The International Journal of Cardiovascular Imaging https://doi.org/10.1007/s10554-017-1295-8

ORIGINAL PAPER

Myocardial perfusion in patients with non-ischaemic systolic heart failure and type 2 diabetes: a cross-sectional study using Rubidium-82 PET/CT Christina Byrne1,2,3 · Philip Hasbak2 · Andreas Kjaer2,3 · Jens Jakob Thune4 · Lars Køber1,3 Received: 16 October 2017 / Accepted: 22 December 2017 © Springer Science+Business Media B.V., part of Springer Nature 2017

Abstract Both patients with non-ischaemic systolic heart failure and patients with type 2 diabetes (T2DM) often have reduced myocardial blood flow without significant coronary atherosclerosis. However, the mechanisms are not fully understood. The aim of this study was to investigate whether perfusion is reduced additionally when the 2 are combined. In a cross-sectional study, we scanned patients with non-ischaemic systolic heart failure with and without T2DM using Rubidium-82 positron emission tomography/computed tomography at rest and adenosine-induced stress, thereby obtaining the myocardial flow reserve (myocardial flow reserve (MFR) = stress flow/rest flow) as a measure of the myocardial vasomotor function; 28 patients with T2DM and 123 without T2DM were included. All patients received heart failure treatment according to guidelines. Multiple regression analysis was performed to assess the association between T2DM and MFR. Age [68 (60–75) years vs. 68 (62–72) years; P = 0.84] and female sex (21% vs. 33%; P = 0.26) were similar between patients with and without T2DM. Patients with T2DM had higher body mass index, (29.9 vs. 26.5 kg/m2; P = 0.02), higher blood glucose (6.2 vs. 5.7 mmol/L; P = 0.03), more often hypertension (50 vs. 27%; P = 0.02) and received more cholesterol lowering medication (61 vs. 35%; P = 0.02). Apart from this, the groups were similar. In a multivariable analysis, MFR was 16% lower in patients with T2DM compared with patients without [estimate − 16%; 95% confidence interval (CI) − 29 to − 0.7%; P = 0.04]. Patients with T2DM and systolic heart failure have lower myocardial flow reserve compared with heart failure patients without T2DM. Keywords  Myocardial perfusion · Positron emission tomography · Non-ischaemic systolic heart · Failure · Diabetes Abbreviations 82 Rb-PET/CT Rubidium-82 positron emission tomography/computed tomography CACS Coronary artery calcium score Electronic supplementary material  The online version of this article (https://doi.org/10.1007/s10554-017-1295-8) contains supplementary material, which is available to authorized users. * Christina Byrne [email protected] 1



Department of Cardiology, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark

2



Department of Clinical Physiology, Nuclear Medicine & PET and Cluster for Molecular Imaging, Rigshospitalet and University of Copenhagen, Copenhagen, Denmark

3

Faculty of Health Sciences, University of Copenhagen, Copenhagen, Denmark

4

Department of Cardiology, Bispebjerg Hospital, University of Copenhagen, Copenhagen, Denmark



CRT​ Cardiac resynchronisation therapy DANISH A DANish randomized, controlled, multicenter study to assess the efficacy of implantable cardioverter defibrillator in patients with non-ischaemic systolic heart failure on mortality MBF Myocardial blood flow MFR Myocardial flow reserve MPI Myocardial perfusion imaging RPP Rate pressure product SDS Summed difference score SRS Summed rest score SSS Summed stress score TID Transient ischaemic dilation

13

Vol.:(0123456789)



Introduction Patients with non-ischaemic systolic heart failure often have reduced myocardial perfusion with regional perfusion defects, even though there is no significant coronary atherosclerosis [1–3]. Yet, there is no definitive explanation for the mechanism behind these perfusion defects. One explanation may be endothelial dysfunction with impaired angiogenesis and hereby impaired perfusion, leading to myocardial ischaemia and heart failure [4]. Another may be cardiac dysfunction due to decreased cardiac output [2]. Patients with type 2 diabetes (T2DM) also have reduced coronary flow reserve, expressing reduced global perfusion, compared with control persons [5–7]. Patients with non-ischaemic systolic heart failure have increased mortality, if they also have T2DM [8]. Therefore it is relevant to investigate this group further. Non-invasive quantification of myocardial blood flow with 82Rb-PET/CT from high-resolution images using a state-of-the-art time-of-flight-capable scanners is now a robust method recently validated with 15O-water–PET [9]. Cardiac PET/CT enables the detection of microvascular disease before structural epicardial coronary artery disease changes occur. Myocardial flow reserve (MFR) describes the vasodilator function of the coronary circulation, and outcome studies have shown that MFR and myocardial blood flow (MBF) are associated with poor prognosis in patients with heart failure [10–12]. It has not previously been investigated whether MFR is further reduced in patients with both non-ischaemic systolic heart failure and T2DM. To gain more knowledge about this group, we investigated whether changes in myocardial perfusion were associated with T2DM in patients with non-ischaemic systolic heart failure, assessed by myocardial flow reserve (MFR) using 82Rb-PET/CT.

