Platelet profi le is associated with clinical complications in patients with vivax and falciparum malaria in Colombia

Introduction: Thrombocytopenia is a common complication in malaria patients. The relationship between abnormal platelet profi le and clinical status in malaria patients is unclear. In low and unstable endemic regions where vivax malaria predominates, the hematologic profi les of malaria patients and their clinical utility are poorly understood. The aim of this study was to characterize the thrombograms of malaria patients from Colombia, where Plasmodium vivax infection is common, and to explore the relationship between thrombograms and clinical status. Methods: Eight hundred sixty-two malaria patients were enrolled, including 533 (61.8%) patients infected with Plasmodium falciparum, 311 (36.1%) patients infected with Plasmodium vivax and 18 (2.1%) patients with mixed infections. Results: The most frequently observed changes were low platelet count (PC) and high platelet distribution width (PDW), which were observed in 65% of patients; thrombocytopenia with <50,000 platelets/μL was identifi ed in 11% of patients. Patients with complications had lower PC and plateletcrit (PT) and higher PDW values. A higher risk of thrombocytopenia was identifi ed in patients with severe anemia, neurologic complications, pulmonary complications, liver dysfunction, renal impairment and severe hypoglycemia. The presence of thrombocytopenia (<150,000 platelets/μL) was associated with a higher probability of liver dysfunction. Conclusions: Young age, longer duration of illness and higher parasitemia are associated with severe thrombocytopenia. Our study showed that thrombocytopenia is related to malaria complications, especially liver dysfunction. High PDW in patients with severe malaria may explain the mechanisms of thrombocytopenia that is common in this group of patients.

Malaria is a disease caused by infection with Plasmodium parasites 1 and is a major health problem worldwide.According to the World Malaria Report, 219 million malaria cases and 660,000 malaria-related deaths occurred worldwide in 2010 2 .
Plasmodium falciparum is the main cause of malaria-related death worldwide; however, other species can also cause serious illness 3,4 .Clinical complications in malaria patients include cerebral malaria, severe anemia (SA), acute kidney failure (AKF), pulmonary edema (PE), severe hypoglycemia, shock, disseminated intravascular coagulation (DIC), acidosis and massive hemolysis.A patient who presents with one or more of these conditions is diagnosed with severe malaria (SM) and has an increased risk of mortality 5 .
Different organs can be affected during a malaria episode, which results in localized or systemic injury.Hematological changes, especially anemia and thrombocytopenia, are common 6,7 .Thrombocytopenia has been associated with Plasmodium knowlesi 8 , Plasmodium falciparum 9 , Plasmodium vivax, Plasmodium ovale 10 and Plasmodium malariae infections 11 .The evidence is mixed regarding the existence of a correlation between platelet count (PC) and parasitemia [8][9][10][12][13][14] . In a sstematic review, Lacerda et al. found no differences in the frequency of thrombocytopenia across species of Plasmodium; the frequency of thrombocytopenia ranged from 24-94% in all studies 6 .The frequency of thrombocytopenia ranged from 8.0-64.2% in Colombian patients [15][16][17] .
The 2013 World Health Organization (WHO) criteria that define complicated malaria do not include severe thrombocytopenia 33 .Studies have postulated that malaria should be suspected in travelers who present with fever and thrombocytopenia 34 .In malaria patients, thrombocytopenia is generally mild or moderate 13,17 , and bleeding is rare 35,36 .One study found that severe thrombocytopenia in malaria was an independent predictor of death (odds ratio (OR)=13.3,95%CI=3.2-55.1) 37, but this finding has been contested 9 .

METHODS
Thrombocytopenia is more frequent in pregnant women with malaria 38 and is especially pronounced in primigravidae.Additionally, there is an inverse relationship between parasite density and PC 39 .The relationship between thrombocytopenia and clinical presentation is not suffi ciently understood, and platelet changes beyond the PC have been either entirely unstudied or not studied in relation to other aspects of the clinical presentation.Moreover, in low and unstable endemic regions such as Colombia, the hematologic profi les of malaria patients and their clinical utility remain poorly understood.The aim of this descriptive study was to characterize the thrombograms of malaria patients from different regions of Colombia and to explore the relationship between clinical status and platelet alterations in ambulatory and hospitalized patients.

