Assessing the prevalence and predicting factors of an abnormal gated myocardial perfusion SPECT in asymptomatic patients with type 2 diabetes

Considering the significant prevalence of silent myocardial ischemia and its related morbidity and mortality in asymptomatic type 2 diabetic patients, it is not well known whether early screening with MPI is cost-effective and predicting factors are not well elucidated. This was a cross-sectional study including 63 asymptomatic patients with type 2 diabetes mellitus (T2DM), with normal ECG and ejection fraction. Patients with any history of documented valvular, congestive or ischemic heart disease, renal or hepatic failure were excluded. At first all patients were interviewed and checked for risk factors and then patients underwent a two-day rest/stress 99mTc-MIBI gated MPI SPECT. Data was assessed by QPS/QGS and 4DM software and evaluated by a nuclear medicine specialist with summed stress score (SSS) of more than 4 defined as CAD. There were 42 females (67%) and 21 males (33%), with a mean age of 61.33 ± 6.98 years and 7.97 ± 4.86 years history of T2DM. CAD was detected in 26 (41.3%) patients and was significantly associated with male gender, smoking and requiring insulin therapy (P-value = 0.019, 0.046, 0.05, respectively). A significant association was found between the duration of diabetes, especially when > 15 years, and the probability of having CAD. Multivariable logistic regression revealed that smoking; male gender and diabetes duration were the strongest independent predictors of abnormal MPI results. We found a high (46%) prevalence of abnormal stress MPI SPECT in patients with type 2 diabetes mellitus, despite being asymptomatic. Asymptomatic patients with a history of smoking, long duration of diabetes, being under insulin treatment and male gender might benefit from MPI for early detection of silent ischemia.


Introduction
The global prevalence of type 2 diabetes mellitus (T2DM), an essential risk factor for coronary artery disease (CAD), is on the rise [1,2]. On the other hand, CAD as a major cause of morbidity, mortality, and financial costs in patients with T2DM, could be asymptomatic in 20-35% of diabetic patients [3]. However, acute myocardial infarction (MI) and sudden cardiac death might be often the initial presentations [4,5]. Developing at a younger age, higher rate of multivessel disease and being more prone to have complications after MI, such as congestive heart failure, are some of the features indicating the overall poorer outcome of CAD in diabetic patients [6][7][8]. This further emphasizes the importance of detecting CAD at an early stage in these patients [9]. It has been suggested that myocardial perfusion imaging 1 3 (MPI) is a reliable technique for risk assessment of CAD in diabetic patients [10][11][12]. It is still under debate whether screening myocardial ischemia is rational in asymptomatic diabetic patients compared to standard therapeutic strategies or which asymptomatic diabetic needs screening [4,13,14]. The aim of this study was to assess CAD risk, prevalence and predicting factors of an abnormal gated myocardial perfusion SPECT in asymptomatic patients with type 2 diabetes. What distinguishes our study compared to similar papers [15][16][17] is that first, we extensively assessed different systolic and diastolic indices to find even subtle evidence of CAD in diabetics, second; apart from reversible or fixed perfusion defects we looked for high risk features like transient ischemic dilatation (TID) or right ventricle (RV) visualization [18] in the MPIs as an evidence of CAD in asymptomatic diabetics, third, we employed both Cedars-Sinai tools (Quantitative Perfusion SPECT (QPS) and Quantitative gated SPECT (QGS)) and corridor 4DM software(University of Michigan/ Mirada Medical) for better interpretation of scans and the last but not the least we also tried to investigate the effect of risk factors on different vascular territories (LAD, RCA or LCX) based on MPI results.

Patient selection
This was a cross-sectional study, conducted at the nuclear medicine department of Shohada Tajrish hospital, Tehran, Iran. In total, 63 asymptomatic patients with type 2 DM, referred for check-up or pre-operative cardiac evaluation entered the study. An ethical approval was obtained from Shahid Beheshti Medical University Ethics committee.
Patients were interviewed on the appointed scanning day and informed about the goals of study in detail. A written informed consent was obtained from all patients before enrolment. The inclusion criteria were patients with proven T2DM, aged 35-80 years, ejection fraction (EF) > 50% (from echocardiography obtained in less than one month), normal ECG and life expectancy of more than one year. Patients with any history of valvular heart disease, ischemic heart disease, abnormal renal or hepatic failure, coronary intervention, arrhythmia, perfusion defect on prior MPI scans or wall motion abnormality on prior gated stress SPECT-MPI were excluded. Demographic characteristics of participants were documented in a questionnaire.