The International Journal of Cardiovascular Imaging

mL) of N-terminal pro-brain natriuretic peptide (NT-proBNP) despite optimal medical treatment. An ischaemic cause of heart failure was excluded by coronary artery catheterisation in 97.4% of patients in this sub study and CT angiogram in the remaining. Exclusion criteria were pregnancy, severe chronic obstructive pulmonary disease (COPD)/asthma, blood pressure above 200/110 mmHg or systolic blood pressure below 90 mmHg, allergy or intolerance to theofyllin or adenosine and inability to adhere to the protocol. A total of 151 patients with non-ischaemic systolic heart failure, of those 28 patients with T2DM, underwent 82Rb-PET between May 2015 and September 2016. Figure 1 shows the inclusion flowchart. The Scientific Ethics Committee of the Capital Region of Denmark and the Danish Data Protection Agency approved the protocol [protocol number H-15000346]. We obtained informed oral and written consents from all patients and the trial was performed in accordance with the principles of the Declaration of Helsinki.

PET imaging All patients underwent PET myocardial perfusion imaging (MPI) at rest and during adenosine-stress in one session. We instructed the patients not to consume caffeine and theophylline-containing substances and medications for 12 h before the scan. In order to obtain the most clinically relevant

Materials and methods Study population In this cross-sectional study, we recruited patients with non-ischaemic systolic heart failure from the DANISH Study (A DANish randomized, controlled, multicenter study to assess the efficacy of Implantable cardioverter defibrillator in patients with non-ischaemic systolic heart failure on mortality) [13]. Patients were included in our sub study in the late follow up period of the DANISH trial. Inclusion criteria were age > 18 years, documented nonischaemic systolic heart failure (left ventricular ejection fraction (LVEF) ≤ 35%), and increased levels (> 200 pg/

13

Fig. 1  Inclusion flow chart. DANISH [13] Centre 1 and 2: Rigshospitalet and Gentofte Hospital, Copenhagen, Denmark. COPD chronic obstructive pulmonary disease

The International Journal of Cardiovascular Imaging

results, all patients continued their medications for heart failure. For rest and stress imaging, patients received 1062 MBq (IQR 1018 − 1273) and 1058 MBq (IQR 1014 − 1269) 82Rb supplied from a CardioGen-82Sr/82Rb generator manufactured for Bracco Diagnostics Inc., Princeton, NJ. We used a Siemens Biograph mCT/PET 128-slice scanner (Siemens Healthcare, Knoxville, TN, USA), and both rest and stress images were acquired ECG-gated in list mode for 7 min from the start of the 82Rb infusion. To stress the patients, we used adenosine infusion (0.14 mg/kg/min) for 6 min, and initiated the stress 82Rb infusion 2.5 min after start of adenosine administration. Additionally, we performed a low-dose CT for attenuation correction before the rest study; this was repeated after the stress study if patient motion was detected. In all patients we acquired coronary artery calcium score (CACS) images from a non-contrast breath-hold CT, as per clinical routine. We calculated the CACS according to the Agatston score system [14].

Quantitative and semi quantitative analysis For MBF quantification of dynamic rest and stress images, we used the method previously described by Armstrong et  al. [15] We performed the MBF quantification using quantitative gated nuclear imaging: QGS/QPS v. 2015.5 (Cedars-Sinai Cardiac Suite, Los Angeles, USA), based on a single-compartment model for 82Rb tracer kinetics proposed by Lortie et al. [16] As a validated standard measure of myocardial perfusion, MFR was defined as MBF during maximal hyperemia by adenosine stress divided by MBF at rest. To correct the MBF at rest for baseline work, MBF was divided with the rate pressure product (RPP), which is the systolic blood pressure, times the heart rate, and multiplied by 10,000 [17]. Further MFR was divided into low (≤ 2.0), borderline (> 2.0–2.5), and normal (> 2.5) [18]. Normal values for resting MBF were defined as 0.82 mL/g/min ± 30% and for stress MBF 3.3 mL/g/min ± 31% [19]. As according to the AHA 17 myocardial segment model, [20] the irreversible and reversible perfusion defects were computed automatically as summed rest score (SRS), summed stress score (SSS) and summed difference score (SDS = SSS − SRS) with the QPET software from Cedars-Sinai. Summed stress score was defined as follows 0–3: normal (< 5% myocardium with perfusion abnormalities), 4–7: mildly abnormal (5–10% myocardium with perfusion abnormalities), > 8 moderately or severely abnormal (> 10% myocardium with perfusion abnormalities) [21]. The software also calculated the LVEF.