Patients and sociodemographic characteristics
A retrospective analysis was conducted using clinical and laboratory data obtained from patients from malaria-endemic regions in Colombia.Patients were treated on an outpatient basis for P. falciparum and P. vivax and were enrolled in clinical and epidemiological studies conducted by the Grupo Malaria (Universidad de Antioquia) between 1997 and 2007 [40][41][42] .A second study group consisted of malaria patients hospitalized between 2005 and 2010 (A.Tobón: unpublished data).The reference population consisted of patients from all age groups who presented at fi rst-or second-tier health-care facilities in the municipalities of Turbo and Necoclí (Urabá Region, Antioquia State), El Bagre (Bajo Cauca Region) and Tumaco, Guapi and Timbiquí (Pacifi c coast), as well as individuals who were treated on an inpatient basis at third-tier hospitals in Medellin with malaria diagnosed microscopically by thick smear (Table 1).For all patients with a complete blood count (CBC), a convenience sample was taken.Eight hundred sixtytwo patients were sampled.

Clinical and laboratory procedures
The thick smear was performed according to the WHO recommendations 43 .Following a medical assessment, a sample of blood was collected in ethylenediaminetetraacetic acid (EDTA) for cellular and biochemical studies, which included a CBC within 2 hours after sampling.The analysis was conducted in third-and fourth-generation automated analyzers (Cell-Dyn 3200, from Abbott Diagnostics®, Canada; Coulter HmX from Beckman Coulter®, United States of America; Celltac F MEK 8222, from Nihon Kohden®, United States of America and Sysmex KX-21N®, from Sysmex America, Inc.United States of America).Reference values for the platelet variables for the Colombian population were as follows: PC 150-400 x 10 3 platelet/µL, mean platelet volume (MPV) 6.5-13.5 fl , platelet distribution width (PDW) 15.4-16.8%and plateletcrit (PT) 0.085-0.287% 44.

Statistical analysis
The Kolmogorov-Smirnov test was used to test quantitative variables for normality; the median values were compared using the Mann-Whitney U test when the variables were not normally distributed.Associations between the qualitative variables were analyzed with the Chi-square test.Spearman's correlation coeffi cient was applied to analyze specifi c relationships.The level of statistical significance adopted was 5% (p<0.05).All analyses were conducted using the IBM® SPSS® statistics program, version 19 (licensed for Universidad de Antioquia).

Ethical considerations
All patients provided their informed consent to enroll in this study as required by the ethics committee of the Medical Research Institute of the Medical School of the Universidad de Antioquia (Acta 31, July 2002 and Acta 05, June 2005).The review of patient medical records was approved by the hospital ethics committees from Pablo Tobón Uribe Hospital and San Vicente Foundation Hospital (Letter 5282, Oct. 2009).

Platelet profi le
The values of PC, PT, MPV and PDW were not normally distributed (Kolmogorov-Smirnov, p=<0.001).Categorical sorting indicated that the most signifi cant alterations were the low PC and high PDW, which were present in approximately 65% of patients (Table 1).
The analysis by species is presented in Table 2.The PT, MPV and PDW were signifi cantly lower in the P. vivax cases compared with the P. falciparum cases (Mann-Whitney U test, p<0.05).PC was lower in the P. vivax cases, but this result was not signifi cantly different from that in the P. falciparum cases (p=0.832).The frequency of thrombocytopenia (<150,000 platelets) in the patients with P. falciparum was 66.6% compared with 63.7% in the P. vivax cases.Severe or profound thrombocytopenia (<50,000 platelets) was observed in 55 (10.3%) of the 533 P. falciparum cases, 34 (10.9%) of the 311 P. vivax cases and 4 (22.2%) of the 18 mixed infections (Chi 2 , p>0.05).
The plateletcrit was higher in mixed infections compared with P. vivax infections (p=0.041), and the PDW was lower in mixed infections compared with P. falciparum infections (p=0.011);additional comparisons by species were not signifi cantly different.

