Laboratory evidence of disseminated intravascular coagulation is associated with a fatal outcome in children with cerebral malaria despite an absence of clinically evident thrombosis or bleeding

Abstract

Background: A procoagulant state is implicated in cerebral malaria (CM) pathogenesis, but whether disseminated intravascular coagulation (DIC) is present or associated with a fatal outcome is unclear.

Objectives: To determine the frequency of overt DIC, according to ISTH criteria, in children with fatal and non-fatal CM.

Methods/patients: Malawian children were recruited into a prospective cohort study in the following diagnostic groups: retinopathy-positive CM (n = 140), retinopathy-negative CM (n = 36), non-malarial coma (n = 14), uncomplicated malaria (UM), (n = 91), mild non-malarial febrile illness (n = 85), and healthy controls (n = 36). Assays in the ISTH DIC criteria were performed, and three fibrin-related markers, i.e. protein C, antithrombin, and soluble thrombomodulin, were measured.

Results and conclusions: Data enabling assignment of the presence or absence of 'overt DIC' were available for 98 of 140 children with retinopathy-positive CM. Overt DIC was present in 19 (19%), and was associated with a fatal outcome (odds ratio [OR] 3.068; 95% confidence interval [CI] 1.085-8.609; P = 0.035]. The levels of the three fibrin-related markers and soluble thrombomodulin were higher in CM patients than in UM patients (all P < 0.001). The mean fibrin degradation product level was higher in fatal CM patients (71.3 μg mL(-1) [95% CI 49.0-93.6]) than in non-fatal CM patients (48.0 μg mL(-1) [95% CI 37.7-58.2]; P = 0.032), but, in multivariate logistic regression, thrombomodulin was the only coagulation-related marker that was independently associated with a fatal outcome (OR 1.084 for each ng mL(-1) increase [95% CI 1.017-1.156]; P = 0.014). Despite these laboratory derangements, no child in the study had clinically evident bleeding or thrombosis. An overt DIC score and high thrombomodulin levels are associated with a fatal outcome in CM, but infrequently indicate a consumptive coagulopathy.

Keywords: blood coagulation; disseminated intravascular coagulation; endothelial cell protein C receptor; fibrin; malaria, cerebral.

Figure 1

Increased plasma markers of intravascular fibrin generation in cerebral malaria (CM). (A) Plasma d‐dimer levels measured on admission/presentation in children with retinopathy‐positive CM (n = 102), retinopathy‐negative CM (n = 31), non‐malarial coma (n = 11), uncomplicated malaria (UM) (n = 55), and mild non‐malarial febrile illness (MF), (n = 28), and in healthy controls (HCs) (n = 27). (B) Plasma fibrin degradation product (FDP) levels in children with retinopathy‐positive CM (n = 74), retinopathy‐negative CM (n = 26), non‐malarial coma (n = 10), MF (n = 34), and UM (n = 64), and in HCs (n = 24). (C) Plasma fibrin monomers in children with retinopathy‐positive CM (n = 114), retinopathy‐negative CM (n = 32), non‐malarial coma (n = 12), MF (n = 25), and UM (n = 51), and in HCs (n = 28). The dashed horizontal line across columns in (C) indicates two standard deviations from the mean in HCs, the threshold for this index in the overt disseminated intravascular coagulation score. The solid horizontal lines in (A)–(C) indicate geometric means, and the bars indicate 95% confidence intervals. Significance was determined by anova with the Tukey honestly significant difference test on log‐transformed data to adjust for multiple comparisons. ***P < 0.001. NS, not significant.

