Professional Documents
Culture Documents
Relationship of Cell-Free Hemoglobin To Impaired Endothelial Nitric Oxide Bioavailability and Perfusion in Severe Falciparum Malaria
Relationship of Cell-Free Hemoglobin To Impaired Endothelial Nitric Oxide Bioavailability and Perfusion in Severe Falciparum Malaria
Background. Hemolysis causes anemia in falciparum malaria, but its contribution to microvascular pathology
in severe malaria (SM) is not well characterized. In other hemolytic diseases, release of cell-free hemoglobin causes
nitric oxide (NO) quenching, endothelial activation, and vascular complications. We examined the relationship of
plasma hemoglobin and myoglobin to endothelial dysfunction and disease severity in malaria.
Methods. Cell-free hemoglobin (a potent NO quencher), reactive hyperemia peripheral arterial tonometry
(RH-PAT) (a measure of endothelial NO bioavailability), and measures of perfusion and endothelial activation
were quantified in adults with moderately severe (n p 78 ) or severe (n p 49 ) malaria and control subjects
(n p 16) from Papua, Indonesia.
Results. Cell-free hemoglobin concentrations in patients with SM (median, 5.4 mmol/L; interquartile range
[IQR], 3.2–7.4 mmol/L) were significantly higher than in those with moderately severe malaria (2.6 mmol/L; IQR,
1.3–4.5 mmol/L) or controls (1.2 mmol/L; IQR, 0.9–2.4 mmol/L; P ! .001 ). Multivariable regression analysis revealed
that cell-free hemoglobin remained inversely associated with RH-PAT, and in patients with SM, there was a signifi-
cant longitudinal association between improvement in RH-PAT index and decreasing levels of cell-free hemoglobin
(P p .047). Cell-free hemoglobin levels were also independently associated with lactate, endothelial activation, and
proinflammatory cytokinemia.
Conclusions. Hemolysis in falciparum malaria results in NO quenching by cell-free hemoglobin, and may
exacerbate endothelial dysfunction, adhesion receptor expression and impaired tissue perfusion. Treatments that
increase NO bioavailability may have potential as adjunctive therapies in SM.
Hemolysis of infected and uninfected red blood cells is falciparum malaria is microvascular obstruction re-
an important cause of anemia in falciparum malaria sulting from cytoadherence of parasitized erythrocytes
[1], but its contribution to other pathophysiological to activated endothelial cells, associated with impaired
pathways in severe malaria (SM) is less well character-
ized. A central process in the pathogenesis of severe
Potential conflicts of interest: N.M.A., D.L.G., and J.B.W. are named as inventors
in a US patent for the use of L-arginine as treatment for severe malaria but have
transferred all their rights to their respective institutional malaria research
collaborations. This patent is issued for US rights only, and no rights are being
Received 23 March 2009; accepted 12 June 2009; electronically published 2 sought in other countries. All other authors report no other conflicting interests.
October 2009. Presented in part: Annual Scientific Meeting of the American Society of Tropical
Reprints or correspondence: Dr Anstey, International Health Division, Menzies Medicine and Hygiene, New Orleans, December 2008 (abstract 1194).
School of Health Research, PO Box 41096 Casuarina, Darwin, NT 0811, Australia Financial support: National Health and Medical Research Council (International
(anstey@menzies.edu.au). Collaborative Research Grant [ICRG] 283321 and practitioner fellowship to N.M.A.),
The Journal of Infectious Diseases 2009; 200:1522–9 Wellcome Trust (ICRG GR071614MA and career development award 074637 to
2009 by the Infectious Diseases Society of America. All rights reserved. R.N.P.), VA Research Service, National Institutes of Health (grants AI55982 and
0022-1899/2009/20010-0007$15.00 AI041764), and the Tudor Foundation.
a
DOI: 10.1086/644641 Present affiliation: Loyola University Medical Center, Maywood, Illinois.
measured by high-performance liquid chromatography (Shi- tralateral index finger without reactive hyperemia. Patients were
madzu), using methods described elsewhere [3]. assessed in a quiet, temperature-controlled environment, with
Endothelial function. Endothelial function was measured arms relaxed and supported by cushions, according to the man-
noninvasively by using peripheral arterial tonometry (Endo- ufacturer’s instructions [3, 19]. The RH-PAT index is at least
PAT) to determine the change in digital pulse wave amplitude 50% dependent on endothelial NO production [20]. Endothe-
in response to reactive hyperemia, giving a reactive hyperemia lial function was measured daily until death or discharge or
peripheral arterial tonometry (RH-PAT) index. With this tech- until the RH-PAT index was above an a priori cutoff value
nique, the arterial pulsatile volume of the index finger at rest (1.67) for 2 consecutive days [7].
is compared with that after an increase in shear stress induced Statistical methods. Statistical analysis was performed with
by 5 min of forearm ischemia (sphygmomanometer inflated to Stata software, version 9.2 (Stata). Intergroup differences were
200 mm Hg). Systemic influences were reduced by simulta- compared by analysis of variance or Kruskal-Wallis test, where
neously performing peripheral arterial tonometry in the con- appropriate. Pearson’s or Spearman’s correlation coefficients
NOTE. Data are means (95% confidence intervals), unless otherwise indicated. P ! .01 for all variables, by analysis of variance (overall)
or 2-sided t test. HRP2, histidine-rich protein 2; ICAM-1, intercellular adhesion molecule 1; IQR, interquartile range; ND, not determined;
RH-PAT, reactive hyperemia peripheral arterial tonometry.