Thrombocytopenia severity and others variables
To explore the relationships between the severity of thrombocytopenia and age, duration of illness and hematological variables, the variables were compared among the patients with no alteration in PC (group A) and the patients with mild (B), severe (C) or profound (D) thrombocytopenia (Table 3).Patient age was not different between the patients with normal or low PCs (p>0.05) but was signifi cantly different between the patients with mild and severe thrombocytopenia (p<0.05), which indicates that younger age could be a risk factor for severe platelet reduction.The median duration of illness was similar between the non-thrombocytopenic patients and the patients with mild changes; however, the duration was signifi cantly higher in the patients with PCs below 50,000/µL (p<0.05).Parasitemia was signifi cantly higher and the erythrocyte count was signifi cantly lower in the patients with PCs below 50,000/µL compared to the other patients groups (Table 3).
The comparison between the patients with normal PC and the patients with any degree of thrombocytopenia showed significant differences in the leukocyte count, with the latter group characterized by signifi cantly lower counts of lymphocytes (p<0.001),monocytes (p=0.003),neutrophils (p=0.010) and eosinophils (p=0.029).
In the cases of severe or profound thrombocytopenia, all variables with the exception of age were signifi cantly different compared with the patients with normal counts.Lower leukocyte and erythrocyte counts, higher coagulation times [prothrombin time (PTi) and partial thromboplastin time (PTT)] and higher parasitemia were observed in the patients with severe or profound thrombocytopenia.
Finally, the comparison between the patients with mild, severe and profound thrombocytopenia revealed no leukocyte count differences (p>0.05).Young age, a longer duration   CM: complicated malaria.
of illness and higher parasitemia were related to severe thrombocytopenia, a lower erythrocyte count and a longer coagulation time.
Patients with complicated malaria (with or without severe thrombocytopenia) exhibited signifi cantly lower median values for PC and PT and higher median values for PDW (Table 4).Median PC values were signifi cantly lower in the patients with complications other than severe anemia or acidosis.In the patients who experienced complications, the median PT was lower and the median PDW was higher compared with the patients who did not experience complications.Furthermore, the PDW was signifi cantly lower in the patients with renal or neurologic complications.The MPV was similar across patients.
Finally, we compared the laboratory data among malaria patients with and without severe thrombocytopenia and stratifi ed by Plasmodium specie (Table 5).The laboratory data were signifi cantly different across categories, with the exception of the white blood cell count, serum creatinine and serum glycemia.In the patients infected with P. falciparum, severe thrombocytopenia was related to abnormal values on all liverfunction tests, BUN, PTi and PTT; in the patients with P. vivax, a relationship was identifi ed between severe thrombocytopenia and abnormal values on liver-function tests, BUN, serum glucose, erythrocyte and white cell counts.