Figure 2

ISTH score for overt disseminated intravascular coagulation (DIC) and its components in cerebral malaria (CM) patients and comatose and non‐comatose controls. (A, B) Prothrombin time (PT) (A) and fibrinogen (B) in children with retinopathy‐positive CM (n = 70), retinopathy‐negative CM (n = 23), non‐malarial coma (n = 11), uncomplicated malaria (UM) (n = 21), and mild non‐malarial febrile illness (MF) (n = 24), and in healthy controls (HCs) (n = 30). The red dashed lines in (A) and (B) indicate a PT prolonged by 3 s over the geometric mean of the HCs (≥ 17.4 s) and a fibrinogen level of > 1 g L⁻¹, the threshold for these indices in the overt DIC score. (C) Platelet count (×10⁶ L⁻¹) in children with retinopathy‐positive CM (n = 138), retinopathy‐negative CM (n = 39), non‐malarial coma (n = 19), UM (n = 88), and MF (n = 77), and in HCs (n = 28). The red dashed line indicates a platelet count of 100 000 × 10⁶ L⁻¹. (D) ISTH DIC scores in children with retinopathy‐positive CM (n = 55), retinopathy‐negative CM (n = 16), and non‐malarial coma (n = 12). For 43 retinopathy‐positive CM patients, 10 retinopathy‐negative CM patients, and two non‐malarial coma patients, data were missing for one or two components, and so a score is not shown, but either the score was already ≥ 5 or it would not reach ≥ 5 regardless of the missing data, and so the presence or absence of overt DIC could still be assigned with certainty. The red dashed line indicates a score of ≥ 5, the threshold for overt DIC. Horizontal lines indicate geometric means, and bars indicate 95% confidence intervals. Significance was determined by anova with the Tukey honestly significant difference test on log‐transformed data to adjust for multiple comparisons. **P < 0.01, ***P < 0.001. The data in (A) and (C) have been published previously 16, but are presented again here for comparison with other coagulation parameters and to enable calculation of the ISTH DIC score. NS, not significant.

Figure 3

Decreased anticoagulant activity in cerebral malaria (CM) and association with outcome. (A) Antithrombin activity measured at admission in children with retinopathy‐positive CM (n = 68), retinopathy‐negative CM (n = 19), non‐malarial coma (n = 11), uncomplicated malaria (UM) (n = 19), and mild non‐malarial febrile illness (MF) (n = 13), and in healthy controls (HCs) (n = 29). (B) Protein C activity in children with retinopathy‐positive CM (n = 68), retinopathy‐negative CM (n = 21), non‐malarial coma (n = 11), UM (n = 20), and MF (n = 18), and in HCs (n = 30). (C) Antithrombin activity at admission in patients with retinopathy‐positive CM grouped according to outcome: those who eventually died (fatal, n = 12) and those who survived (non‐fatal, n = 54). (D) Protein C activity in fatal (n = 12) and non‐fatal cases (n = 56) of retinopathy‐positive CM. Horizontal lines indicate means, and bars indicate 95% confidence intervals. In (A) and (B), significance was determined by anova with the Tukey honestly significant difference test to adjust for multiple comparisons. In (C) and (D), significance was determined with a two‐tailed t‐test. *P < 0.05, ***P < 0.001. ATIII, antithrombin; NS, not significant.

Figure 4

Increased plasma soluble thrombomodulin (sTM) levels in children with cerebral malaria (CM) and association with fatal outcome. (A) Plasma sTM levels measured at admission in children with retinopathy‐positive CM (n = 115), retinopathy‐negative CM (n = 23), non‐malarial coma (n = 10), uncomplicated malaria (UM) (n = 103), and mild non‐malarial febrile illness (MF) (n = 95), and in healthy controls (HCs) (n = 36). (B) Admission sTM levels in fatal (n = 24) and non‐fatal (n = 91) cases. Horizontal lines indicate means, and bars indicate 95% confidence intervals. Significance in (A) was determined by anova with the Tukey honestly significant difference test to adjust for multiple comparisons. Significance in (B) was determined with a two tailed t‐test. *P < 0.05, ***P < 0.001. The data from the HCs and from the MF and UM patients have been published previously 22, but are presented again here for comparison with new data from non‐malarial comatose, retinopathy‐negative CM and retinopathy‐positive CM patients, and in the context of outcome.