Reduced Cardiac Index Reserve and Hypovolemia in Severe Falciparum Malaria

Abstract Background Impaired microvascular perfusion is central to the development of coma and lactic acidosis in severe falciparum malaria. Refractory hypotension is rare on admission but develops frequently in fatal cases. We assessed cardiac function and volume status in severe falciparum malaria and its prognostic significance. Methods Patients with severe (N = 101) or acute uncomplicated falciparum malaria (N = 83) were recruited from 2 hospitals in India and Bangladesh, and healthy participants (N = 44) underwent echocardiography. Results Patients with severe malaria had 38% shorter left ventricular (LV) filling times and 25% shorter LV ejection times than healthy participants because of tachycardia; however, stroke volume, LV internal diameter in diastole (LVIDd), and LV internal diameter in systole (LVIDs) indices were similar. A low endocardial fraction shortening (eFS) was present in 17% (9 of 52) of severe malaria patients. Adjusting for preload and afterload, eFS was similar in health and severe malaria. Fatal cases had smaller baseline LVIDd and LVIDs indices, more collapsible inferior vena cavae (IVC), and higher heart rates than survivors. The LVIDs and IVC collapsibility were independent predictors for mortality, together with base excess and Glasgow Coma Scale. Conclusions Patients with severe malaria have rapid ejection of a normal stroke volume. Fatal cases had features of relative hypovolemia and reduced cardiac index reserve.

Severe falciparum malaria caused an estimated 435 000 deaths globally in 2017 [1]. Vital organ dysfunction develops in patients with high parasite biomass infections [2]. Reduced microcirculatory perfusion resulting from microvascular obstruction by sequestered parasites is central to the development of coma and lactic acidosis [3,4]. Although dehydration is common and hypovolemia may occur [5][6][7][8], large volume fluid resuscitation is not associated with resolution of acidosis [6] and may result in pulmonary edema [6]. Rapid administration of large fluid volumes is associated with secondary hypotension and increased mortality in children [9]. These findings suggest that microvascular obstruction and dysfunction (rather than hypovolemia or insufficient cardiac output) lead to inadequate tissue perfusion [4,10].
The causal chain of events resulting in death from severe malaria has multiple links. The factors that cause severe disease to develop (including sequestration of a high parasite biomass infection) are not necessarily equivalent to the factors affecting subsequent survival. Although hypovolemia does not cause coma or acidosis, macrovascular stability could determine survival after the start of treatment because refractory hypotension is a common terminal event in patients with severe malaria [11]. The cardiovascular system in malaria is hyperdynamic, with an increase in the ratio of oxygen consumption to delivery resulting from increased oxygen consumption and anemia [12].
There have been few studies of cardiac function in adult patients with severe malaria [13,14]. Post-mortem studies of the heart in adults and children dying from severe falciparum malaria have shown that sequestration of parasitized erythrocytes in the cardiac microvasculature is common [15,16].
We studied the relationships between disease severity, cardiac function, and volume status in malaria. We hypothesized that reduced cardiac index reserve resulting from hypovolemia and impaired systolic function are associated with increased mortality.

Patients
Consecutive consenting patients admitted to the adult medical wards during the malaria seasons at Chittagong Medical College Hospital (CMCH), Chittagong, Bangladesh in 2011-2014 and ISPAT General Hospital (IGH), Rourkela, India in 2011 with severe or uncomplicated falciparum malaria as defined previously [17] were enrolled into a pathophysiology study. There was no age cutoff, and adolescents admitted to the adult wards were also included. Patients with severe pre-existing cardiovascular disease were excluded from this analysis. Healthy subjects with no known acute or chronic illnesses were recruited locally. Further details can be found in the Supplementary Methods. Written informed consent was obtained from all participants or, for patients lacking capacity to provide consent, their representatives, or those who were <18 years old. The study was approved by the Oxford Tropical Research, Chittagong Medical College and IGH ethical committees. All patients had a medical history taken, full clinical examination, blood draw, and echocardiography performed on enrollment and were followed up at least once every 6 hours. Between 2011 and 2012, a focused echocardiographic exam measuring only cardiac index and inferior vena cava (IVC) diameters was conducted, whereas from 2013 to 2014 a detailed exam was completed. Full details of echocardiographic procedures are provided in the Supplementary Methods.