Statistical analysis For the statistical analysis, we used SAS version 9.4 (SAS Institute, Cary, NC, USA). Continuous variables were expressed as medians and interquartile ranges. Categorical

variables were expressed in percentages. We analysed differences between groups with unpaired t test, after logtransformation of variables necessary to obtain a normal distribution, or Wilcoxon two-sample test for continuous variables and Chi square test for categorical variables. We used the general linear model (GLM) procedure for multiple regression analysis of explanatory variables. Further, we performed a sensitivity analysis to adjust for differences between groups, matching for age, sex, BMI and atrial fibrillation during scan. P values < 0.05 were considered statistically significant.

Results The 151 patients in our study were largely comparable to the rest of the study population of the main study (Supplementary Table S1). The major differences were a significantly lower NT-pro-BNP in patients included in our study (851 vs. 1230; P < 0.01) and higher estimated glomerular filtration rate (eGFR) (78 vs. 72 mL/min/1.73 m2; P = 0.02). Fewer were treated with cardiac resynchronisation therapy (CRT) (50 vs. 59%; P = 0.03), and by design none of the patients included in our study had COPD (0 vs. 14%). Except from these variables, our study population was similar to the remaining patients in the main study. Demographics of the 151 patients are summarised in Table 1. Patients with T2DM had higher body mass index (BMI) than patients without T2DM (29.9 vs. 26.5 kg/m2; P = 0.02). Blood glucose measured right before the scan was higher among patients with T2DM (6.2 vs. 5.7 mmol/L; P = 0.03) and hypertension was more common in this group (50 vs. 27%; P = 0.02). Also, more patients with T2DM were treated with cholesterol lowering medication (61 vs. 35% P = 0.02). Apart from these variables, the two groups were similar.

PET/CT data Table 2 shows findings from the 82Rb PET/CT scan. RPP and LVEF were similar in the two groups as was the amount of 82Rb administered. Patients with T2DM received higher doses of adenosine (77 vs. 69 mg; P < 0.01) due to their higher BMI and coronary calcium score was significantly higher in the T2DM group (311 vs. 73; P = 0.03).

Global myocardial blood flow and myocardial flow reserve Myocardial flow reserve tended to be lower in patients with T2DM, although this was not significant (2.01 vs. 2.27; P = 0.15). MFR was in the borderline interval for

13



The International Journal of Cardiovascular Imaging

Table 1  Characteristics of study population Median age (IQR) (year) Female sex—no. (%) Median body-mass index (IQR) (kg/m2) Median NT-pro-BNP (IQR) (pg/mL) Median e-GFR (IQR)—mL/min/1.73 m2 Median P-glucose (IQR) (mmol/L) Median LVEF baseline (IQR) (%) Coexisting conditions—no. (%)  Hypertension  Left bundle branch block—no. (%) Cause of heart failure no. (%)  Idiopathic  Valvular  Hypertension  Other Medications—no. (%)  ACE-inhibitor or ARB  Betablocker  Aldosterone receptor antagonist  Calcium antagonist  Long lasting nitrates  Cholesterol lowering medication  Anticoagulation treatment  Acetylsalicylic acid Device therapy—no. (%)  CRT​  ICD

Patients with T2DM (N = 28)

Patients without T2DM P value (N = 123)

68 (60–75) 6 (21) 29.9 (27.0-34.7) 998 (550-1,590) 66 (53–79) 6.2 (5.4–9.9) 25 (20–30)

68 (62–72) 41 (33) 26.5 (24.5–30.2) 824 (438-1,939) 65 (52–80) 5.7 (5.3–6.2) 25 (20–31)

0.84 0.26 0.02 0.51 0.98 0.03 0.22

14 (50) 16 (59)

33 (27) 67 (60)

0.02 1.0

20 (71) 1 (4) 3 (11) 4 (14)

74 (60) 8 (7) 8 (7) 33 (27)

0.57

28 (100) 28 (100) 13 (46) 1 (4) 1 (4) 17 (61) 15 (54) 7 (25)

123 (100) 115 (94) 64 (62) 3 (2) 2 (2) 43 (35) 48 (39) 26 (21)

– 0.35 0.68 0.56 0.46 0.02 0.20 0.62

15 (54) 14 (50)

69 (56) 66 (53.7)

