This study was a retrospectively planned observational study for non-culprit coronary lesions in patients with CKD who underwent IVUS-guided PCI.

Figure 1 shows a breakdown of patients included in this study. Between April 2010 and March 2014, we had 543 consecutive PCI cases. IVUS-guided PCI was performed for 476 lesions (466 patients) in this period. After 8 months of standard medical treatment including statin, follow-up coronary angiography and IVUS examination were performed for 410 lesions (401 patients), irrespective of the presence or absence of symptoms. After excluding patients who had undergone follow-up examination using other modalities (25 patients with 25 lesions), serial IVUS images for 385 lesions (376 patients) obtained at both baseline and follow-up using identical IVUS catheters (ViewIT™, Terumo, Tokyo, Japan, or Atlantis™ SR Pro 2, Boston Scientific, Natick, MA, USA) were available for analysis. A target NCL in this study was defined as 5-mm segment which was 3 mm distal from ostium of each coronary artery (LAD, LCX, RCA), with percent diameter stenosis <50 % on quantitative coronary angiography (QCA). And also, a target NCL must be at least 10-mm proximal to the stented segment to remove stent edge effects (Fig. 2). Therefore, additional 187 lesions (182 patients) were excluded because we could not identify NCLs. After excluding 36 patients (38 lesions) with acute myocardial infarction, 12 patients (12 lesions) with left main trunk lesions, and 33 patients (35 lesions) with poor IVUS image quality, we enrolled 113 patients (113 NCLs) in this study. The patients were classified into 4 groups based on CKD stage: CKD-1: eGFR ≧90 ml/min/1.73 m²; CKD-2: eGFR <90 and ≧60 ml/min/1.73 m²; CKD-3: eGFR <60 and ≧30 ml/min/1.73 m²; and CKD4–5: eGFR <30 ml/min/1.73 m² with/without hemodialysis. The eGFR was calculated using the Diet in Renal Disease equation modified with the Japanese coefficient: eGFR (ml/min/1.73 m²) = 194 × serum creatinine (mg/dl)^−1.094 × age (years)^−0.287 (× 0.739 for female patients).

Standard medical treatment including statin was performed during the 8-month follow-up. Medication was left to the discretion of the treating physician. Dyslipidemia was controlled to the levels of low-density lipoprotein (LDL) cholesterol <100 mg/dl and high-density lipoprotein (HDL) cholesterol >40 mg/dl. All patients were prescribed statins (rosuvastatin, atorvastatin, pitavastatin or pravastatin) to manage the lipid profile to the target control level, unless contraindicated. Diabetes mellitus was controlled to HbA1c level <7.0 %. Hypertension was controlled to the systolic blood pressure (SBP) <130 mmHg and diastolic blood pressure (DBP) <80 mmHg. A study protocol was approved by the appropriate institutional review committee, and informed consent was obtained from all patients.

After the administration of 1–2 mg of intracoronary isosorbide dinitrate, IVUS images were obtained from the distal site of the culprit lesion using a 40-MHz IVUS catheter (ViewIT™ or Atlantis SR Pro 2™) attached to an IVUS imaging system (VISIWAVE™ or iLab™) with a motorized pullback at 0.5 mm/s. IB-IVUS images were obtained using software supplied with 1 of 2 IVUS machines (VISIATRAS™, Terumo, Japan).

Eight months later, a follow-up IVUS examination was repeated in the same coronary segment imaged at the baseline examination, regardless of clinical symptoms. Corresponding images were identified by the distance from the fiduciary side branches.

Each conventional IVUS and IB-IVUS parameter were measured at baseline and after 8 months. One experienced investigator who was unaware of the patient group allocation performed the quantitative IVUS analysis.

IVUS analyses were performed according to the criteria described in the American College of Cardiology Clinical Expert Consensus document on IVUS [13]. Cross-sectional lumen area, cross-sectional vessel area within the external elastic membrane, and plaque area (external elastic membrane area minus lumen area) were determined using software attached to each IVUS machine (VISIATRAS or EchoPlaque™, Indec System, Mountain View, CA, USA).

