Is there a role for bedside ultrasound in malaria? A survey of the literature

Abstract

Purpose

Point-of-care ultrasound (POCUS) has proven utility in the evaluation and treatment of many tropical diseases. Its role in malaria has been studied, but its value for the clinician at the bedside is unclear. Our review aimed at summarizing the existing studies to assess the usefulness, if any, of POCUS in treating malaria.

Methods

We used Boolean operators using keywords “malaria”, “acoustic”, “ultrasound”, “echography”, and “ultrasonography” to search PubMed, Scopus, and Science Direct in three languages (Italian, French, and English).

Results

We found 22 eligible references. Organs explored include the liver, spleen, heart, optic nerve sheath diameter (ONSD), kidney, lungs, and cerebral vasculature. Multiple pathologic findings by ultrasound are reported, but few demonstrate clinical utility. Current studies involve small numbers of patients, and a few trends emerge when studies are compared. The ability to combine study results is limited due to the significant heterogeneity that exists between studies in regards to both methods of evaluation and the reporting of organ pathology and malaria severity.

Conclusions and assessment

A review of the current literature indicates that the use of ultrasound by clinicians adds little to the diagnostic evaluation of patients with malaria. Our review did find that measurements of the spleen, lungs, optic nerve sheath diameter, and cerebral blood flow have potential utility in specific patient populations. Further studies are needed to evaluate whether this utility persists when a larger sample size is used.

Introduction

Malaria is a disease caused by parasites belonging to the Plasmodium genus. In 2015, P. falciparum malaria was responsible for 214 million cases and 438,000 deaths, 90% of which occurred in Africa. African children under the age of 4 bear 90% of the burden of mortality from malaria, with estimates of around 306,000 deaths/year [1]. Rapid diagnostic tests (RDTs) and microscopy are usually employed to diagnose malaria. Microscopy is currently the gold standard for malaria diagnosis. However, it can be difficult to use in some settings due to the lack of trained personnel [1]. Criteria for the diagnosis of severe malaria have been established by the WHO, and include clinical parameters, parasitological findings, and laboratory results [1].

The clinical presentation of malaria can vary widely based on both environmental factors and patient factors. Clinical findings in malaria include the presence of myalgia; arthralgia; fever; hypotension; chills; respiratory distress; GI tract symptoms such as diarrhea or vomiting; and, in the case of cerebral malaria, seizures or alteration of consciousness, which can lead to coma. In the case of severe renal failure, oliguria may also be present, and urine might become dark (the so-called blackwater fever) [2]. Laboratory findings include a normocytic normochromic anemia, hypoglycemia, and decreased platelet count. White blood cells are often normal, though they can be markedly elevated. Procalcitonin, CRP, and fibrinogen levels are usually elevated, as are fibrin levels [2]. Signs of a coagulopathy may also be present. Increases in serum lactate, creatinine, bilirubin, urate, and muscle enzymes can be found. Acidosis may also be present [2]. Splenomegaly due to RBC sequestration is a common finding [3].

Patients who have lived their entire lives in hyperendemic malaria zones may only present with fever and body aches, while the disease can cause life-threatening end-organ damage in young children or travelers who have never been previously exposed. Particularly problematic for the clinician is that many other diseases that exist in tropical settings produce similar symptoms of fever, headache, confusion, and circulatory collapse. The diagnosis of severe malaria in endemic areas can be complicated: signs of severe malaria are nonspecific [4–6] and can be the consequence of other illnesses. Advanced laboratory testing is often not available where the burden of malaria is the highest. The presence of parasites does not necessarily translate into pathology: in high-transmission areas, the prevalence of asymptomatic parasitemia in the population can be as high as 70% [7, 8]. While rapid diagnostic tests and blood smears may help to identify if a patient has a malaria infection, they do not prove that the malaria is the cause of the patient’s presenting symptoms. Co-infection with other viral or bacterial pathogens may make it more difficult to tell in the diagnostic workup if malaria is incidental or if it is the primary culprit of disease in an endemic malaria zone [6].

Imaging or testing that would allow the clinician to more accurately identify and risk-stratify patients with severe malaria infections in low-resource settings would be immensely helpful. Because malaria causes splenic sequestration and can lead to end-organ damage in the liver, kidney, lungs, and brain, an imaging modality that allowed a clinician to view pathology in these organs might have the potential to help in the diagnosis and treatment of infected patients. Given this known organ involvement, the research question which we sought to address in undertaking this review was: Does the current body of literature on ultrasound and malaria support the use of POCUS by clinicians as a diagnostic test in patients with suspected or confirmed malaria? Ultrasound is a tool well suited to this purpose, as it has been shown to have value in the diagnosis of other tropical diseases that affect the abdominal and thoracic organs, such as tuberculosis, schistosomiasis, and cystic echinococcosis. It has also shown promise in the evaluation of dengue [9–11]. US has been shown to be an effective instrument for the clinical management of patients with tropical diseases, as it allows either the diagnosis of patients or their follow-up [9–11]. Current ultrasound technology has advanced to the point where ultrasound machines are extremely portable and have excellent image quality. The nature of ultrasound allows for advanced imaging that is minimally invasive and highly repeatable [9, 10, 12].

