The main results of our study is that the prevalence of HU in CR patients is 19.5% (with the classic cut-off) similar to what has been founded in in ACS (23%) [3] and CCS hospitalized patients (22.9%) [23]. As already mentioned, patients that are normally referred to CR are the one with multiple CV events, multiple cardiac and non-cardiac comorbidities and the one with uncontrolled risk factor and the need for therapies optimization [10]. Even if this may suggest the possibility to find a higher prevalence of HU, our study didn’t confirm this speculation.

The second point is that, when the newer and lower URRAH cut-off was used, HU prevalence significantly increases to 47.4% of the population. This new cut-off is of importance since UA begins to negatively impact the CV system at lower values than the classical one that is, conversely, principally related to UA crystal deposition, and so, to articular and kidney gout. As obvious, the lower we move the cut-off, the higher will be the prevalence of hyperuricemia (in our study raises from 19.5 to 47.4%). Nevertheless, this could be of interest if a clinical and a therapeutic benefit is determined by this re-classification. This matter is still under investigation as the last trial on hypouricemic agents (ALL-HEART study [24]) suggests that treating asymptomatic HU does not yield a substantial benefit on subsequent CV events in secondary prevention patients. As discussed elsewhere [25], this trial has strong limitation and other trials (evaluating also febuxostat and not only allopurinol, enrolling also recent ACS and not only CCS, selecting only UA overproduction patients) are needed before definitive conclusions can be drawn.

Another result of our study deserved to be mentioned, i.e. the fact that HU was not associated with functional improvement (Δ6MWT). Only one previous study evaluate this association [18] founding a direct correlation between elevated UA levels and a higher Δ6MWT in a population of patients who underwent CR after heart surgery. However, the correlation was found only in the subpopulation aged over 65 years. 47% of the subjects of this study underwent valve surgery with no coronary artery disease, so a completely different cohort in comparison to ours.

Conversely, 6 studies evaluated the relationship in the opposite direction, looking for the effects of cardiac rehabilitation on UA serum levels. Unfortunately, since we didn’t have a second UA value, we cannot add new information on this topic. However, all the study published were on very few patients (n = 26–90) showing heterogeneous results with 2 of them in favour of a decrease of UA after CR [14, 15] while 2 of them showed no significant difference [13, 17]. The only one with an appropriate number of patients (n = 693) showed the absence of a significant differences in UA before and after CR [12]. Finally, one study showed an increase in UA during CR [16]. However, none of the cited studies described baseline hypouricemic agents nor if this were started or dosage were increased during the CR.

Differences in these studies results are probably driven by three factors: (1) population (only ACS patients or also CCS patients); (2) time of CR initiation after the index events; (3) pro-oxidant and anti-oxidant properties of UA.

Regarding the latter it is well established that oxidoreductive impairment plays a crucial role in CV diseases. Whether UA stands as a scavenger mechanism or may worsen the balance is still disputed. The origin of this dualism lies in the final biochemical transformation of hypoxanthine in UA; this passage could be catalyzed by two different enzymes, xanthine-dehydrogenase (XDH) and xanthine-oxidase (XO). The balance between these two enzymes will define whether UA help to reduce the oxidative state or worsen it. Some authors suggest that an ischemic drive may facilitate the conversion from XDH to XO and therefore may enhance the production of reactive oxygen species [26].

Another point of our study deserved to be mentioned, i.e. the relationship between UA and EF. Our results are in line with the few studied already published, mostly focused on HF, in which a significant association was found [19,20,21, 27]. The association was confirmed also in CCS [20, 28] and ACS patients [3].

Regarding the mechanism that could link UA to EF, all the cited studies are cross-sectional doesn’t give us the opportunity to draw definite conclusion on the direction of the associations found. So, it could be possible that UA can determine an EF reduction but also that a reduced EF is able to determine an increase in UA. Data in favour of both hypotheses have been published.

UA could increase due to many conditions that are present when EF reduce such as an increased purine degradation determined by hypoxia and tissue catabolism [29], increase in xanthine-oxidase activity during HF [30, 31], diuretics use (common in the settings of reduced EF) and acute kidney failure that could complicate an ACS or an HF hospitalization.

On the contrary UA could determine a decrease in EF due to an increase local inflammation, angiotensin release and increase in oxidative stress. However, longitudinal data are needed in order to give a definitive direction to this strong and important association.

Coming back to the cut-off issue, the absence of a significant association between UA and EF with the URRAH cut-off, in the absence of longitudinal studies, raise the hypothesis that higher values are needed for UA in order to damage left ventricular systolic function or that, on the opposite direction, when EF reduce UA increase in a significant very manner.

Furthermore, our patients are in a late stage of atherosclerotic disease (already with at least one CV events) and it is also possible that UA is more able to damage endothelium and arteries in a first phase when they are still healthy. When the disease progress to a more advanced stage higher values are needed to further determine a damage because the effect of lower values can be overshadowed by the other CV risk factors (aging process, previous ACS or revascularization for a CCS and comorbidities).

The results of our study are influenced by some limitations, the first of which is the cross-sectional nature of the study. As already mentioned, because of this, the direction of the observed association could not be determined. Secondly, ours is a monocentric study and so generalizability could be influenced by local patients’ selection and procedures. Furthermore, the lack of a second uric acid dosage (at discharge) didn’t provide us any information on its changes determined by the CR. Finally, the use of allopurinol or febuxostat was not reported.