DISCUSSION
This study evaluated the association between MBFR, a measure of coronary microvascular function and LV diastolic function, measured via septal mitral annular E/e’ in a group of patients presenting with chest pain and unobstructed coronary arteries. Our study showed that there was no association found between MBFR and LV diastolic function in both univariate and multivariate regression analysis after taking traditional cardiovascular risk factors into account.
Contrary to the findings from Taqueti,4 we did not find any correlation between septal mitral annular E/e’ and MBFR. In their study, they included patients with a similar profile (patients presenting with chest pain and no evidence of flow limiting coronary artery disease (CAD) on PET) yet found that impaired coronary flow reserve (CFR) in these patients was independently associated with diastolic dysfunction.4 In addition, they also used a similar echocardiographic definition for LV diastolic dysfunction to our study (E/e’ septal>15) and demonstrated a significant direct relationship between CFR and septal e’ and an inverse relationship between CFR and E/e’ septal. Their findings were clinically important as those patients with evidence of both CMD and elevated septal E/e’ ratio had a five-fold increased risk for HFpEF hospitalisation after a median follow up of 4 years.4In our study, only 5 out of the 55 (9%) patients with CMD had evidence of raised LV filling pressures. It is not clear from the study by Taqueti how many patients with CMD also had echocardiographic evidence of elevated LV filling pressures (septal mitral annular E/e’ ratio ≥ 15), although the published median and quartiles for E/e’ were 13.0 (9.3-16.1) in patients with CMD vs 10.8 (8.7-12.8) in patients without CMD.4
A larger prospective multicenter study by Shah enrolled 202 patients with a confirmed diagnosis of HFpEF and measured the prevalence of CMD via left anterior descending artery Doppler flow signals at rest and during adenosine infusion.11 CMD was present in nearly 75% of patients and associated with abnormal longitudinal myocardial function assessed by TTE. However, the main limitation of their study was the lack of exclusion of macrovascular CAD which would have led to a lower measurement of CFR. 11,12
There are several potential reasons that might explain the discrepancy between the findings in our study and previous studies. Firstly, in the study by Taqueti, their patient cohort included a greater proportion of comorbidities known to affect both coronary microvascular function and myocardial stiffness.4 The difference between their study population and ours were the higher numbers of the patients who were elderly (median age of 66 vs 59.7), proportions female (64.7% vs 48%), hypertensive (75.6% vs 39%) and had diabetes mellitus (32.8% vs 11%).4 Additionally, it is possible that despite the presence of CMD, our patient cohort with its lower prevalence of co-morbidities, may be earlier on the spectrum of myocardial stiffness and not yet manifesting detectable diastolic dysfunction.11,12
Notably, approximately 25% of patients who met the criteria for HFpEF in the study by Shah did not have evidence of CMD.11This is interesting as it could be explained by the heterogenous nature of HFpEF driven by non-cardiac specific causes of fluid overload, and that besides endothelial dysfunction, there may be other factors causing HFpEF syndrome in these patients.11 This implies that the presence of CMD does not necessarily equate to the reduction of diastolic function, hence the absence of correlation between CMD and LV diastolic function in our study.
The main question of the causal mechanistic links of whether 1) CMD leads to LV diastolic dysfunction and subsequently HFpEF or 2) LV diastolic dysfunction observed in patients with HFpEF leads to CMD, remains unclear. As alluded to, it is possible that our patient cohort is at an early stage of CMD, prior to the development of diffuse myocardial fibrosis and subsequent increase in LV stiffness via recurrent micro-infarctions secondary to coronary microvascular ischaemia.13 This hypothesis is supported by the finding of an exacerbated diastolic dysfunction when there was a detectable troponin in Taqueti’s study.4,13 Limited by the retrospective nature of this study, it would be interesting to observe prospectively how many of our patients with CMD eventually developed evidence of LV diastolic dysfunction or HFpEF.
Previous studies have also postulated that CMD is a consequence of a systemic pro-inflammatory state which occurs in medical conditions such as diabetes or obesity.1,6,7,14 Higher levels of Hs-CRP have also been observed in patients with HFpEF than in patients with heart failure with reduced ejection fraction (HFrEF), supporting a link between inflammation and HFpEF.7 However, in our study, we did not find an association between reduced MBFR and raised Hs-CRP. This potentially is a result of a lower prevalence of medical co-morbidities in the patients recruited in our study.