In the case of malaria, ultrasound would make it possible for the clinician to find the evidence of liver damage causing change in liver echo-texture, as well as evidence of RBC sequestration in the spleen. Thoracic ultrasound has advanced considerably in the past 15 years, and finding pulmonary edema or pulmonary infiltrates in malaria patients might also be possible [13]. Ultrasound has also proven to be valuable in screening for elevated intracranial pressure by looking at optic nerve sheath diameter. Such pressure is a known problem in cerebral malaria, and is particularly devastating in pediatric patients [14]. We hoped to survey the literature to see if any of these ultrasound exams would be useful in the clinical care of malaria patients, particularly with regard to risk stratification.

Materials and methods

The main data sources used to carry out the literature search were three bibliographic databases: PubMed, Scopus, and Science Direct. The computer search was conducted in three languages (English, French, and Italian) by combining the topic-related keywords “malaria”, “acoustic”, “ultrasound”, “echography”, and “ultrasonography”, and using Boolean operators. We excluded papers on malaria in pregnancy from our review, as the utility of ultrasound in evaluating the effects of malaria infections on fetal growth throughout pregnancy is well documented [15–17]. Fetal biometry that is used to detect growth restrictions is performed and interpreted by a physician with advanced fetal ultrasound credentialing and training. This is due to the fact that small errors in measurement can change fetal growth calculations substantially. We felt that fetal ultrasound has clearly proven benefits when performed by experts. However, fetal biometry is beyond the ability of the average clinician performing bedside ultrasound in malaria-endemic areas, and so fell outside the scope of our research.

The last online search was performed on 06 January 2019. Our initial search yielded 1822 potentially relevant references. The first screening, based on title and/or abstract, eliminated 1656 publications. During the review process, papers dealing with malaria in pregnancy were excluded for the reasons outlined above. Papers were excluded when they discussed the study populations of less than ten individuals. Finally, review papers were excluded, as they were summaries of primary literature sources that were already being considered. No meta-analysis paper was found. After the review process (Fig. 1), 22 references were obtained.

Fig. 1.

Figure detailing the search algorithm used for the study

Patient cohorts

Out of 22 included papers, 10 (45%) included only adults [13, 18–26], 11 (49.5%) only children [14, 27–36], and 1 (4.5%) both children and adults [37]. Only 5 (22.7%) papers included both severe and non-severe malaria cases [13, 18–21], while 5 (22.7%) did not stratify their populations [22–26]; 6 (27.2%) focused only on severe patients [23, 24, 26, 29, 32, 36], and 3 (13.5%) distinguished between severe and cerebral malaria (CM) [28, 33, 35]. The inclusion criteria among the papers that evaluated cases of severe malaria differed: out of the 14 papers considering severe malaria patients, 11 (78.6%) based their inclusion criteria on WHO guidelines [18, 20, 21, 27–31, 33], while one paper divided patients into severe and non-severe on the basis of hemoglobin levels only [19]. Two other papers used a Blantyre Coma Scale ≤ 2 to help further stratify patients [34, 35]. A control group was included only in 5 (22.5%) of the papers [14, 18, 27, 28, 31]. The diagnosis of malaria was based on microscopy in 20 papers (90.9%) [13, 14, 18–26, 29–37], while the methodology was not specified for one paper (4.5%) [32]. In one case, the diagnosis of malaria was based solely on RDT positivity [33]. Of the studies considered, 2 (9%) were carried out in Europe [19, 21], 3 (13.5%) in Oceania [30, 34, 37], 8 (36.4%) on the Indian subcontinent [13, 20, 22–26, 32], and 9 (40.9%) in Africa [14, 18, 27–29, 31, 33, 35, 36] (Table 1).

Table 1.

