Haemodynamic response to COVID-19 and its outcome implications

Abstract

COVID-19, the syndrome caused by the severe acute respiratory syndrome coronavirus 2, has dominated the global medical landscape for the past 2 years. The clinical picture and outcomes have been fairly well characterized, and vaccination proven to prevent the development of infection and attenuate the severity of its manifestations. While the majority of individuals infected with COVID-19 will experience only a mild, influenza-like illness and recover uneventfully, the requirement for hospitalization portends a much worse outcome. Hospitalized COVID-19 patients with (pre-existing) cardiac disease have a high rate of thrombo-embolism, septic shock, and death. Despite large numbers of individuals having been hospitalized and their cardiovascular systems supported artificially, the haemodynamic profile of the disease has not been adequately characterized. Furthermore, the prognostic implications of various haemodynamic patient subsets have not been comprehensively described. In the current issue of the journal, Taieb et al. investigated the non-invasive haemodynamic profile of 531 hospitalized patients with COVID-19. Reassuringly, 44% of these patients had normal left ventricular and right ventricular haemodynamics. Bilateral ventricular function was stratified into four groups: (i) normal filling pressure and normal output, (ii) normal filling pressure and low output, (iii) high filling pressure and normal output, and (iv) high filling pressure with low output. While abnormal right ventricular haemodynamics correlated with clinical status, it was not associated with worse outcome. Perturbation of left ventricular haemodynamics (stroke volume index, E/e0, and stroke work index), however, was associated with worse outcome. The non-invasively defined haemodynamic profile (four abovementioned groups) of COVID-19 patients was independently associated with mortality and demonstrated incremental value beyond a well-recognized clinical risk score [the Modified Early Warning Score (MEWS)]. Few data have been published on the non-invasive haemodynamic profile of COVID-19 patients. In a study of 23 individuals infected with COVID-19, non-invasively measured haemodynamics revealed a high cardiac index and a low systemic vascular resistance index, similar to what would typically characterize a hyperdynamic state. A variety of underlying mechanisms were proposed, including hypoxaemia, viral sepsis, and peripheral shunting. The near-normal PaO2 found in the majority of patients in this study suggests that hypoxaemia was not the primary driver for the hyperdynamic circulatory response. Evidence for hepatic arteriovenous shunting is indirect and originates from portal vein dilatation seen in autopsy studies. Viral sepsis secondary to COVID-19 infection with direct vasodilatory effects is an attractive theory to explain most of the haemodynamic observations, but unfortunately is not a well-defined entity. Interestingly, despite pulmonary involvement being the common denominator of COVID-19 infection, in the series described by Busana et al., pulmonary pressures and the total pulmonary resistance were not elevated. This unexpected observation may have been caused by pulmonary capillary recruitment or intrapulmonary neoangiogenesis. In a series of 21 mechanically ventilated COVID-19 patients, invasive measurements revealed a haemodynamic profile similar to the patients described by Busana et al., although comparison with the current study is made challenging due to the very different risk profile of the study population. The haemodynamic profile of COVID-19 patients in the study of Busana et al. contrasts sharply with a large number of those in the current analysis, where the indexed stroke volume was impaired in 46% of patients, high systemic vascular resistance was observed in 61% and pulmonary hypertension in 76%. These conflicting data may be reconciled by the fact that Busana et al. described the haemodynamic profile of COVID-19 early during the course of the disease (mean of 8.9± 6.6 days after symptom onset), while time to symptom onset is not mentioned by Taieb et al. Patients included in the current analysis may have been more severely ill: 44% required supplemental oxygen administration and/or non-invasive ventilation, while mechanical cardiorespiratory support or the institution of