0.84 0.83

IQR Interquartile range, NT-proBNP N-terminal pro-brain natriuretic peptide, e-GFR estimated glomerular filtration rate, LVEF left ventricular ejection fraction, ACE angiotensin-converting enzyme, ARB angiotensin-receptor blocker, CRT​ cardiac resynchronisation therapy, ICD implantable cardioverter-defibrillator. Two-sided P value determined by Wilcoxon two-sample or Chi square test

both groups (2.0–2.5). The resting MBF was comparable in the two groups (0.96 vs. 0.94 mL/g/min; P = 0.89) and within the normal interval (0.82 mL/g/min ± 30%). This did not change after correction for cardiac work expressed as rate-pressure product (RPP). During stress, the MBF tended to be lower in the group with T2DM (1.96 vs. 2.28 mL/g/min; P = 0.09), (Table 2). The stress MBF was lower than the normal interval in the group with T2DM and low normal in the group without T2DM (normal range 3.3 mL/g/min ± 31%). MFR also tended to increase with increasing BMI [estimate 1% / (kg/m 2); 95% confidence interval (CI) − 0.04 to 3%; P = 0.06] (Fig. 2). MBF at rest tended to be lower with increasing BMI (− 0.01 mL/g/min 95% CI − 0.06 to 0.00; P = 0.06) whereas MBF at stress did not change with changes in BMI (0.00 mL/g/min 95% CI − 0.02 to 0.02; P = 0.99).

13

Multiple regression analysis In a multivariable analysis controlling for sex, age, BMI, atrial fibrillation during scan, left bundle branch block, hypertension, NT-pro-BNP, left ventricular mass, and LVEF at baseline we found a significant lower MFR in patients with T2DM (estimate − 16%; 95% CI − 29 to − 0.7%; P = 0.04). In addition, MFR was significantly lower in patients with atrial fibrillation during the scan (estimate − 31%; 95% CI − 42 to − 17%; P < 0.001) and MFR increased significantly with increasing BMI (estimate 2% / (kg/m 2); 95% CI 0.4–4%; P = 0.02) and decreased significantly with increasing left ventricular mass (estimate − 2% / (10 g); 95% CI − 3 to − 0.3%; P = 0.02), (Table 3).

The International Journal of Cardiovascular Imaging Table 2  Scan data

Atrial fibrillation/flutter during scan—% CACS (IQR) Left ventricular mass (IQR) (g) Rest scan  Rubidium-82 (IQR) (Mbq)  Median systolic blood pressure (IQR) (mmHg)  Median diastolic blood pressure (IQR) (mmHg)  Median heart rate (IQR) (bpm)  Median rate-pressure product (IQR)  Median LVEF (IQR) (%)  Median end diastolic volume (IQR) (mL)  Median end systolic volume (IQR) (mL) Stress scan  Rubidium-82 (IQR) (Mbq)  Adenosin (IQR) (mg)  Median systolic blood pressure (IQR) (mmHg)  Median diastolic blood pressure (IQR) (mmHg)  Median heart rate (IQR) (bpm)  Median rate-pressure product (IQR)  Median LVEF (IQR) (%)  Median end diastolic volume (IQR) (mL)  Median end systolic volume (IQR) (mL) Myocardial perfusion data  Median global rest MBF (IQR) (mL/g/min)  Median global stress MBF (IQR) (mL/g/min)  Mean MFR* (95% CI)  Median TID (IQR)  Median SRS (IQR)  Median SSS (IQR)  Median SDS (IQR)

Patients with T2DM (N = 28)

Patients without T2DM (N = 123)

P value

7 (28) 311 (31–694) 179 (113–233)

23 (20) 73 (0–307) 170 (148–202)

0.42 0.03 0.13

1081 (1040–1297) 111 (99–118) 65 (60–69) 71 (66–77) 7628 (6750–8814) 39 (28–55) 143 (126–209) 90 (60–144)

1062 (1014–1273) 107 (98–119) 62 (56–69) 66 (60–72) 7035 (6161–8208) 46 (35–56) 128 (101–174) 69 (44–108)

0.45 0.83 0.56 0.02 0.08 0.22 0.03 0.10

1079 (1031–1291) 77 (70–87) 112 (105–119) 63 (58–70) 74 (69–80) 8536 (7457–9530) 43 (32–60) 162 (137–227) 90 (61–139)

1058 (1014–1269) 69 (61–79) 108 (99–122) 62 (55–68) 74 (66–81) 7760 (6831–9250) 51 (37–62) 139 (112–188) 68 (44–108)

0.51 < 0.01 0.56 0.75 0.70 0.31 0.18 0.03 0.07

0.96 (0.77–1.10) 1.96 (1.36–2.42) 2.01 (1.72–2.36) 1.08 (1.02–1.17) 3.0 (0.0–5.0) 4.5 (3.0–7.0) 2.0 (1.0–4.5)

0.94 (0.75–1.09) 2.28 (1.72–2.79) 2.27 (2.12–2.43) 1.07 (1.02–1.14) 1.0 (0.0–3.0) 4.0 (2.0–7.0) 2.0 (1.0–5.0)

0.89 0.09 0.15 0.81 0.15 0.40 0.76

IQR interquartile range, LVEF left ventricular ejection fraction, MBF myocardial blood flow, MFR myocardial flow reserve CI confidence interval, TID transient ischaemic dilation, SRS summed rest score, SSS summed stress score, SDS summed difference score *Myocardial flow reserve back transformed from log2(MFR). Two-sided P value determined by unpaired T test, Wilcoxon two-sample test or Chi square test

