The sample consisted of 87 children suffering from severe malaria with conscious impairment hospitalized in the CHDU-B pediatric unit. Two groups were set up:

- Group 1 consisted of 37 children suffering from severe malaria with unrousable coma or “pernicious malaria” or CM;

- Group 2 consisted of 50 children suffering from severe malaria with conscious impairment but without unrousable coma. We have termed these patients as severe malaria with impaired consciousness.

The sampling was exhaustive, with consecutive selection following the order of admission to the pediatric unit.

A malaria diagnosis was adopted in all cases based on a blood smear containing asexual forms of Plasmodium falciparum.

The diagnosis of CM is adopted when there was a nonreactive coma i.e. an absence of reaction to nociceptive stimuli (Blantyre coma score ≤ 2 for children under 5 years; pediatric Glasgow coma score ≤ 9 for those above 5 years of age). That coma must last more than one hour after the end of a generalized seizure, after the administration of an anti-convulsant treatment, or after the adequate treatment of hypoglycemia. An absence of other etiologies of encephalopathy (metabolic disorders, hypoglycemia, meningitis, meningoencephalitis, severe anemia) in order to adopt that diagnosis [19] .

The diagnosis of severe malaria with conscious impairment without unrousable coma (severe malaria with impaired consciousness), was adopted when state of consciousness was altered (pediatric Glasgow coma score < 15 and > 9 for children above 5 years of age or Blantyre coma score < 5 and >2 for those under 5 years of age) with one or several signs of neurological severity (repeated convulsions, lethargy, prostration, sign of decerebrate rigidity, sign of decorticate posturing, opisthotonos or other sign) [19]

For treatment, 85 children were treated with quinine salts by intravenous perfusion (10 mg/kg/8 h in a serum glucose of 20 ml/kg in 4 hours every 8 hours till awakening) and 2 children, all in Group 2, were treated with injectable artemisinin derivatives (4 mg/kg on the first day and 2 mg/kg from the second to the 5th day). All the children benefitted from perfusion of 10% hypertonic serum glucose. Sixteen children (18.4%), 5 (13.51%) from Group 1 and 11 (22%) from Group 2, received a transfusion of packed red blood cells. All the children from Group 1 and 12 children (24%) from Group 2 were put on oxygen. Two months after the children’s discharge a clinical follow-up was done to monitor any sequelae of the illness or treatment.

All the children hospitalized for severe malaria with conscious impairment during the study period were involved in the study. The study excluded all children with an obito-ocular disease or any other cause impeding the performance of optic nerve utra-sound, those with a history of psychomotor development disorders or neurological deficits and those with another disease likely to explain altered state of consciousness.

Information was entered based on children’s medical record and then transcribed on data collection forms drafted for that purpose. An additional questionnaire for children’s parents and/or nursing staff was sometimes necessary. ONSD ultrasound measurements were directly transcribed on those forms.

An ophthalmic ultrasound with optic nerve measurement was performed among all children involved within a maximum of twelve (12) hours after admission. Examinations were performed using a multi-frequency linear probe (5 - 10 MHz) of a Mindray Digi Prince DP 8800 Plus^Ò machine. For ONSD measurement, the child was put in supine position and after applying a gel layer on the upper eyelid, the operator used a linear probe on the latter in the middle position. Two sections (sagittal and transverse) of optic nerve were performed and the ONSD was measured at 3 mm down the back of the eyeball with the electronic cursors placed on the external limits of the nerve sheath perpendicularly to its axis (Figure 1). We carried out two measurements for each eye: a measure in the transverse plane and a measure in the sagittal plane.

The older children were required to keep their gaze directed at the zenith, thus enabling positioning of the papilla and the optic nerve in front of the probe. In the cases where children’s cooperation could not be obtained, particularly among the youngest ones, the ultrasound probe was handled until an adequate image of the optic nerve was obtained. In those cases, the ‘‘loop’’ (or cinema) mode was used.

The optic nerve is measured three millimeters posterior to the disc (Figure 1) perpendicular to its axis sliders being positioned on the external boundaries of the sheaths.

A patient’s ONSD value corresponds to the average of the following four measures:

ONSD = (RESO + REAO + LESO + LEAO)/4

RESO = Right Eye Sagittal ONSD; REAO = Right Eye Axial ONSD

LESO = Left Eye Sagittal ONSD; LEAO = Left Eye Axial ONSD

An ONSD higher than 4 mm in children under 48 months of age and higher than 4.40 mm in those older was considered raised and symptomatic of a raised ICP [20] .

The study variables were age, sex, optic nerve sheath diameter (ONSD), and altered consciousness. The latter was classified according to pediatric Blantyre or Glasgow coma scores as the case may be. In addition, a study variable was also the clinical outcome of sick children (deaths, recovery with or without sequelae): cured, with sequelae, deceased.

The data collected were entered with Epi data software and processed with STATA software. Fisher’s exact test was used to compare the proportions and Kruskal-Wallis non-parametric test to compare the averages. A p-value lower than 0.05 was considered as significant. The correlations were assessed by means of Bravais-Pearson r correlation co-efficient.