Image acquisition
All patients underwent a 2-day rest/stress 99mTc-MIBI gated MPI SPECT, according to the European Association of Nuclear Medicine (EANM) guidelines for MPI scan [19]. The rest scan was performed 60 min after IV injection of 555 MBq 99mTc-sestamibi. On the day after, myocardial stress was achieved by exercise (Bruce protocol) or Dipyridamole (0.14 mg/kg/min over 4 min), with 740 MBq 99mTc-sestamibi injected at peak stress. All patients consumed 120 mL whole milk 10 min after tracer injection, as per routine. Imaging was performed in both supine and prone positions. SPECT images were acquired with a variable angle dual head Evo-Voxel Siemens gamma camera (64 projections; 25 s/projection, matrix 64 × 64, zoom of 1.46, 8-frame gated study, low-energy high resolution collimator, in a noncircular orbit).

Image interpretation
Images were first reconstructed with 3D flash algorithm using Siemens Syngo® MI Apps (E.soft) and were analyzed and scored visually and then reconstructed images were checked in QPS, QGS and 4DM software. In case of any inconsistency, MPIs were rechecked to achieve agreement and if any discrepancy remained between visual assessment and results of mentioned software, another qualified nuclear medicine specialist (blind to initial judgment and software results) was asked to evaluate the scan. Scoring of the left ventricular myocardial perfusion during rest and stress was done using 17 segments model. Normal perfusion is indicated on the scale as a score of 0. Mild and moderate perfusion impairment is indicated by 1 and 2 points, respectively. A score of 3 points implies substantial perfusion defect, while a score of 4 points is used to show absent perfusion. The Summed Stress Score (SSS), the Summed Rest Score (SRS) were calculated as a sum of the individual scores from the 17 segments of the polar map obtained during stress and rest and then Summed Difference Score (SDS) can be calculated by subtracting the SRS from the SSS (SDS = SSS − SRS). SSS more than 4 was considered abnormal and was regarded as a CAD [20]. Usually There are three main coronary arteries, the right coronary artery (RCA), left circumflex artery (LCX) and left anterior descending (LAD) Each segment of the myocardium is attributed to one of the these arteries, so the pattern of perfusion defects helped us to record the possible involved coronary artery [20]. Additional findings on MPI, such as transient ischemic dilation (TID) and right ventricular visualization, which are recognized as markers of extensive CAD and predictive of adverse outcomes, were also recorded [21]. TID is defined as a transient dilatation of left ventricle(LV) most likely due to sub-endocardial ischemia. In our study, we divided LV volume at stress to LV volume at rest (volumes were extracted using QPS software) and patients with this ratio of more than 1.13 were regarded to have TID [22]. Visualization of the right ventricle in the stress phase is also proposed as a high risk feature and may be due to overall decrease in LV tracer uptake at stress phase [23].
To assess systolic and diastolic function, as more parameters were available in QGS software, indices from only QGS were only recorded and analyzed.

Statistical analysis
Descriptive data were reported as mean and standard deviation for quantitative variables and number and percentage for qualitative variables. T-test was used to compare the means between the two groups. Kruskal Wallis analysis and multivariate logistic analysis were also employed to compare different groups and assess risk factors, respectively. For all tests, a P-value below 0.05 was considered statistically significant. SPSS version 20 (SPSS Inc., Chicago, IL, The USA) was used for data analysis.

Results
Of 63 patients, 42 were females (67%) and 21 males (33%), with a mean age of 61.33 ± 6.98 years (range 42-78). The mean ± SD duration of diabetes was 7.97 ± 4.86 years. Baseline characteristics of participants are presented in Tables 1  and 2. Fifty-four patients underwent stress with Dipyridamole, two with Dobutamine and seven with exercise.
According to the obtained values of MPI (true SSS > 4, without attenuation artifacts), CAD was detected in 26 (41.3%) patients. Presence of CAD was significantly associated with male gender, and smoking (P-values; 0.019, 0.046, respectively). While hypertension and dyslipidemia were more common in CAD patients compared to those without CAD, no statistically significant association was noted (P-value > 0.05) (Table 3).
Moreover, patients were divided into five groups based on the duration of diabetes. Kruskal Wallis analysis revealed a significant correlation between the duration of diabetes, especially when more than 15 years, and the probability of CAD, based on SSS, SDS and especially in the Left Anterior Descending artery (LAD) territory (P-value = 0.048) (Fig. 1). Besides, requiring Insulin therapy was significantly more prevalent in diabetic patients with CAD compared to those without CAD (P-value = 0.042). (Fig. 2).
Mean End Diastolic Volume (EDV) and End Systolic Volume (ESV) were higher in CAD patients compared to diabetic patients without CAD (68.52 ± 3.41 and 29.35 ± 2.14 mL versus 52.72 ± 2.23 and 18.52 ± 1.47 mL). However, the difference was not statistically significant (P-value > 0.05).
According to multivariate logistic analysis, smoking had a significant association with extent of ischemia (higher SDS, P-value = 46%), perfusion defect in Right Coronary Artery (RCA) territory (P-value = 0.017) and right ventricle (RV) visualization (P-value = 0.034). In addition, duration of DM was significantly associated with higher SSS (P-value = 0.003), higher SRS (P-value = 0.017) and RCA/LAD territory involvement (P-value = 0.008 for RCA, P-value = 0.023 for LAD) as well more frequent TID (P-value = 0.010