Matched study

Fig. 2  Myocardial flow reserve (MFR) on log2 scale as function of body mass index (BMI) in patients without and with type 2 diabetes. MFR increased with increasing BMI both in patients with and without type 2 diabetes

In addition to the multivariable analysis, we performed matched analyses, where we compared 24 of the patients with T2DM with 24 without T2DM matched on age, sex, BMI and atrial fibrillation during scan. Supplementary Tables S2 and S3 summarise the characteristics and scan parameters of the matched study population with no significant differences between groups. In the matched study MFR was significantly lower in patients with T2DM (2.00 vs. 2.47; P = 0.04), (Supplementary Table S3). In a multivariable analysis of the matched population, controlling for sex, age, BMI, atrial fibrillation during scan, left bundle branch block, hypertension, NT-pro-BNP, left ventricular

13



The International Journal of Cardiovascular Imaging

Table 3  Factors associated with myocardial flow reserve Univariable

Myocardial flow reserve  Type 2 diabetes  Sex  Age (10 years)  Body mass index (kg/m2)  Atrial fibrillation during scan  Left bundle branch block  Hypertension  Log2(NT-pro-BNP)  Left ventricular mass (10 g)  LVEF at baseline (10%)

Multivariable

Percent change per unit* (95% CI)

P value

Percent change per unit* (95% CI)

P value

− 11.1 (− 24.3; 4.5) − 8.5 (− 20.0; 4.6) − 8.0 (− 13.5; − 2.1) 1.3 (− 0.04; 2.7) − 27.6 (− 37.7; − 16.0) 1.2 (− 11.3; 15.5) 0.8 (− 12.0; 15.4) − 3.8 (− 8.1; 0.7) − 2.0 (− 3.3; − 0.6) − 6.9 (− 15.7; 2.9)

0.15 0.19 < 0.01 0.06 < 0.0001 0.85 0.91 0.10 < 0.01 0.16

− 16.1 (− 29.2; − 0.7) 3.0 (− 11.1; 19.4) − 2.9 (− 9.9; 4.7) 1.9 (0.4; 3.5) − 30.6 (− 42.1; − 16.9) − 11.5 (− 23.5; 2.3) − 0.3 (− 13.0; 14.2) 0.4 (− 5.0; 6.1) − 1.8 (− 3.2; − 0.3) − 7.3 (− 17.1; 3.7)

0.04 0.69 0.44 0.02 < 0.001 0.10 0.96 0.90 0.02 0.18

CI confidence interval, NT-pro-BNP N-terminal pro-brain natriuretic peptide, LVEF Left ventricular ejection fraction *Estimated differences are expressed in relative terms, i.e., as a percentage

changed from − 16 to − 12% and at the same time the number of patients in the multivariable analysis fell from 126 to 117. Figure 3 illustrates decreasing MFR with increasing CACS in patients with and without T2DM.

Myocardial perfusion defects

Fig. 3  Myocardial flow reserve (MFR) on log2 scale as function of coronary calcium score (CACS) in patients without and with type 2 diabetes. At low levels of CACS, MFR was decreased in patients with type 2 diabetes compared to patients without. However this relationship seemed to change with increasing CACS

mass and LVEF at baseline we still found that MFR was significantly lower in patients with T2DM (estimate − 17%; 95% CI − 30 to − 1%; P = 0.04). Also, MFR was still significantly lower in patients with atrial fibrillation during scan (estimate − 32%; 95% CI − 47 to − 13%; P < 0.01) and MFR significantly increased with increasing BMI (estimate 3%; 95% CI 1–5%; P < 0.01) and decreased with increasing left ventricular mass (estimate − 2% / (10 g); 95% CI − 5 to − 0.2%; P = 0.03), (Supplementary Table 4).

Coronary artery calcium score (CACS) We found that MFR was significantly negatively correlated with CACS (R2 = 0.23; P < 0.01). If we added CACS to the multivariable model instead of T2DM, the effect of CACS remained significant (P = 0.03). When adding both CACS and T2DM to the multivariable model, both variables became non-significant. CACS: P = 0.06 and T2DM: (estimate − 12%; 95% CI − 26 to 4%; P = 0.14). The estimate

13

In all patients, we measured the summed rest-, stress- and difference scores. We found no difference in the three scores between the group with and without T2DM although summed rest score and summed stress score tended to be a little higher in the T2DM group (Table 2). In both groups the median summed stress score reached the level of mildly abnormal (score 4–7).