Statistics
Groups were compared using Mann-Whitney U tests (2 groups) or Kruskal-Wallis tests (more than 2 groups). Correlations were assessed using Spearman's rank. Linear regression was used to adjust indices of ventricular function for preload, afterload, and ejection time [18]. Within the group of patients with severe malaria, the prognostic significance of variables was assessed by calculation of the area under the receiver operator curve (AUROC). Multivariate logistic regression was used to assess whether echo variables that were significant univariate predictors of mortality were independent predictors of mortality alongside the established predictors; Glasgow Coma Score (GCS) and base excess. All analyses were conducted using Stata, version 15 (StataCorp, College Station, TX).

Baseline Characteristics and Recruitment
Data from detailed and focused echocardiographic examinations, respectively, were available from 52 and 49 patients with severe malaria, 30 and 53 patients with uncomplicated malaria, and 32 and 12 healthy participants, respectively. Of these patients, 5 with uncomplicated malaria and 3 with severe malaria had a history of cardiovascular disease (hypertension only). The healthy control group had larger body surface areas than the malaria groups (both P < .001) ( Table 1). The age range in the severe malaria group was from 10-to 70-year-olds with 8% (8

Heart Rate, Stroke Volume, and Cardiac Output
Despite the increases in heart rate with increasing malaria severity (P < .001 test for trend) ( Table 1), and the corresponding reductions in ejection time ( Figure 1A), stroke index values were similar in severe and uncomplicated malaria patients compared with healthy participants ( Figure 1B). Therefore, cardiac index ( Figure 1C) was increased in malaria in proportion to the rise in heart rate. Values of stroke work index, a measure of the external work done by the heart, were mildly reduced in malaria relative to healthy persons ( Figure 1D, Supplementary Table 1), whereas values of stroke power index, a measure of the rate at which work is being done by the heart, increased modestly in malaria as the shortening in ejection time outweighed the decrease in mean arterial pressure in malaria (Supplementary Table 1). Cardiac power index values were increased similarly by approximately 35% in severe malaria relative to health because of the higher heart rates (Supplementary Table 1). Fatal cases had higher heart rates (AUROC, 0.71; 95% confidence interval [CI], 0.6-0.81) ( Table 1) than survivors.

Cardiac Chamber Dimensions
Median chamber dimensions in patients with uncomplicated malaria were mostly within the normal range but were larger than in healthy participants ( Figure 2, Supplementary Table  2). Patients with severe malaria tended to have smaller cardiac dimensions than in uncomplicated malaria, similar to healthy participants ( Figure 2, Supplementary Table 2). The IVC inspiratory diameter was smaller in malaria, and IVC collapsibility, a predictor of fluid responsiveness, was greater ( Figure 3A and B). The IVC collapsibility did not correlate significantly with respiratory rate in severe malaria (rho = 0.17, N = 85, P = .170) and was not associated with dyspnea in the severe malaria group (P = .10).

Systolic Function
In severe malaria values for average systolic mitral annular tissue velocity (S') ( Figure 3C) and maximal LV outflow tract velocity (LVOT Vmax ) ( Figure 3D), both measures of peak LV contraction rate were increased, whereas stroke index values were similar relative to healthy participants ( Figure 1B). Increased peak ejection velocity was associated with a reduction in afterload in severe malaria, with average mitral annular S' correlating inversely with mediorotational end systolic stress (MRESS) (rho = −0.47, N = 51, P ≤ .001) and systemic vascular resistance index (SVRI) (rho = −0.42, N = 47, P = .004). Endocardial fractional shortening (eFS) ( Figure 4A) values were low (<25%) in 7% (2 of 30) of patients and high (>45%) in none of the patients with uncomplicated malaria, whereas in severe malaria, eFS values were low in 17% (9 of 52) of patients and high in 13% (7 of 52). In severe malaria, eFS was not associated with mortality when analyzed as a continuous variable (Supplementary Table 3 Left ventricular systolic performance was also assessed in relation to preload (LVIDd) ( Figure 5A) and afterload (MRESS) ( Figure 5B). Patients with severe malaria performed both above and below the prediction interval derived from the healthy participant data. In a linear regression model for eFS using LVIDd, MRESS, sex (a known determinant of eFS), and participant category (Supplementary Table 4), eFS was similar in severe malaria to healthy persons but lower in uncomplicated malaria than healthy persons. When stroke work index was considered ( Figure 5C), after adjusting for LV end diastolic volume index (LVEDVI) and sex, stroke work index was similar in uncomplicated and lower in severe malaria than in healthy persons (Supplementary Table 4). Neither stroke work index nor eFS take work rate into account, whereas stroke power does; when stroke power index was considered, after adjusting for LVEDVI and gender, stroke power index values were similar in uncomplicated malaria and higher in severe malaria than healthy participants (Supplementary Table 4).