DISCUSSION
In addition to anemia, a reduction in the number of platelets is one of the more well-known hematologic changes observed in patients with malaria.While platelet-count changes are particularly associated with P. falciparum malaria 9,12,45 , this pathology has also been identified in malaria caused by other species, such as P. vivax 36,46 .Changes in other platelet parameters, such as PT, MPV or PDW, have not been well described in malaria patients.
Some studies have found that patients infected with P. vivax or P. falciparum develop anemia, monocytosis, eosinopenia and lymphopenia 45 .This study found a progressive reduction in hemoglobin levels and white blood cell counts with reduced PCs, which resulted in anemia, monocytopenia, neutropenia and lymphopenia.Thrombocytopenia presented frequently in the patients infected with P. falciparum (67%) and P. vivax (64%).The PC did not differ by species, which suggests that a similar mechanism governs this interaction in P. vivax and P. falciparum infections.Signifi cant differences in PT and MPV were observed among species, but the values remained within the normal range.The PDW was signifi cantly higher in the patients infected with P. falciparum compared with those infected with P. vivax; however, both median values were above the normal range.These fi ndings and the results of various studies that have demonstrated a rapid recovery of PCs following the initiation of antimalarial treatment suggest that thrombocytopenia is related to a systemic endothelial and platelet activation rather than medullary suppression.Systemic endothelial activation likely mediates platelet sequestration and does not depend exclusively on the formation of platelet aggregates and parasitized as observed in P. falciparum infections 47 .
Interestingly, we found that the patients with P. vivax infection and severe thrombocytopenia had more signifi cant changes in leukocyte and erythrocyte counts compared with the patients without thrombocytopenia.These changes, which were not identifi ed in the patients infected with P. falciparum, favor a stronger infl ammatory response in severe P. vivax infections that can affect bone marrow via the production of tumor necrosis factor alpha (TNF-α).This hypothesis is supported by a study that identifi ed increased production of TNF-α in P. vivax malaria and a stronger host immune response in P. vivax malaria compared with P. falciparum malaria 48 .
Various mechanisms have been proposed to explain thrombocytopenia in malaria; however, its clinical signifi cance remains uncertain.Thrombocytopenia can be interpreted as evidence of damage and has been associated with clinical complications and high parasitemia 9,12,37 .However, some studies have suggested that thrombocytopenia has a protective effect, which reduces the probability of red blood cell aggregation 28 or mediates parasite destruction.The latter effect has been observed in vitro in the early stages of erythrocytic infection with P. falciparum and in a mouse model with Plasmodium chabaudi 49 .In this study, we identifi ed a signifi cant relationship between severe thrombocytopenia and other clinical complications, but we were not able to establish causality.
Thrombocytopenia has been associated with disturbances in coagulation and has been implicated in cases of disseminated intravascular coagulation (DIC) 50 .In children with severe malaria, prolonged PTi, PTT and thrombin time (in 47.5, 35.0 and 62.5% of patients, respectively) may lead to platelet hypoaggregation 46 .Activation of coagulation may also be responsible, in part, for thrombocytopenia 6 .Our study showed that the patients infected with P. falciparum, but not P. vivax, had more prolonged PTi and PTT when the thrombocytopenia was more severe.This difference may be the result of increased procoagulant activity in P. vivax compared with P. falciparum infections, which develops in response to increased endothelial activity induced by TNF-α levels and the consequent formation of thrombin-antithrombin III 48 .
In this study, patients with severe or profound thrombocytopenia showed higher parasitemia compared with patients without thrombocytopenia, which is in agreement with previous fi ndings.Low PCs have been associated with increased parasite load.This association may result from sensitization induced by the parasitized red blood cells in the platelet, with consequent increased platelet sensitivity to adenosine diphosphate (ADP) and higher dense-granule secretion 45,46 .These changes could ultimately promote platelet aggregation on the endothelium, which has been demonstrated in cerebral malaria 27 .
Plateletcrit is a measure of platelet mass, the clinical signifi cance of which should be interpreted in terms of the number and size of the platelets.Whitfi eld et al. 51 demonstrated that in healthy subjects, there was an inverse relationship between the platelet size (MPV) and the platelet number (PC) and proposed that the total platelet mass is closely regulated by changes on MPV or PC.We found that the PT was low in 36% of patients with malaria.This fi nding may be explained by a combination of low PC with normal MPV (63%), low PC and MPV (35%) and normal PC with low MPV (2%).The reduction in circulating platelet mass is best explained by a reduction in platelet number rather than a reduction in platelet size.It is suspected that in malaria, sequestration leads to pseudothrombocytopenia 52 .Some studies have identifi ed a relationship between the decrease in PC and the appearance of clinical complications, such as anemia, respiratory distress syndrome and cerebral malaria 37 .We found that the patients with severe (25,000-49,999 platelets/µL) or profound thrombocytopenia (<25,000 platelets/µL) had more pathological changes in their clinical status compared with the patients without thrombocytopenia.Furthermore, parasitemia, the duration of disease, and TP and TPP increased with an increasing severity of thrombocytopenia (mildmoderatesevereprofound), whereas erythrocyte and leukocyte counts decreased.
