Obstructive sleep apnea is associated with increased coronary plaque instability: an optical frequency domain imaging study

Obstructive sleep apnea (OSA) is associated with coronary artery disease (CAD) and with an increased risk for myocardial infarction, stroke or death due to cardiovascular disease. Optical frequency-domain imaging (OFDI) is a useful modality for evaluating the characteristics of atherosclerotic plaque. The purpose of the study was to use OFDI to investigate the association of OSA with coronary plaque characteristics in patients undergoing percutaneous coronary intervention (PCI). We retrospectively analyzed OFDI data for coronary artery plaques from 15 patients with OSA and 35 non–OSA patients treated between October 2015 and October 2018. Plaque morphology was evaluated for 70 lesions, including 21 from patients with OSA and 49 from non–OSA patients. Compared with the non–OSA group, patients with OSA had significantly higher prevalences of thinned cap fibroatheroma (TCFA) (67% vs. 35%, P = 0.014) and microchannels (86% vs. 55%, P = 0.014); a significantly higher mean lipid index (1392 ± 982 vs. 817 ± 699, P = 0.021), macrophage grade (8.4 ± 6.4 vs. 4.8 ± 4.5, P = 0.030), and maximum number of microchannels (1.5 ± 1.0 vs. 0.7 ± 0.7, P = 0.001); and a significantly lower mean minimum fibrous cap thickness (69.4 ± 28.7 vs. 96.1 ± 51.8 μm, P = 0.008). This OFDI analysis suggests that OSA is associated with unstable plaque characteristics in patients with CAD. More intensive medical management for stabilization of coronary atherosclerotic plaque is required in patients with OSA. Electronic supplementary material The online version of this article (10.1007/s00380-019-01363-8) contains supplementary material, which is available to authorized users.


Introduction
Obstructive sleep apnea (OSA) is a common disorder with an estimated prevalence of 10-50% in men and 3-24% in women [1][2][3][4]. OSA is an independent risk factor for coronary artery disease (CAD), and up to 70% of patients with CAD have undiagnosed OSA [5]. Several studies have shown that OSA is associated with an increased risk for cardiovascular diseases such as myocardial infarction, stroke or death from cardiovascular disease [6][7][8][9].
Computed tomography (CT) in patients with OSA has revealed larger coronary plaque burdens and a larger lipid core compared to patients without OSA [20,21]. IVUS has shown that patients with OSA have a larger coronary atherosclerotic plaque volume than patients without OSA [22]. OCT and OFDI are superior to these modalities due to their higher resolution in visualizing in-depth microstructure. In the current study, the association of OSA with features of plaque stability was examined using OFDI in patients undergoing percutaneous coronary intervention (PCI), as the first such study of this association.

Sample population
Data were analyzed retrospectively for 50 consecutive patients (37 men, 13 women) aged > 30 years who underwent OFDI-guided PCI at Hokkaido Cardiovascular Hospital between October 2015 and October 2018. For each participant, polysomnography (PSG) monitoring was performed within 6 months of OFDI-guided PCI. PSG monitoring was conducted in a standardized fashion. Overnight PSG was performed using an SAS-3200 (Nihon Kohden, Tokyo, Japan). A total of 21 lesions were examined in 15 patients with OSA based on an apnea hypopnea index (AHI) ≥ 15, and 49 lesions were examined in 35 patients without OSA (AHI < 15). The mean ages in the OSA and non-OSA groups were 71.8 ± 10.4 and 69.7 ± 8.9 years, respectively, with no significant difference between the groups. Patients presenting with left main CAD and cardiogenic shock were already excluded from 50 consecutive patients who underwent OFDI-guided PCI. Patients who accepted continuous positive airway pressure therapy (CPAP) as a treatment for OSA were also excluded. The study was approved by the Ethics Committee of Hokkaido Cardiovascular Hospital and was performed in compliance with the Declaration of Helsinki and ethical principles for medical research involving human subjects. All patients gave written informed consent.

Coronary angiography
Coronary angiograms were analyzed by offline quantitative coronary angiography (GE ver. 5.10.1, Pie Medical Imaging BV. Maastricht, The Netherlands). Reference diameter, minimum lumen diameter, diameter stenosis, and lesion length were measured.

