Prognostic value of intravascular ultrasound early after heart transplantation

Previous heart transplantation (HTx) studies have established that an increase in maximal intimal thickness (MIT) ≥0.5 mm within year 1 post-HTx predicts long-term outcome.1–3 Selected intravascular (IVUS) images (3–7 frames) were utilized in these studies, but current IVUS recommendations4 advocate volumetric IVUS analysis of a longer coronary segment (40–50 frames).5,6 Also, percent atheroma volume (PAV) is emerging as a favoured cardiac allograft vasculopathy (CAV) parameter,5 but the prognostic value of PAV has not been compared with MIT in a large study. Hence, this study determined the relationship between volumetric IVUS evaluation and long-term HTx outcome. This study included de novo adult HTx recipients aged 18–70 years from four transplant centres (Oslo University Hospital Rikshospitalet, Norway; Stanford University School of Medicine, USA; Asan Medical Center, Seoul, Korea; and Sahlgrenska University Hospital, Sweden). The centres’ independent ethics committee approved the study. Demographic data were collected at the time of baseline IVUS. Mortality and major adverse cardiac event (MACE) data [cardiac death, re-HTx, myocardial infarction, heart failure (ejection fraction <45%), or coronary revascularization] was obtained from national registries/hospital records. Serial IVUS examination of the left anterior descending artery was performed at baseline (within 12 weeks post-HTx) and at year 1. Percent atheroma volume defined as ∑(external elastic membrane (EEM)area − Lumenarea)/∑(EEMarea) × 100% was calculated. Normalized total atheroma volume (TAV) defined as ∑((EEMarea − Lumenarea)/number of frames analysed) × median number of frames in the study cohort was the secondary endpoint. MedCalc® statistical software (version 19.8) was used for plotting receiver operating characteristic (ROC) curves for change in PAV and MIT and all-cause mortality. Pairwise comparison of the area under the curve (AUC) was conducted using the method of Delong et al. (1988). Youden index J (defined as: J = sensitivity + specificity − 1) was used to determine the optimal cut-off values for PAV, MIT, and TAV in relation to all-cause mortality. Time to event was assessed by Kaplan–Meier statistics. Multivariate Cox regression analysis identified independent risk factors for survival/MACE. All statistical tests were two-sided at a 5% significance level. This study included 243 patients. Mean follow-up time for all-cause mortality was 8.1 ± 3.1 years. There were 44 (18%) deaths, 2 (0.8%) cases requiring re-HTx, 35 (14%) cardiac deaths, 10 (4%) patients underwent percutaneous coronary intervention/coronary artery bypass grafting, and 32 (13%) cases of ejection fraction <45%. The median (interquartile range) number of IVUS frames analysed was 47.3 (39.1–50.1). Mean PAV at baseline, year 1, and increase at year 1 was 19.5 ± 8.0, 23.5 ± 9.6, and 4.0 ± 5.2%, respectively. Increase in mean PAV was significantly greater amongst non-survivors vs. survivors (5.3 ± 6.0% vs. 3.7 ± 5.0%; P = .04). Percent atheroma volume increase ≥7.0% (n = 50, 21%) was the optimal cut-off for all-cause mortality with a sensitivity of 38.6% and specificity of 83.4% [Youden index J = 0.22, AUC 0.61 (0.55–0.67), P < .01]. There was significantly higher mortality amongst patients with PAV increase ≥7.0% vs. <7.0% (34% vs. 14%, P = .01) (Figure 1). Similarly, MACE (predominantly fatal cardiac events) was significantly higher amongst patients with PAV increase ≥7.0% compared with <7.0% (40% vs. 21%, P = .04) (Figure 1). Top (left): volumetric intravascular ultrasound (IVUS) analysis of a long coronary segment performed according to consensus recommendations to assess the increase in percent atheroma volume (PAV). Top (right): regression analysis confirms that increase in PAV ≥7.0% at year 1 is an independent predictor of mortality, whereas maximal intimal thickness (MIT) increase ≥0.5 mm is not. Bottom: Kaplan–Meier analysis shows that PAV increase ≥7.0% at year 1 is associated with