The definition of IB parameters for each histological category were determined by comparing histologic images reported in the previous study [14]. Plaque properties were diagnosed into 4 types according combining spectral parameters of posterior scattering signal: lipid; fibrous; dense-fibrous; or calcification. The area of each component was automatically measured in each plaque.

The volume of each conventional IVUS and IB-IVUS parameters for 5-mm length in the NCL was calculated using integration. A representative case is shown in Fig. 2.

A total of 113 NCLs were measured in 113 patients without significant stenosis. NCLs were categorized as follows: CKD-1, 18 lesions; CKD-2, 42 lesions, CKD-3, 29 lesions, CKD4–5, 24 lesions (Fig. 1).

Table 1 shows baseline clinical characteristics. Age in CKD-3, unstable angina pectoris (UAP) and DBP in CKD-1, and diabetic patients in CKD4–5 were higher than the other groups. Otherwise, there were no significant differences in other parameters among the groups.

The anti-atherosclerotic medications use and the degree of risk factor control at follow-up are summarized in Table 2. Statin treatment was prescribed for more than 80 % of each group. The patients with statin treatment initiated before PCI (continue statin group: CS) included 22 % in CKD-1, 33 % in CKD-2, 34 % in CKD-3, and 46 % in CKD4–5. The patients with statin treatment initiated after PCI (initiate statin group: IS) included 61 % in CKD-1, 62 % in CKD-2, 55 % in CKD-3, and 42 % in CKD4–5. The frequencies of CS and IS were comparable among 4 groups. Beta-blockers were more likely to be used in CKD-3 and CKD4–5. CKD4–5 had significantly higher HbA1c levels at follow-up, and CKD-3 and CKD4–5 had significantly higher SBP than CKD-1 and CKD-2. The achievement ratio of the HbA1c target level in the CKD4–5, and the target SBP in CKD-3 and CKD4–5 were significantly lower than those in the other groups.

Table 3 shows the volume data using conventional gray-scale IVUS. The EEM volume was significantly greater in CKD4–5 than the other groups at both baseline and follow-up. The lumen volume showed no significant differences among groups at baseline; however, at follow-up, it was significantly greater in CKD-1. The plaque volume was significantly greater in CKD4–5 than the other stages at both baseline and follow-up. The EEM volume significantly increased in CKD4–5, whereas it did not in the other groups. The lumen volume significantly decreased in CKD-3 and CKD4–5, but increased in CKD-1. The plaque volume significantly increased in CKD-3 and CKD4–5, whereas it significantly decreased in CKD-1 and CKD-2. Furthermore, in the ANCOVA models including age, gender, BMI, unstable angina pectoris, hypertension, hyperlipidemia, diabetes, follow-up LDL-cholesterol and HDL-cholesterol levels, ACE-I/ARB use, beta-blocker use, and statin use, did not change the results, reflected by still greater plaque progression in moderate to advanced CKD patients (Fig. 3). Univariate linear regression analysis showed the baseline eGFR correlated significantly with plaque volume change (R = 0.65, p < 0.0001) (Fig. 4). On univariate and multivariate logistic regression analyses, after considering CKD stage, confounding factors (age, gender, and unstable angina pectoris), coronary risk factors and medications, CKD stage 3–5 demonstrated the strongest association with plaque progression (Tables 4, 5).

Serial IB-IVUS images were acquired from 93 patients with 93 NCLs: CKD-1, 15 lesions; CKD-2, 33 lesions; CKD-3, 23 lesions; CKD4–5, 22 lesions. Baseline clinical characteristics and risk factor control were similar to gray-scale IVUS analysis.

In IB-IVUS analyses, although percent calcification was larger in CKD4–5 than the other groups at baseline, there were no significant differences in percent of each component (relative values) among groups at follow-up. However, the volumes of calcification, dense-fibrosis, and lipid were significantly greater in CKD4–5 than the other groups at both baseline and follow-up. Although fibrous volume showed no significant difference among groups at baseline, it was greater in CKD4–5 than the other groups at follow-up (Table 6).

The calcification and dense-fibrosis showed no significant changes in all groups during 8 months of follow-up. Fibrous plaque decreased in CKD-1, but increased in CKD4–5. Lipid volume significantly increased in CKD-3 and CKD4–5 during follow-up, although it decreased in CKD-2 and CKD-3 (Fig. 5). Representative cases are shown in Fig. 6.

