A simple four-grade score for aortic arch calcification by posteroanterior chest X-ray is associated with cardiovascular disease in haemodialysis patients

Sir, We found the recent article by Noordzij et al. [1] extremely interesting read. The authors found that progression of aortic arch calcification (AoAC), which was associated with hyperparathyroidism, as assessed by lateral chest X-ray (CXR), was related to mortality from cardiovascular disease (CVD). While an annual lateral CXR may be recommended in other countries, such procedures are rarely implemented in daily dialysis practice in Japan. Instead, based on a manual published in the late 1990s, only posteroanterior CXR (PA-CXR) is routinely performed and interpreted by dialysis physicians to monitor cardiothoracic index for dry weight assessment [2]. As such, we feel that the medical community could benefit greatly from easily accomplished assessment of AoAC via routine PA-CXR. We conducted a cross-sectional study involving maintenance haemodialysis patients [N = 51; mean age: 67 ± 12 years; median dialysis vintage: 4.5 years (interquartile range: 1.9–8.9)] at an outpatient dialysis centre. Using PA-CXR, AoAC was graded as follows: Grade 0, no visible calcification; Grade 1, small spots of calcification or thin calcification on the aortic knob; Grade 2, one or more areas of thick calcification and Grade 3, circular calcification on the aortic knob [3]. CVD was considered present if the patient had a past history of ischaemic heart disease (diagnosed by coronary angiogram), brain infarction with neurological symptoms (confirmed by brain imaging) or peripheral arterial disease (diagnosed by angiography or ankle–brachial index < 0.90). We examined the prevalence of CVD by AoAC grade and determined the association between AoAC and presence of CVD using binomial regression. CVD prevalence by AoAC grade was 14% (N = 14), 42% (N = 19), 64% (N = 14) and 75% (N = 4) for Grades 0–3, respectively (P for trend < 0.01; Figure 1). Univariate analysis revealed that CVD was associated with AoAC Grades 2 and 3 [versus Grade 0, prevalence difference (PD): 0.50 and 0.61, respectively; Table 1]. After adjustment for age and diabetes as a cause of dialysis, CVD was found to be associated with AoAC Grades 1, 2 and 3 (versus Grade 0, PD: 0.25, 0.43 and 0.62, respectively). Factors associated with presence of CVD on binomial regressiona Binomial regression was performed by STATA version 11.0. The regression procedure in STATA fits generalized linear models estimating PD. Statistically significant values P < 0.05 in bold; all P < 0.20 in italics. Possible confounders with P < 0.20 in univariate analysis were entered into multivariate analysis. CI, confidence interval. Factors associated with presence of CVD on binomial regressiona Binomial regression was performed by STATA version 11.0. The regression procedure in STATA fits generalized linear models estimating PD. Statistically significant values P < 0.05 in bold; all P < 0.20 in italics. Possible confounders with P < 0.20 in univariate analysis were entered into multivariate analysis. CI, confidence interval. Prevalence of CVD by AoAC grade in haemodialysis patients (N = 51). The proportion of patients with total number of CVD presence (sum of ischaemic heart disease, brain infarction and peripheral arterial disease numbers) is shown for each grade (range: 0-3). Despite the relatively small population and limited covariates available in our analyses, our results suggest that AoAC grade as assessed by using PA-CXR may be associated with CVD in haemodialysis patients, and CVD prevalence may increase with progression of AoAC. This simple four-grade system has been shown to predict CVD events in the general population [4] and appears applicable in haemodialysis facilities across Japan. A large longitudinal study is needed to clarify whether or not this grading system based on PA-CXR is able to predict CVD outcomes as well as lateral CXR in dialysis patients. Conflict of interest statement. None declared. Editorial Note: Dr Noordzij et al. had no further comments on this letter.