increased all-cause mortality (left), while MIT increase ≥0.5 mm is not (right). Mean MIT at baseline, year 1, and increase at year 1 was 0.62 ± 0.39, 0.89 ± 0.78, and 0.28 ± 0.54 mm, respectively. The greatest MIT increase was 0.50 ± 0.61 mm and 0.23 ± 0.51 mm amongst non-survivors and survivors, respectively (P = .04). Receiver operating characteristic found that matched MIT increase ≥0.5 mm (n = 56, 23%) yielded a sensitivity and specificity of 31.8% and 78.9%, respectively, for all-cause mortality [Youden index J = 0.11, AUC 0.55 (0.49–0.62), P = .16]. There were 14 (25%) and 30 (16%) deaths in patients with and without MIT increase ≥0.5 mm, respectively (P = .74) (Figure 1). Similarly, MACE was not significantly different in patients with and without MIT increase ≥0.5 mm (36% and 22%, respectively, P = .41). There was a non-significant elevated risk of death amongst patients with normalized TAV increase above vs. below 30.0 mm3 (27% vs. 15%, P = .06). There was no difference in MACE incidence according to normalized TAV increase (30% vs. 23%, P = .38). In univariate analysis, chronic renal disease was not associated with an increased risk of mortality (P = .98). Percent atheroma volume increase ≥7.0%, diabetes mellitus, and lack of mycophenolate mofetil (MMF) therapy were associated with all-cause mortality (P < .05). These three variables were included in a Cox regression model, together with age and gender. In this model, increase in PAV ≥7.0%, diabetes mellitus, and MMF therapy at baseline were found to be independent predictors of all-cause mortality with an adjusted hazard ratio (HR) of 2.60 [95% confidence interval (CI) 1.37–4.93], 2.42 (95% 1.25–4.72), and 0.31 (95% CI 0.14–0.67), respectively (all P < .01). Increase in PAV ≥7.0% was also an independent predictor of MACE [adjusted HR 1.99 (95% CI 1.15–3.45), P = .01]. Lack of everolimus/sirolimus therapy and donor male gender were independent risk factors for PAV increase ≥7.0% with an adjusted HR of 3.0 (1.3–7.2; P = .01) and 2.4 (1.1–5.5; P = .04), respectively. This is the first multi-centre study to date evaluating the prognostic value of volumetric IVUS early after HTx in relation to long-term outcome. Percent atheroma volume increase ≥7.0% from baseline to 1 year (absolute percentage units) was a strong predictor of long-term survival and MACE. The previously established threshold of MIT increase ≥0.5 mm was not associated with mortality or MACE in this IVUS study. The threshold of MIT increase ≥0.5 mm is based on previous landmark studies utilizing select cross-sectional images. However, current IVUS recommendations4 advise against such selected analysis. This study utilized the recommended method of volumetric IVUS analysis.5,6 The median number of matched images per patient was 47 (compared with 3–7 matching frames in previous landmark studies). Increase in PAV ≥ 7.0% was the only IVUS variable to independently predict both all-cause mortality and MACE. Lack of autopsy data is a potential limitation of this study. This study is larger than preceding ones and has used the consensus IVUS methodology, but the relatively small number of events and the discriminatory value of an AUC of 0.61 require further validation. In conclusion, early CAV development, as defined by PAV increase ≥7.0% at year 1, is an independent predictor of subsequent all-cause mortality and MACE. Increase in MIT ≥0.5 mm is often used in current CAV trials, but in this study, it was not associated with an adverse long-term outcome. All authors declare no conflict of interest for this contribution. Data will be made available upon request in adherence with transparency conventions in medical research and through reasonable requests to the corresponding author. All authors declare no funding for this contribution. The independent ethics committee for each centre approved the study protocol. The pre-registered clinical trial numbers are NCT01266148 (SCHEDULE), NCT01796379 (HITS), NCT01078363 (ACE Inhibition and Cardiac Allograft Vasculopathy study), and NCT01366404 (IRIS registry).