This research work got the approval of the local ethics committee and parents of children involved in the study. They were given information in their own language and we got their verbal and sometimes written informed consent.

The characteristics of children’s age and sex are shown in Table 1. There was no

^aStandard Deviation.

difference between the two groups (p = 0.135) regarding the sex. Group 1 children were older than the ones from Group 2 (p = 0.0149).

ONSD was accurately measured in all the children involved (see Figure 2).

Table 2 provides a summary of ONSD values of both groups. The ONSD of Group 1 children was significantly higher than the one of Group 2 children (Chi2 = 28.61; p = 0.0001).

Among the 87 children in the sample, 42 had an ONSD measurement higher than the expected value i.e. a raised ICP of 48.28%. In Group 1, that prevalence was 83.78% (31 children) compared with 22% (11 children) in Group 2 (p < 0.0001).

Figure 3 shows the distribution of ONSD values according to age in Group 1 (r = 0.27 and n = 37) and Figure 4 the one in Group 2 (r = 0.66 and n = 50).

In our study, we had no case of coma at stage. Table 3 gives a summary of ONSD values in mm according to depth of coma (stage) in the sample.

In Group 1, 22 children (59.5%) were cured without sequelae, 8 (21.6%) recovered but with sequelae and 7 (18.9%) died. As sequelae, we observed: muscle contracture in the limbs in 4 cases, loss of ambulation as psychomotor retardation, deafness, aphasia and blindness in each of the cases. There was a difference between the NOSD measures in the different groups of children (p = 0.0014). The NOSD measure was smaller for children who recovered without sequelae than for those who had sequelae (p = 0.023) and those who died (p = 0.002). No difference was noted between children with sequelae and those who died (p = 1). Table 4 summarizes the ONSD of children in this group according to evolution.

*Standard Deviation.

^aStandard Deviation.

^aStandard Deviation.

In Group 2, 49 children (98%) had a satisfactory clinical progression, their average ONSD was 3.87 ± 0.17 mm with extremes of 3.44 and 4.30 mm. In this group, one child died and his ONSD was 3.91 mm (p = 0.7552).

In this study, optic nerve ultrasound determined an ONSD average value of 4.39 ± 0.94 mm and identified a raised ICP in 48.28% of the subjects. This raised ICP prevalence is similar to the one found by Beare et al. in Malawi in 2012 (49%) within a population of children with severe malaria, all forms combined [5] . A prevalence lower than 33.33% was identified by Murphy et al. in Uganda in 2011 in children suffering from malaria (uncomplicated forms included) [18] .

ONSD average and raised ICP prevalence was higher (p < 0.0001) in Group 1 (5.09 ± 1.09 mm and 83.78%) than in Group 2 (3.87 ± 0.17 mm and 22%). These findings confirm raised ICP role in the severity of neurological disorders in severe malaria with conscious impairment.

Raised ICP role in the disorders of consciousness in severe malaria with conscious impairment affected children is also suggested in this study by the significant increase (p < 0.0001) of ONSD average measurement associated with severity of coma (Table 3).

In healthy children, ONSD raises with age and this growth occurs especially within the first 48 months of life [11] [20] [26] . That correlation continues during severe cerebral malaria without unrousable coma (Group 2) with an average correlation of r = 0.66; whereas in the series of children suffering from CM (Group 1) that correlation between ONSD and age becomes low (r = 0.27). This reflects the fact that in children suffering from CM, as clinically defined, raised ICP occurs in varying degrees.

Clinically, raised ICP can also be identified through the presence of papilledema in the fundus. Unfortunately, this is a late onset sign and its reliability is limited. Several research works have noted the absence of correlation between presence of papilledema and an unfavorable clinical progression [7] [33] . However, other studies have demonstrated that the presence of papilledema is highly associated with death [34] . Furthermore, papilledema is more common among children with raised ONSD [5] . Papilledema may be associated with the following signs of malaria retinopathy: retinal pallor, vascular changes and retinal hemorrhages. Those malaria retinopathy signs are common and encountered in most cases of CM [35] . In malaria-endemic areas, these are signs that help differentiate comas due to malaria from other etiologies of coma [35] [36] . Besides, in developing countries the most sensitive technique for identifying those signs (indirect ophthalmoscopy) is expensive and its availability is limited [35] .

In our series, the ONSD of children with sequelae or those who died was significantly higher than the one of children who are cured (p < 0.0001). But no difference was noted between the ONSD of deceased children and those of children with sequelae. There are serious reasons to fear adverse clinical progression (sequelae or death) in any child with severe cerebral malaria with an ONSD greater than or equal to 5 mm (Table 4). Newton et al. had also reported in 1997 that children with a severe raised ICP had more frequent sequelae [21] . In 2012, while concluding their research work conducted on CM, Beare et al. also reported a significantly raised ONSD identified in children discharged with sequelae compared to children who recovered [5] .