Discussion
Many studies have been conducted to evaluate the prevalence of silent myocardial ischemia in asymptomatic diabetic patients, reporting a prevalence of 9-57% [24]. Silent ischemia is characterized by confirmed myocardial ischemia objectively without symptoms of angina or angina equivalents [25]. The autonomic neuropathy in diabetes has been proposed to be the reason for blunted perception of chest pains, justifying a higher proportion of silent ischemia in these patients [26]. The incidence of cardiovascular events and death are significantly higher in type 2 diabetic patients with MPI SPECT abnormalities [17]. Various diagnostic tests are implemented to diagnose CAD, each having their own limitations. For example, rest ECG could yield normal findings in more than a half of patients with chronic stable angina [16]. Also, no agreement has been achieved on which asymptomatic diabetic patients should be screened [27,28]. Scholte et al. reported a 13% rate of cardiac deaths and cardiovascular events in patients with type 2 diabetes during a 3-year follow-up. None of the known cardiovascular risk factors have been shown to have enough sensitivity and specificity to predict silent myocardial ischemia in diabetic patients [29].
However, our study revealed a 41.3% prevalence of silent ischemia in asymptomatic type 2 diabetic patients, higher than 37% reported by Al-Humaidi et al. [15]. They reported that perfusion defects were related to disease duration, insulin use, nephropathy, and neuropathy. Their results regarding the duration of diabetes or insulin use were in accordance with our findings which should be examined in further studies as well.
CAD was diagnosed in our patients despite having no compromise in LVEF in echocardiography or electrocardiographic (ECG), highlighting the superior sensitivity of MPI SPECT in detecting silent ischemia/CAD. This is in accordance with the findings of Mohagheghie et al. who reported that CAD was detected by MPI in one third of asymptomatic diabetic patients with normal ECG and no evidence of peripheral arterial disease [30]. Moreover, CAD was significantly more prevalent in patients with diabetes mellitus of more than 15 years, consistent with other studies which indicate an increased risk of cardiac events in prolonged diabetes mellitus [31][32][33].
Wu et al. reported that asymptomatic postmenopausal women with DM had a higher SSS and SDS than agematched men [34]. While our study found that male gender, regardless of smoking and requiring insulin treatment, is a strong predictor of myocardial perfusion abnormalities in asymptomatic diabetic patients. This might be due to poor control of blood sugar in men and intentional non-adherence to medications compared to women. However, women in our study were mainly pre-menopausal (34 out of 42).
Moreover, Salehi et al., concluded that DM was independently associated with abnormal MPI in asymptomatic patients with a normal LVEF, also claiming that smoking and longer duration of DM were correlated with myocardial perfusion abnormalities, which is consistent with our findings.   However, they found no significant difference between males and females in terms of abnormal MPI findings [35]. This might be due to different study population or inclusion criteria. MPI SPECT is a readily accessible, easily performed modality compared to other evolving modalities such as CTangiography, MRI or newly discovered biomarkers, as evidence is scarce in supporting their incremental values in predicting further cardiac events. Moreover, MPI SPECT offers an automated evaluation and is less operator-dependent [36].
Our study had some limitations. Our sample size was quite small. Besides, the effect of some of possible confounding factors such as underlying liver or kidney disease (abnormal liver function tests or serum creatinine levels) was not assessed in our study, as we excluded such patients from our study. Further, prospective studies, with larger population and longer follow-ups are necessary to determine the prognostic value of an abnormal MPI in asymptomatic diabetic patients.

Conclusion
The current study revealed a high (46%) prevalence of abnormal stress MPI SPECT in patients with type 2 diabetes mellitus, despite having a normal ECG and EF in echocardiography, which in turn indicates a higher risk of fatal cardiac events, if not detected early. Smoking, duration of diabetes, requiring insulin treatment and male gender were independent predictors of abnormal MPI, proceeding to silent ischemia. These findings may assist cardiologists in choosing asymptomatic diabetic patients for early screening of CAD, thereby lowering the risk of CAD-related morbidity and mortality.