OFDI method and analysis
An OFDI imaging catheter (FastView™, Terumo Corp., Tokyo, Japan) was advanced using a 0.014-inch guidewire with the help of a 6-or 7-Fr guiding catheter, and the imaging core was placed at a site distal to the lesion. Before PCI, OFDI was performed with continuous flush of contrast media at a rate of 4 mL/s, and the OFDI wire was pulled back at a rate of 20-40 mm/s. OFDI was generally performed without dilation by a balloon catheter, but the lesion was dilated with a small-sized balloon if the OFDI catheter could not pass through the lesion because of severe stenosis. For patients with acute coronary syndrome (ACS) without spontaneous recanalization, aspiration thrombectomy was performed prior to OFDI. The plaque morphology of culprit lesions was studied. Other lesions in the same epicardial artery with > 50% stenosis in angiography were also included in OFDI analysis. After identifying the most stenotic cross-section, the 5-mm proximal and distal crosssections (total length: 10 mm) were examined. Cross-sectional images were then analyzed every 1 mm and evaluated for the presence of plaque rupture, plaque erosion, luminal thrombus, and categorized into lipid-rich plaque, TCFA, or a lesion with macrophage infiltration, cholesterol crystals, microchannels or calcification. OFDI analysis was conducted by two independent investigators (T.K. and R.K.) who were blinded to the clinical course of each patient. When there was discordance between the investigators, a consensus reading was obtained.

Definitions of OFDI findings
OFDI analysis revealed the presence of plaque rupture, plaque erosion, intraluminal thrombus, lipid-rich plaque, TCFA, macrophage accumulation, microvessels, and cholesterol crystals in the plaque. Plaque rupture was defined as the presence of fibrous cap discontinuity leading to communication between necrotic tissue and the lumen (Fig. 1a) [23,24]. Plaque erosion was defined as a lesion without fibrous cap disruption and the presence of thrombus [25]. A thrombus was defined as a well-delineated mass with a high signal attached to the luminal surface or floating within the lumen [24]. Lipid-rich plaques were defined as lesions with a lipid arc > 180° (Fig. 1b) [24]. The lipid arc was measured within a lipid-rich plaque, and the maximum value was recorded (Fig. 1b). Lipid-core length was defined as the length of plaque with > 90° of lipid and was measured on the longitudinal view. The lipid index is the maximum lipid arc multiplied by the lipid-core length [26]. A cholesterol crystal was defined as a thin, linear region of high signal intensity within the lipid plaque, without backscattering ( Fig. 1c) [24]. TCFA was defined as a fibrous cap thickness (FCT) < 65 μm, where FCT is the minimum thickness of a signal-rich layer from the coronary artery lumen to the inner border of the underlying lipid in the culprit lesion (Fig. 1d) [24]. Macrophage accumulation was defined as increased signal intensity within the fibrous cap, accompanied by heterogeneous backward shadows (Fig. 1e), and macrophages were semiquantified based on axial and circumferential distribution, as follows: grade 0, no macrophages; grade 1, localized macrophage accumulation; grade 2, clustered accumulation in < 1 quadrant; grade 3, clustered accumulation in ≥ 1 and < 3 quadrants; and grade 4, clustered accumulation in ≥ 3 quadrants (Fig. 2) [24,27]. The range for the macrophage score was 0-40, based on summation of the 0-4 grades across all slices. Microchannels were defined as small vesicular or tubular structures with diameters of 50 to 300 μm within the intima (Fig. 1f) [24]. The number of microchannels was also counted at the cross-section with the highest number of microchannels [19]. Calcification was defined as well-delineated and low backscattered heterogeneous regions [24]. Spotty calcium The maximum lipid arc was measured within a lipid-rich plaque (yellow lines). c A cholesterol crystal was defined as a thin, linear region of high signal intensity within the lipid plaque, without backscat-tering. d Measurement of fibrous cap thickness. The thickness was measured at the thinnest point three times, and the average was taken (white arrows). e Macrophage accumulation was defined as increased signal intensity within the fibrous cap, accompanied by heterogeneous backward shadows (arrowheads). f Microchannels were defined as small vesicular or tubular structures with diameters of 50-300 μm within the intima b grade 1, localized accumulation; c grade 2, clustered accumulation in < 1 quadrant; d grade 3, clustered accumulation in ≥ 1 and < 3 quadrants; and e grade 4, clustered accumulation in ≥ 3 quadrants deposits were defined as those with length < 4 mm and maximal arc < 90°, and deposits not meeting these criteria were classified as large calcium deposits [28,29]. Microcalcification was defined as a small microcalcification with a maximal calcium angle < 22.5° and a maximal calcification length < 1 mm [30].