In this study, the patients who experienced complications had lower PC and PT values and higher levels of PDW compared with the patients who did not experience such complications.Specifi cally, the median PCs were signifi cantly lower in the patients with hepatic complications (90,500 vs. 129,000; p≤0.001), renal complications (80,000 vs. 114,000; p=0.017), neurologic complications (42,000 vs. 122,000; p≤0.001), pulmonary complications (55,000 vs. 122,000; p=0.001) and severe hypoglycemia (40,500 vs. 119,000; p=0.004).The low PC may result from the adhesion of platelets to the sites of vascular injuries and the mediation of malaria-parasitized red blood cells (pRBC) sequestration to the endothelium through von Willebrand-factor strings 53 .
When platelets decrease in number, bone marrow megakaryocytes are stimulated by thrombopoietin, and their nucleus becomes hyperlobulated with a higher deoxyribonucleic acid (DNA) content.These stimulated megakaryocytes produce larger platelets.Thus, platelets with a higher MPV are expected to be present in autoimmune thrombocytopenia when megakaryocytic stimulation is present.Conversely, platelets with a lower MPV are expected in thrombocytopenic states associated with marrow hypoplasia or aplasia 54 .Interestingly, 83% of the patients studied had normal MPV, and it appears that neither of these two mechanisms associated with thrombocytopenia was signifi cantly activated.
The examination of PDW with MPV provides information regarding thrombocyte volume disturbance.Thrombocyte volume heterogeneity occurs because of the production factors in the bone marrow.PDW is an index of thrombocyte volume heterogeneity similar to erythrocyte distribution, and both MPV and PDW are markers of platelet immaturity and platelet activation 55 .A high PDW predominated in the patients studied (65%), and we identifi ed signifi cantly higher PDW in the total sample of severe patients; however, the analysis for each complication indicated confl icting results, which could a result of the differential patterns of infl ammation in the spectrum of clinical complications or statistical effects based on the low number of cases when the complications were individually analyzed.
In healthy human subjects, no correlation was identifi ed between PCT, MPV or PDW values and platelet aggregation results 54 ; however, these relationships could be modifi ed in the presence of infl ammation.
It has been reported that PDW increases over storage time because of the formation of abnormally small and large platelets 54 , and it is well known that pseudo-thrombocytopenia occurs as a result of exposure to EDTA when used in blood sampling 56 .There is a need for multi-center and prospective studies with large sample sizes to elucidate the utility and clinical importance of these parameters in malaria.
Our results indicate that thrombocytopenia accompanies malaria complications, especially liver dysfunction.Major changes in total bilirubin, direct bilirubin, aspartate aminotransferase, alanine aminotransferase, prothrombin time and partial thromboplastin time were observed in severe thrombocytopenia (<50,000 platelets/µL).The PCs were significantly lower in the patients who experienced liver, renal, neurologic, and pulmonary complications and severe hypoglycemia.
The liver is a key regulator of platelet production through the synthesis of TPO.Thus, hepatic dysfunction may explain, in part, the thrombocytopenia observed in these patients; however, more data are required because TPO appears to rise during infection by P. falciparum, even when liver function is compromised 23 .
Differences between P. falciparum and P. vivax infections may arise from the different mechanisms by which the physiopathology of thrombocytopenia is mediated, i.e., via clumping in P. falciparum and via medullar suppression in P. vivax.A better understanding of these mechanisms is still needed.
Despite the inherent limitations of retrospective analyses, these results may contribute to the understanding of severe thrombocytopenia as a negative prognostic marker in patients with malaria.Our fi ndings suggest a need to study hepatic function and to establish close clinical monitoring of malaria patients.Furthermore, there is a substantial need for prospective studies to elucidate the utility and clinical importance of these parameters in malaria.

TABLE 3 -Thrombocytopenia level stratifi ed by quantitative variables values).
PTi: Prothrombin time; PTT: Partial thromboplastin time.aPTi and PTT in seconds; b Mann-Whitney U test: between A and D, p<0.05 for all comparisons.Martinez-Salazar EL et al. -Platelet profi les and clinical complications in malaria

TABLE 5 -Median values of hematological laboratory tests in relation to severe thrombocytopenia.
a Including mixed infections; b <50,000 platelets/µL; c Mann-Whitney U test.