Discussion In the present study, we investigated MFR in patients with non-ischaemic systolic heart failure with and without T2DM using 82Rb PET imaging. In an adjusted analysis, we found that MFR was significantly lower in patients with T2DM. We reproduced this finding in a matched design with 24 patients with T2DM and 24 patients without T2DM matched on age, sex, BMI and atrial fibrillation during scan. This finding is congruent with previous findings in patients with T2DM without heart failure [6, 7]. In both our groups with heart failure with and without T2DM we found a decreased MFR compared to normal values, most prominent in the T2DM group. However, average MFR in our patients was higher than the 1.89 ± 0.68 previously found in patients with non-ischaemic systolic heart failure in a the study by Majmudar et al. [10] This may be explained by a more optimal treatment with angiotensin-converting enzymes/

The International Journal of Cardiovascular Imaging

angiotensin-receptor blockers and beta-blockers in our study population (99% vs. 54% and 96% vs. 65% respectively). Another explanation could be that the sickest patients of our main study population died prior to start of inclusion to the 82Rb-PET sub study and that mean MFR would have been lower if we had scanned a random selection of patients at baseline. This theory is supported by the fact that our patients had lower NT-pro-BNP and higher e-GFR than the rest of the main study population. Stress MBF was also reduced in both groups, whereas resting MBF was within the normal interval. There is no clear explanation for the lower myocardial perfusion either in patients with non-ischaemic heart failure or in patients with T2DM. An MFR less than 2.5 has been defined as reduced [18]. Applying this definition to our study, 69% of patients with T2DM in addition to heart failure and 56% of patients without T2DM had reduced MFR. It has previously been shown that the lower MFR in patients with T2DM, without heart failure, could be caused by a higher resting perfusion [5, 22] that may be caused by a higher oxygen demand at rest. We have previously found higher resting perfusion in patients with T1DM [22]. However, our findings in the current study showed no difference in resting perfusion between patients with non-ischaemic heart failure with or without T2DM. This might be explained by the general perfusion impairment observed in this group of patients with non-ischaemic heart failure. Part of the explanation for the lower MFR in the group with T2DM could be related to epicardial coronary artery disease, hence higher CACS. CACS was significantly higher in the group with T2DM and MFR significantly decreased with increasing CACS. This confirms earlier findings [7, 23]. Moreover, the significant correlation between T2DM and reduced MFR was attenuated when adjusting for CACS and the reduction in MFR in T2DM patients changed from 16 to 12%. However, this indicates that CACS explained only some of the reduced MFR in the T2DM group. The fact that CACS also became non-significant in the multivariate analysis with T2DM indicates an interaction between the two. Figure 3 suggests that the effect of T2DM on MFR was larger in patients with low CACS and that the effect of T2DM became masked with increasing CACS. Reduced MFR in patients with T2DM could be due to arterial stiffness which is also correlated with higher CACS [24] and small vessel disease not measurable on CACS. We found no significant difference between groups in MBF at rest or at stress but only in MFR. This also suggests that the difference is to be found in impaired vasomotor function e.g. due to arterial stiffness. In patients with T2DM, Nahser et al. found significantly reduced maximal coronary vasodilation combined with impaired regulation of coronary flow in response to increases in myocardial demand. The authors found that differences in

resting haemodynamics or incidence of hypertension did not contribute to differences in coronary microvascular function between patients with T2DM and patients without [6]. These findings are in line with ours. Our results of myocardial perfusion defects are congruent with previous findings of increased summed stress score in patients with T2DM [18]. In addition, we found that MFR was significantly lower in patients with atrial fibrillation during scan. This finding concurs with former results by Range et al. [25, 26]. Changes in haemodynamics in patients with atrial fibrillation could, apart from acceleration in heart rate, be explained by irregularity of the ventricular cycle length [27]. In the current study, we found that MFR significantly increased with increasing BMI. We showed this in both the overall study and in the matched sub study. Also the MBF at rest tended to be lower in the group with T2DM. This is in contrast to previous studies in obese patients without clinically known heart disease that have found reduced MFR [28]. Our findings might somehow be related to the obesity paradox, where increased BMI has a potentially protective effect on mortality in patients with chronic HF [29]. The patients with T2DM had significantly higher BMI than patients without T2DM. This could explain that our results showed significantly lower MFR in patients with T2DM in the matched study and in the multiple regression analysis without reaching the level of significance in the unadjusted test. Our study provides additional information about reduced myocardial perfusion in patients with non-ischaemic systolic heart failure and T2DM compared with patients with nonischaemic systolic heart failure and no history of T2DM. This may indicate that the disease progression of the myocardium has advanced further in patients with combined non-ischaemic heart failure and T2DM, and might explain part of the findings of increased mortality in patients with T2DM as comorbidity to non-ischaemic systolic heart failure [8]. Further studies including long term studies are needed to explore the prognostic importance of MFR in relation to other risk factors. Secondly, it will be interesting to investigate whether potential treatment strategies with future angiogenic treatment targeted at the microvascular perfusion could improve the prognosis in patients with T2DM as comorbidity to heart failure [30].

