other centres or patient populations. Third, the small patient numbers and consequently few events observed result in limited power for a multivariable analysis. Fourth, although it would be ideal to assess the relationship between post-MPI lipid levels and MACE, these results were not included because we observed significant variation in the timing of lipid testing after statin change, frequency of testing and whether fasting was required when the patients were cared for by different primary care physicians after discharge from the cardiology clinic after MPI. As such, the lipid data was deemed to be less reliable for reporting. Finally, although we showed an association between post-MPI MHIS and MACE, we were unable to show that the reduction in MACE is directly caused by initiation or escalation to MHIS after MPI. This is partly because half of the patients in our cohort were on MHIS at baseline. Because statins can increase CAC by inducing plaque calcification,[23] the actual impact of previous statin use on the observed CAC measured at the time of MPI as well as eventual 24-month MACE cannot be fully evaluated. The high prevalence of cardiovascular risk factors such as hypertension and diabetes mellitus among our patients most likely explains why there were many patients on pre-MPI MHIS. A similar situation was observed in the landmark Multi-Ethnic Study of Atherosclerosis study, where 16.3% of the participants were already on lipid lowering medications.[3] Although the study showed the utility of CAC as an independent predictor of coronary disease, patients with coronary events had significantly higher rates of baseline lipid lowering medication use (28.4% vs. 16%).[3] Although we were not able to account for the varying duration of MHIS use before MPI because of the retrospective nature of this study, we attempted to evaluate for confounding effects of pre-MPI MHIS on CAC by showing that baseline CAC was not significantly different in those with or without pre-MPI MHIS and that pre-MPI MHIS status did not eventually predict the occurrence of MACE. Given the above, we acknowledge that there are complex and poorly understood interactions between statin, CAC and MACE that are inherent limitations in the methodology of observational studies like ours.

In conclusion, patients with CAC >300 are at significant risk of MACE at 24 months after normal stress MPI. Post-MPI MHIS use is associated with significantly lower MACE rates and is an independent negative predictor for MACE. Physicians should be proactive in starting MHIS in these patients, despite the apparently reassuring stress MPI results.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.