Statistical analysis
Continuous variables are shown as means ± standard deviation (SD) or median (interquartile range) and categorical variables as counts and percentages. The normality of distributions was assessed by the Kolmogorov-Smirnov test. Between-group differences were examined by Pearson chisquare or Fisher exact test for categorical variables and Student t test or Mann-Whitney U test for continuous variables, as appropriate. P < 0.05 was considered to be significant. All data were analyzed with SPSS 25.0 (IBM Corp., Armonk, NY).

Clinical characteristics
The clinical characteristics of the OSA and non-OSA groups are compared in Table 1. The mean AHI was significantly higher in the OSA group (30.9 ± 12.6 vs. 7.2 ± 3.8, P < 0.001). The mean LDL-C did not differ significantly between the groups (107 ± 29 vs. 104 ± 45 mg/dl, P = 0.732). All other characteristics, including medications and concomitant diseases, were similar in the two groups. All patients who had statin therapy prior to PCI were treated with strong statin (atorvastatin 10 mg/day, rosuvastatin 2.5-5 mg/day or pitavastatin 2 mg/day).

Angiographic findings
Plaque location and angiographic data are shown for the two groups in Table 2. The OSA group had a significantly higher stenosis diameter (87.7% ± 14.6% vs. 79.7% ± 15.4%, P = 0.044). There were no significant differences in plaque location or in other angiographic data between the two groups.

Risk factors for higher lipid index, TCFA, macrophage invasion, and microchannels
Multiple logistic regression analyses were performed to identify risk factors for higher lipid index, TCFA, macrophage invasion, and microchannels. Because history of PCI or CABG, history of myocardial infarction, prior aspirin use and prior clopidogrel use were correlated with prior statin use, they were excluded from the multiple variable analysis. We similarly excluded LDL-cholesterol (LDL-C) when we entered LDL-C to HDL-cholesterol (HDL-C) ratio in the multiple variable model. In multiple variable analyses, AHI, prior statin use, and glucose concentration were independently associated with a higher lipid index (Table 4); AHI and LDL-cholesterol (LDL-C) to HDL-cholesterol (HDL-C) ratio were associated with TCFA (Table 5); AHI and prior statin use were associated with greater macrophage grading (Table 6); and AHI, hemoglobin level and HDL-C were associated with greater microchannels (Table 7).

Observer variabilities
OFDI images were analyzed by two independent observers. The inter-observer reliabilities and intra-observer reproducibilities measured by the Pearson coefficient were r = 0.90 and 0.91 for lipid index, r = 0.92 and 0.94 for minimum FCT, r = 0.95 and 0.93 for macrophage grading, and r = 0.92 and 0.95 for maximum number of microchannels, respectively.

Discussion
The main findings in this study were (1) that patients with OSA had a larger lipid burden, thinner fibrous cap, greater macrophage accumulation, and more microchannels compared to those without OSA; (2) lipid index, minimum FCT, macrophage accumulation and microchannels were positively or inversely correlated with AHI; and (3) in patients undergoing PCI, AHI, prior statin use and glucose concentration were predictors of lipid index; AHI and LDL-C to HDL-C ratio were predictors of TCFA; AHI and prior statin use were predictors of macrophage grading; and AHI, hemoglobin and HDL-C were predictors of greater microchannels. To the best of our knowledge, this study is the first in depth comparison of coronary artery plaques in patients with and without OSA, with analysis of correlations of AHI with characteristics of unstable plaque using OFDI in patients who underwent PCI.
These observations improve understanding of the pathophysiology of OSA, and may have important implications for management of patients with OSA presenting with CAD.