Study limitations Some limitations should be considered in interpreting our findings. The study was conducted in a minor group of patients with non-ischaemic systolic heart failure with different comorbidities in addition to diabetes. Nevertheless, the study population was similar to the general population

13



of patients with non-ischaemic systolic heart failure and we reproduced our findings in the matched study with no difference in comorbidity except T2DM. The time span from coronary angiography to 82Rb-PET scan differed between individual patients. Conversely, none of the patients complained of angina at the time of the scan, no one had received interventional treatment and only one person in each group had received long lasting nitrates in the interim.

The International Journal of Cardiovascular Imaging

7.

8.

9.

Conclusion In this study of myocardial flow reserve in patients with nonischaemic systolic heart failure and T2DM using quantitative 82 Rb PET imaging, we found that patients with T2DM in addition to their heart failure had lower MFR compared with patients with heart failure without diabetes. Our results confirm that also in patients with non-ischaemic systolic heart failure, T2DM remains an independent impairing factor to the myocardial vasomotor function.

10.

Funding  Københavns Universitet (Copenhagen, DK), Hjerteforeningen (Copenhagen, DK), Arvid Nielssons Fond (Copenhagen, DK), Grosserer Valdemar Foersom og hustru Thyra Foersoms Fond (Copenhagen, DK), Snedkermester Sophus Jacobsen og hustru Astrid Jacobsens Fond (Copenhagen, DK), and Eva og Henry Frænkels Mindefond (Holte, DK).

13.

Compliance with ethical standards  Conflict of interest  The authors declare that they have no conflict of interest.

References 1. O’Neill JO, McCarthy PM, Brunken RC et al (2004) PET abnormalities in patients with nonischemic cardiomyopathy. J Card Fail 10:244–249. https://doi.org/10.1016/j.cardfail.2003.09.007 2. de Jong RM, Tio RA, van der Harst P et  al (2009) Ischemic patterns assessed by positron emission tomography predict adverse outcome in patients with idiopathic dilated cardiomyopathy. J Nucl Cardiol 16:769–774. https://doi.org/10.1007/ s12350-009-9130-9 3. Iskandrian AS, Hakki AH, Newman D (1985) The relation between myocardial ischemia and the ejection fraction response to exercise in patients with normal or abnormal resting left ventricular function. Am Heart J 109:1253–1258 4. De Boer RA, Pinto YM, Van Veldhuisen DJ (2003) The imbalance between oxygen demand and supply as a potential mechanism in the pathophysiology of heart failure: the role of microvascular growth and abnormalities. Microcirculation 10:113–126. https:// doi.org/10.1038/sj.mn.7800188 5. Picchi A, Limbruno U, Focardi M et al (2011) Increased basal coronary blood flow as a cause of reduced coronary flow reserve in diabetic patients. AJP Heart Circ Physiol 301:H2279–H2284. https://doi.org/10.1152/ajpheart.00615.2011 6. Nahser PJ, Brown RE, Oskarsson H et al (1995) Maximal coronary flow reserve and metabolic coronary vasodilation in patients

13

11. 12.

14.

15.

16.

17. 18.

19.

20.

21.

22.