Left Ventricular Diastolic Function
With the exceptions of the isovolemic relaxation time ([IVRT] a measure of the rate at which the ventricle relaxes) ( Figure  4B) and mitral valve ejection time (MVET) ( Figure 4C), which were both low in severe and uncomplicated malaria, variables related to diastole including E' (a measure of LV relaxation rate) ( Figure 4D) were relatively unchanged by malaria and did not predict mortality (Supplementary Table 5). A low E' (septal <7 or lateral <10 cm/s) was present in 3 of 28 (10%) healthy persons, 5 of 30 (17%) of patients with uncomplicated malaria, and 13 of 51 (25%) of patients with severe malaria (P = .23). The relationship between filling time and LVIDd is shown in Figure 5D. In severe malaria, LVIDd was correlated positively with average mitral peak early diastolic tissue velocity (E') (rho = 0.37, N = 51, P = .007), MVET (rho = 0.37, N = 51, P = .007), but not early mitral peak E-wave velocity (E), IVRT, or IVC diameter or collapsibility (all P > .05), suggesting that in the severe malaria group, filling time and possibly relaxation are the main determinants of LVIDd. The IVRT and average E' were inversely associated (rho = −0.6, N = 47, P < .001). The LVIDs was not associated with E, average E' , or IVRT (all P > .05).

Multivariate Prediction of Mortality
The predictors of mortality in the univariate analysis were combined with the previously established predictors of mortality, base excess, and GCS, to assess whether macrovascular parameters were independent prognostic indicators. The RA area index, relative wall thicknesses, and LVIDd index were not significant independent predictors (P > .05), whereast heart rate, IVC collapsibility, IVC inspiratory diameter, and LVIDs index were significantly associated with a fatal outcome independent of base excess and GCS (Supplementary Table 6).