CKD is well known to be associated with high mortality resulting from coronary artery disease (CAD) compared to mortality from progression to end-stage renal dysfunction. Many previous studies showed that CKD patients have severe atherosclerosis, and that they are at high risk for CAD [15‐17]. A previous study that included VH-IVUS analysis demonstrated that plaque volumes in NCLs were comparable among patients with varying levels of renal function [10, 18]. On the other hand, it was also reported that plaque volumes in patients with moderate CKD were greater than those in patients with normal renal function [11]. However, those studies analyzed only 1 timeframe. In one serial IVUS analysis, Nozue et al. [19] found that patients with normal-to-mild renal dysfunction under strong statin treatment showed significant decreases in EEM volume and plaque regression. Little is known about the impact of advanced CKD on serial change in plaque volume. In the present study, plaque progression occurred in patients with moderate to advanced CKD (CKD stage 3–5), whereas plaque regression occurred in CKD-1 or CKD-2. In this study, IS included 40–50 % of all patient, and they showed approximately 20 % reduction in LDL-cholesterol level. In CKD1–2, plaque regressions in IS were significantly greater than those in CS (6.5 ± 1.7 mm³ in IS vs. 2.2 ± 0.8 mm³ in CS, p = 0.01). Therefore, it is suggested that plaque regression in CKD-1 and CKD-2 was mainly due to initiate statin treatment after PCI. Previous study reported that plaque regression after newly initiated statin therapy in patients with normal to mild renal dysfunction was 1.3 (±9.1) % [19]. In the present study, plaque regressions of IS in CKD1–2 were greater compared with the previous study. It is suggested that the difference may be due to the patient characteristics. Our study included patients with mildly impaired renal function as well as those with advanced CKD, our results may suggest that plaque regression during standard medical treatment gradually decelerated with decreasing renal function. A previous randomized study failed to demonstrate clinical benefit of the rosuvastatin in patients with hemodialysis [20]. Results of the present study may support the results. An ANCOVA model and multivariate logistic model considering confounding factors, coronary risk factors, and medications, demonstrated that CKD stage have the strong association with plaque volume change. As for the multivariate logistic regression analysis, we integrated CKD-3 with CKD4–5 to perform plausible multivariate analysis. In this analysis, moderate to advanced CKD was definitely a strong independent predictor of plaque progression. Strict risk management for plaque regression is necessary, especially in patients with moderate to advanced CKD after PCI. In addition, we did not observe any clinical events associated with NCLs during this study period. Longer follow-up is needed to confirm the relationship between cardiovascular events and CKD stages.

A previous postmortem study showed that a lower eGFR is associated with a high percentage of advanced coronary atherosclerosis, defined as American Heart Association type IV (atheroma), type V (fibro-atheroma), and type VI (complicated plaque) [21]. Moreover, Miyagi et al. [8] found renal dysfunction to be significantly associated with an increased percentage of lipid volume and a decreased percentage of fibrous volume in the non-target coronary lesions of 89 patients using IB-IVUS. In addition, Hayano et al. [11] demonstrated greater lipid and smaller fibrous volumes in NCLs in a series of 201 patients using IB-IVUS. In the present study, we identified several differences in atherosclerotic plaque composition among patients with various stages of CKD. Lipid plaque regression occurred in CKD-1 and CKD-2, whereas lipid plaque gain occurred in CKD-3 and CKD4–5. Fibrous plaque also increased in CKD4–5. We speculate that the reason for these findings is the inclusion of patients who were on hemodialysis. A previous autopsy study showed that coronary plaques in hemodialysis patients are characterized by increased media thickness [17]. This finding might reflect fibrous plaque gain in hemodialysis patients during the follow-up period. The relationship between coronary plaque characteristics and mineral disorder in patients with end-stage CKD should be analyzed. Large lipid cores are considered histologic markers for plaque vulnerability, which is directly related to the risk of plaque rupture [22, 23]. Therefore, large lipid cores are considered one of the reasons for the increased risk of cardiovascular events in patients with CKD.