with diabetes mellitus. Circulation 91:635–640. https://doi. org/10.1161/01.CIR.91.3.635 von Scholten BJ, Hasbak P, Christensen TE et al (2016) Cardiac 82Rb PET/CT for fast and non-invasive assessment of microvascular function and structure in asymptomatic patients with type 2 diabetes. Diabetologia 59:371–378. https://doi.org/10.1007/ s00125-015-3799-x Johansson I, Dahlström U, Edner M et al (2016) Prognostic implications of type 2 diabetes mellitus in ischemic and non-ischemic heart failure. J Am Coll Cardiol 68:1404–1416. https://doi. org/10.1016/j.jacc.2016.06.061 Germino M, Ropchan J, Mulnix T et al (2016) Quantification of myocardial blood flow with 82Rb: validation with 15O-water using time-of-flight and point-spread-function modeling. EJNMMI Res. https://doi.org/10.1186/s13550-016-0215-6 Majmudar MD, Murthy VL, Shah RV et al (2015) Quantification of coronary flow reserve in patients with ischaemic and non-ischaemic cardiomyopathy and its association with clinical outcomes. Eur Heart J Cardiovasc Imaging 16:900–909. https:// doi.org/10.1093/ehjci/jev012 Neglia D, Michelassi C, Trivieri MG et al (2002) Prognostic role of myocardial blood flow impairment in idiopathic left ventricular dysfunction. Circulation 105:186–193 Hagemann CE, Ghotbi AA, Kjær A, Hasbak P (2015) Quantitative myocardial blood flow with rubidium-82 PET: a clinical perspective. Am J Nucl Med Mol Imaging 5:457–468 Køber L, Thune JJ, Nielsen JC et al (2016) Defibrillator implantation in patients with non-ischemic systolic heart failure. N Engl J Med. https://doi.org/10.1056/NEJMoa1608029 Agatston AS, Janowitz WR, Hildner FJ et  al (1990) Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol 15:827–832. https://doi. org/10.1016/0735-1097(90)90282-T Armstrong IS, Tonge CM, Arumugam P (2014) Impact of point spread function modeling and time-of-flight on myocardial blood flow and myocardial flow reserve measurements for rubidium-82 cardiac PET. J Nucl Cardiol 21:467–474. https://doi.org/10.1007/ s12350-014-9858-8 Lortie M, Beanlands RSB, Yoshinaga K et al (2007) Quantification of myocardial blood flow with 82Rb dynamic PET imaging. Eur J Nucl Med Mol Imaging 34:1765–1774. https://doi. org/10.1007/s00259-007-0478-2 Czernin J, Muller P, Chan S et al (1993) Influence of age and hemodynamics on myocardial blood flow and flow reserve. Circulation 88:62–69. https://doi.org/10.1161/01.CIR.88.1.62 Schindler TH, Schelbert HR, Quercioli A, Dilsizian V (2010) Cardiac PET imaging for the detection and monitoring of coronary artery disease and microvascular health. JACC Cardiovasc Imaging 3:623–640. https://doi.org/10.1016/j.jcmg.2010.04.007 Renaud JM, DaSilva JN, Beanlands RSB, deKemp RA (2013) Characterizing the normal range of myocardial blood flow with 82rubidium and 13N-ammonia PET imaging. J Nucl Cardiol 20:578–591. https://doi.org/10.1007/s12350-013-9721-3 Cerqueira MD (2002) Standardized myocardial segmentation and nomenclature for tomographic imaging of the heart: a statement for healthcare professionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the American Heart Association. Circulation 105:539–542. https://doi.org/10.1161/ hc0402.102975 Dilsizian V, Bacharach SL, Beanlands RS et al (2016) ASNC imaging guidelines/SNMMI procedure standard for positron emission tomography (PET) nuclear cardiology procedures. J Nucl Cardiol 23:1187–1226. https://doi.org/10.1007/ s12350-016-0522-3 Byrne C, Jensen T, Hjortkjær H et al (2015) Myocardial perfusion at rest in patients with diabetes mellitus type 1 compared with

The International Journal of Cardiovascular Imaging

23.

24.

25.

26.

healthy controls assessed with multi detector computed tomography. Diabetes Res Clin Pract 107:15–22. https://doi.org/10.1016/j. diabres.2014.10.011 Wang L, Jerosch-Herold M, Jacobs DR et al (2006) Coronary artery calcification and myocardial perfusion in asymptomatic adults. J Am Coll Cardiol 48:1018–1026. https://doi. org/10.1016/j.jacc.2006.04.089 Djaberi R, Beishuizen ED, Pereira AM et al (2008) Non-invasive cardiac imaging techniques and vascular tools for the assessment of cardiovascular disease in type 2 diabetes mellitus. Diabetologia 51:1581–1593. https://doi.org/10.1007/s00125-008-1062-4 Range FT, Paul M, Schafers KP et al (2009) Myocardial perfusion in nonischemic dilated cardiomyopathy with and without atrial fibrillation. J Nucl Med 50:390–396. https://doi.org/10.2967/ jnumed.108.055665 Range FT, Schafers M, Acil T et al (2007) Impaired myocardial perfusion and perfusion reserve associated with increased

27.

28.

29. 30.

coronary resistance in persistent idiopathic atrial fibrillation. Eur Heart J 28:2223–2230. https://doi.org/10.1093/eurheartj/ehm246 Clark DM, Plumb VJ, Epstein AE, Kay GN (1997) Hemodynamic effects of an irregular sequence of ventricular cycle lengths during atrial fibrillation. J Am Coll Cardiol 30:1039–1045. https://doi. org/10.1016/S0735-1097(97)00254-4 Tona F, Serra R, Di Ascenzo L et al (2014) Systemic inflammation is related to coronary microvascular dysfunction in obese patients without obstructive coronary disease. Nutr Metab Cardiovasc Dis 24:447–453. https://doi.org/10.1016/j.numecd.2013.09.021 Oreopoulos A, Padwal R, Kalantar-Zadeh K et al (2008) Body mass index and mortality in heart failure: a meta-analysis. Am Heart J 156:13–22. https://doi.org/10.1016/j.ahj.2008.02.014 Ylä-Herttuala S, Bridges C, Katz MG, Korpisalo P (2017) Angiogenic gene therapy in cardiovascular diseases: dream or vision? Eur Heart J 38:1365–1371. https://doi.org/10.1093/eurheartj/ ehw547

13