DISCUSSION
Macrovascular cardiovascular parameters predicted outcome in this cohort of patients despite the fact that coma and lactic acidosis in severe falciparum malaria was not caused by inadequate cardiac output or total oxygen delivery. Differences in cardiac function parameters between severe and uncomplicated malaria were small (smaller atria, shorter ejection time, and higher heart rate in severe disease), but more substantial changes separated severe malaria survivors from fatal cases. Patients who died had higher heart rates, with smaller LV and RA diameters, increased IVC collapsibility, and high LV fractional shortening. Patients presenting with these cardiovascular profiles are expected to have lower cardiac index reserve and hence might be less able to compensate further peripheral vasodilation or other hemodynamic stresses. There was no evidence that systolic LV or RV dysfunction predicted outcome.
Previous studies of hemodynamics in severe malaria have lacked power to assess the association between macrovascular hemodynamic parameters and mortality. In severe malaria, no clinically significant acute change in LVIDd was noted in 2 studies in children [8,19] and 1 study in adults [13], and baseline eFS or ejection fraction was usually normal [8,10,13,14,19,20]. In children, insulin sensitivity (SI) decreased on recovery in one study in which there was a corresponding increase in hematocrit [19] and increased in another [8]. An increase in IVC collapsibility has been reported in children [8]. One study in children found evidence of increased tricuspid regurgitant jet velocity, suggesting that pulmonary hypertension may occur [20]. Impaired cardiac function in falciparum malaria is rare, presenting as myocarditis, reduced ejection fractional, or acute coronary syndrome in children [21] and adults [22][23][24][25].
In this study, malaria was associated with an increase in heart rate, peak ejection velocity, and S' , and a decrease in LVET and MVET, but preservation of LV end diastolic dimension, stroke volume, and eFS. There was no decrease in LV end systolic dimension, suggesting that although the rate of peak LV contraction is increased, this is counterbalanced by shorter ejection times, and therefore it does not increase stroke volumes. This more rapid ejection of a normal stroke volume could be caused by an increase in contractility due to sympathetic stimulation, circulating catecholamines, or the Treppe phenomenon (finding of contractility increasing with heart rate) [26]. Alternatively, because indices of ventricular performance such as eFS and S' may be dependent on preload and afterload, alterations in LV loading conditions could be responsible. Examining the indices of preload and afterload, LVIDd and MRESS were similar in severe malaria to health, but SVRI and systolic blood pressure were reduced in malaria. When eFS was used as the measure of performance, adjusting for preload and afterload by regression [18], it was lower in uncomplicated malaria but not severe malaria than in healthy participants. When stroke power was used, which takes into account the reduced ejection time in malaria, this was higher in severe malaria cases than in healthy cases. Thus, when duration of contraction, preload, and afterload are taken into consideration, ventricular performance appears moderately elevated in severe malaria relative to healthy people at rest.
Within the severe malaria group, tachycardia, which was associated with fever but not with anemia, inversely correlated with LV dimensions and stroke volume. Anemia was associated with a larger fractional shortening and stroke volume, likely maintained by a more rapid ventricular ejection (LVOT vmax and S') and filling (larger E). These findings are in keeping with studies in children with malaria and anemia, where cardiac index has been found to increase [19], predominantly because of an increase in stroke volume [10]. After accounting for preload and afterload, hematocrit was not an independent predictor of eFS, suggesting that anemia leads to increased stroke volume predominantly by affecting preload and/or afterload as opposed to being associated with increased contractility. Consistent with this, hematocrit inversely correlated with vascular resistance as noted previously (H. W. F. K., manuscript in preparation), and on multivariate analysis anemia was associated with a larger LVIDd.
Hyperlactatemia, a consistent feature of severe malaria, was associated with tachycardia, a reduction in LV but not IVC dimensions, and an increase in eFS and a preserved SI, consistent with sympathetic stimulation causing a hyperdynamic state. Fatal cases had features consistent with relative hypovolemia and reduced cardiac index reserve, ie, smaller cardiac chamber volumes, increased relative wall thicknesses despite no increases in mass (pseudohypertrophy [27]), and more collapsible IVC. Patients with less cardiac index reserve may be unable to maintain or increase their cardiac index in response to stresses and would therefore be more likely to develop shock. Previous studies have shown that mortality in severe malaria increases with age, which is known to correlate with a decrease in cardiovascular reserve [28]. In healthy individuals, markers of myocardial relaxation deteriorate with aging [29]. This is consistent with the observation in this study that myocardial relaxation (E' and IVRT) worsened significantly with age in the severe malaria group.
In bacterial sepsis, as in severe malaria, vascular resistance is typically reduced and the circulation is typically hyperdynamic (H. W. F. K., manuscript in preparation) [30]. Left ventricular systolic, diastolic, and RV function have been assessed in a wide range of ways in sepsis [31], although not typically in the context of preload and afterload. Where loading conditions have been considered, strong associations of contractility indices with afterload were observed in sepsis, confounding their interpretation in isolation [32]. Reduced indices of ventricular function have been reported more frequently in sepsis than we observed in severe malaria [33]. Patients described in studies of cardiac function in sepsis tend to be much older, with a mean age 50-60 [34] as opposed to median age 30 in this study, which may partially explain the differences in the frequency of abnormalities we observed. In addition, patients may have already received intensive care support including inotropic agents when the study began, which may have complicated assessment of contractility [35]. Left ventricular ejection fraction has not been consistently linked to mortality in sepsis, but smaller nonindexed LV dimensions have been linked in meta-analyses.
Using a noninvasive approach maximized recruitment into the study. However, this meant that chamber pressures and central venous oxygen saturations were not measured. In particular, in assessment of contractility and diastolic function, the assessment of pressure-volume loops would be informative. We have not assessed ventriculo-arterial coupling, a factor that can affect cardiac performance [36]. Adolescents were included alongside adults; severe malaria presents similarly in children over 10 years old as in adults [28]. Dimensions were adjusted for body size. Detailed studies were performed at only 1 of the study sites, and additional studies will be required to assess the generalizability of these findings in other settings. Inclusion of exercising or anemic healthy participants would be of value for controlling for the effects of heart rate and hematocrit.

CONCLUSIONS
In conclusion, our study shows that impaired systolic function is typically absent in malaria and does not predict mortality. Diastolic LV function in malaria patients is usually not impaired and varies with age. Stroke volume index is maintained in adult patients with malaria despite a reduction in LV filling and ejection time. Patients who died from severe malaria had evidence of hypovolemia at baseline and evidence of reduced cardiac index reserve. Future studies should investigate whether tailoring fluid therapy to optimize volume status in these patients is beneficial.

Supplementary Data
Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author. Notes