Lipid-rich plaque
A large lipid core is an important contributor to plaque rupture through mechanically increasing the tension of the fibrous cap covering the lipid core, leading to disruption   [35]. In patients with OSA, intermittent hypoxia (IH) during sleep can increase oxidative stress, leading to oxidative modification of lipoproteins and other molecules [36][37][38]. These oxidized particles cause endothelial surface injury and promote accumulation of cholesterol in atherosclerotic plaque [39,40]. CT studies have shown larger coronary plaque burdens and a larger lipid core in non-culprit lesions of patients with OSA compared to those in patients without OSA [20,21], which is consistent with our data ( Table 3, Fig. 3A).

TCFA and FCT
Previous reports have shown that a thin fibrous cap is one of the most important features of unstable plaque in coronary and carotid artery [41][42][43]. Since matrix metalloproteinases (MMPs) released by macrophages induce thinning of fibrous caps of atherosclerotic plaques through collagen breakdown [44], more macrophage accumulation in the fibrous cap might lead to a thinner fibrous cap (Supplemental Fig. 2). Therefore, more macrophage accumulation in the plaque might contribute to a higher prevalence of TCFA and a lower FCT in patients with OSA (Table 3, Fig. 3b).

Macrophage accumulation
In a murine model of OSA, IH during sleep caused recruitment of more macrophages to the aortic wall [45,46], and IH during sleep for 6 weeks led to a significant increase in the percentage of macrophages in the aortic wall (6.4% ± 0.3% vs. 8.1% ± 0.3%, P = 0.003) [46]. Macrophages are thought to be extravasated from intraplaque microvessels into plaque tissue [47]. The number of microchannels in the plaque was weakly, but positively, correlated with the macrophage grade (Supplemental Fig. 1). Therefore, more intraplaque microvessels and direct invasion from the lumen of coronary arteries might contribute to greater macrophage accumulation in patients with OSA (Table 3, Fig. 3c).

Microchannels
Serum concentrations of vascular endothelial growth factor (VEGF) and endothelin-1, which have important roles in angiogenesis, are increased in patients with OSA [48,49].
Oxidative stress due to hypoxia in these patients induces endothelial dysfunction, leading to endothelial-derived microparticle (MP) formation [50]. MPs are small plasma  [49]. These mechanisms might explain the finding of more microchannels in patients with OSA and CAD (Table 3, Fig. 3d).

Unstable plaque in OSA and clinical perspectives
Several studies have shown that OSA might cause or accelerate atherosclerosis [53,54]. In this study, multiple regression analysis showed that AHI is independently correlated with features of unstable plaque ( Table 4, 5, 6, 7). These results support the hypothesis that OSA plays an important role in development of atherosclerotic plaque and plaque instability. However, in the Sleep Apnea Cardiovascular Endpoints (SAVE) trial, CPAP did not result in a lower incidence of the primary end point than usual care alone [55]. In this trial, the mean duration of CPAP use was only 3.3 h per night, which might not be sufficient for preventing cardiovascular events. After propensity-score matching, the incidence of a cerebrovascular event was significantly lower in 561 patients who used CPAP for more than 4 h per night, compared to that in the usual-care group (hazard ratio, 0.52, P = 0.02) [55]. Furthermore, there is evidence that treatment of OSA with CPAP improves endothelial function and reduces the risk for cardiovascular events [56], and a randomized controlled trial showed that CPAP therapy resulted in a significant reduction of carotid intima-media thickness [57]. Therefore, CPAP therapy and stringent management of other coronary risk factors might be effective for stabilization of unstable plaque and secondary prevention of adverse cardiovascular events in patients with OSA.

Limitations
The sample size of this retrospective, cross-sectional study conducted at a single medical center was small, and the results require confirmation in a prospective study including a larger number of patients. Second, we could not analyze microchannels of < 50 μm due to the limit of the device resolution. Third, there is an inherent discrepancy between characteristics assessed by OFDI and actual histopathological findings [58]. Further analyses using higher resolution OFDI might enable more detailed assessment of intraplaque microstructures in patients undergoing PCI. Fourth, we did not have information on inflammatory biomarkers such as high-sensitivity CRP, ICAM-1 and E-selectin. Fifth, the analysis of the risk factors was done on a per lesion basis although the patient characteristics were analyzed on a per patient basis.

Conclusion
This OFDI analysis suggests that OSA is associated with plaque instability in patients with CAD. More intensive medical management is required for patients with OSA for stabilization of coronary atherosclerotic plaques.