NOTE. df, degrees of freedom; HRP2, histidine-rich protein 2; ICAM-1, intercellular adhesion molecule
1; IL, interleukin; RH-PAT, reactive hyperemia peripheral arterial tonometry; TNF, tumor necrosis factor.
a
Correlation with cell-free hemoglobin level.
b
TNF and IL-6 levels were measured only in patients with severe malaria.
nase and cell-free hemoglobin remained significantly inverse- SM, with median plasma concentrations of 2.4 pg/mL (IQR,
ly associated with RH-PAT both in all patients with malar- 1.5–4.2 pg/mL) for TNF and 84 pg/mL (IQR, 15–501 pg/mL)
ia (P p .02 and P ! .001, respectively) and in those with SM for IL-6. In patients with SM, cell-free hemoglobin level was
(P p .046 and P p .02, respectively). In a longitudinal mul- correlated significantly with both TNF level (r p 0.33; P p
tivariable, mixed-effects model in patients with SM who sur- .04) and IL-6 (r p 0.41; P p .01) (Table 3).
vived 124 h after enrollment, there were significant associa- Cell-free hemoglobin, LDH, and markers of organ dys-
tions between improvement in RH-PAT index and both de- function. The LDH levels were correlated significantly with
creasing cell-free hemoglobin (P p .047) and increasing l-ar- levels of cell-free hemoglobin (rs p 0.65; P ! .001), creatinine
ginine (P p .001) concentrations. (rs p 0.68; P ! .001), and creatine kinase (rs p 0.48; P ! .001).
Cell-free hemoglobin and endothelial activation. Relative These associations remained significant after stratification by
to patients with MSM, patients with SM had significantly higher disease severity.
plasma concentrations of soluble ICAM-1, E-selectin, and an- Cell-free hemoglobin, plasma arginase activity, and markers
giopoietin 2 (Table 2). Cell-free hemoglobin was correlated with of hepatic function. Plasma arginase activity levels were in-
all 3 parameters among all patients with malaria: ICAM-1 creased in patients with SM (P ! .001) (Table 3). Among all
(r p 0.4; P ! .001), E-selectin (r p 0.33; P ! .001), and angio- patients with malaria, plasma arginase activity was correlated
poietin 2 (r p 0.42; P ! .001) (Table 3). After adjustment for with levels of cell-free hemoglobin (r p 0.29 ; P p .005) and
plasma HRP2 and disease severity, cell-free hemoglobin re- alanine transaminase (r p 0.29; P p .01), suggesting potential
mained significantly associated with ICAM-1. contributions to circulating arginase from both erythrocytic
Cell-free hemoglobin and biomarkers of severity. Com- and hepatic sources.
pared with patients with MSM, those with SM had higher con-
centrations of lactate (P ! .001) and plasma HRP2 (P ! .001) DISCUSSION
(Table 2). Blood lactate was correlated with cell-free hemoglo-
bin, both in all patients with malaria (rs p 0.38 ; P ! .001) and The plasma cell-free hemoglobin concentration increases with
those with SM (rs p 0.29; P p .04) (Table 3). Although plasma malaria disease severity, and levels are associated with impaired
HRP2 was correlated with cell-free hemoglobin (r p 0.5; P ! endothelial NO bioavailability, endothelial activation, increased
.001), this association was not significant in the subgroup with parasite biomass, and impaired tissue perfusion, as measured
SM. In multivariable analysis including all patients with ma- by blood lactate concentrations. Quenching of NO by cell-free
laria and controlling for disease severity, lactate was associated hemoglobin probably contributes to impaired endothelial ho-
with plasma HRP2 (P ! .001) and cell-free hemoglobin (P p meostasis with microvascular dysfunction, increased adhesion
.05). In a longitudinal mixed-effects model, the fall in lactate receptor expression, increased microvascular sequestration of
during clinical recovery from SM was significantly associated parasitized erythrocytes, and tissue hypoxia. Endothelial dys-
with the decrease in cell-free hemoglobin concentration (r p function was independently associated with increases in plasma
⫺0.5; P ! .001). cell-free hemoglobin and plasma arginase activity. This suggests
TNF and IL-6 levels were measured only in patients with that, as in SCD, both of these consequences of hemolysis con-