Populations of considered studies

Study and country	Cases (%)*	Controls (%)*	Uncomplicated (%)**	Severe (%)**	Cerebral (%)**	Total
Adult patients
Zha et al. [18]—Tanzania	56 (100%)	0	16 (29%)	40 (71%)	NA	56
Kachawaha et al. [23]—India	33 (100%)	0	NA	NA	NA	33
Kochar et al. [25]—India	190 (100%)	0	NA	NA	NA	190
Ray et al. [30]—India	27 (100%)	0	NA	NA	NA	27
Hati et al. [37]—India	63 (100%)	0	63 (100%)	0	0	63
Shah et al. [31]—Pakistan	62 (100%)	0	62 (100%)	0	NA	62
Singh et al. [29]—India	82 (100%)	0	NA	NA	NA	82
Ohmae et al. [26]—Solomon Is.	50	0	NA	NA	NA	50
Richter et al. [22]—Germany	118 (100%)	0	NA	NA	NA	118
Franzen et al. [24]—Germany	22 (100%)	0	NA	NA	NA	22
Pediatric patients
Laman et al. [19]—Papua N. G.	25 (100%)	0	10 (40%)	15 (60%)	NA	25
Atalabi et al. [20]—Nigeria	170 (56.5%)	131 (43.5%)	85 (50%)	85 (50%)	NA	301
Murphy et al. [21]—Uganda	33 (100%)	0	16 (48.4%)	12 (36.4%)	5 (15.2%)	33
Beare et al. [34]—Malawi	112 (100%)	0	112 (100%)	0	101 (90%)	112
Yacoub et al. [35]—Kenya	30 (100%)	0	30 (100%)	0	NA	30
Newton et al. [32]—Kenya	60 (34.3%)	115 (65.7%)	0	60 (100%)	60 (100%)	175
Kotlyar et al. [33]—Uganda	104 (35.9%)	186 (64.1%)	0	104 (100%)	15 (14.4%)	290
Nguah et al. [27]—Ghana	183 (100%)	0	0	183 (100%)	NA	183
Nayak et al. [28]—India	100 (100%)	0	100 (100%)	0	NA	100
Ohmae et al. [26]—Solomon Is.	34 (100%)	0	NA	NA	NA	34
Corkill et al. [36]—Papua N. G.	61 (100%)	0	61 (100%)	0	NA	61

Only one study enrolled both adults and children [26] and is thus reported twice

* % calculated using the total study population as reference; ** % calculated using the total number of malaria cases as references

Organ systems

The eligible papers examined seven different organ systems. The spleen was assessed in 10 (45%) articles [18, 19, 21–23, 25, 26, 31, 33, 36], the liver in 7 (32%) [18, 22, 23, 25, 29, 31, 37], the optic nerve sheath diameter (ONSD) in 3 (14%) [18, 21, 34], the kidney in 1 (4.5%) [20], the heart in 6 (27%) [21, 24, 27, 28, 30, 35], the cerebral vessels in 3 (14%) [21, 32], and the lungs in 1 (4.5%) [14]. Some of the papers did not indicate the timing of assessment [18, 23, 25, 26, 28–31, 37], while others employed multiple assessments in a matter of days [13, 14, 19] or hours [34]. Some papers compared the US findings at the time of diagnosis and after recovery, over periods range from 21 days to 9 months [19, 22, 24, 27]. Probe frequencies ranged from 1 to 14 MHz. Nine (41%) papers did not specify the probes used in the related studies [14, 19, 25, 26, 29–31, 35, 37]. Table 2 summarizes the ultrasound methodologies described in the reviewed studies. There was a significant heterogeneity between probe types (when listed), probe frequencies, and timing of assessments (Table 2).

Table 2.

Summary of organ systems

Article	Organs	Probes	Time of assessment
Zha et al. [18]	Spleen, liver, ONSD	5–10-MHz linear probe for ONSD 3.5–5-MHz convex probe for abdomen	Not indicated
Laman et al. [19]	Spleen	Not indicated	On days 0, 1, 2, 3, 7, 14, 28, and 42 from begin of treatment
Atalabi et al. [20].	Kidney	5–8-MHz curved array probe	Within 24 h from admission
Murphy et al. [21]	Spleen, ONSD, kidney, heart, cerebral vasculature	1–5-MHz, 17-mm broadband probe for spleen and cardiac imaging A 7–14-MHz, 38-mm broadband linear array probe for the examination of the ONSD	Within 12 h from admission
Richter et al. [22]	Spleen, liver	3–8-MHz convex probe	At admission and after 21 days
Beare et al. [34]	ONSD	7-MHz curved array probe	Repeated hourly if pt. still in deep coma
Yacoub et al. [35]	Heart	Not indicated	Two examinations: at the earliest opportunity and at discharge
Newton et al. [32]	Cerebral vasculature	2-MHz probe	Within 3 h from admission
Ohmae et al. [26]	Spleen	Not indicated	Not indicated
Kachawaha et al. [23]	Spleen, liver	3.5–5-MHz probe	Not indicated
Kochar et al. [25]	Spleen, liver	Not indicated	Not indicated
Kotlyar et al. [33]	Spleen	C5-1 curvilinear probe (1–5 MHz)	At admission and at 24 h
Franzen et al. [24]	Heart	3.5-MHz imaging probe	Within first 3 days after diagnosis and 9 months after recovery
Nguah et al. [27]	Heart	3.5-, 5-, or 7.5-MHz probes	At day 9 and at follow-up after 42 ± 3 days after recruitment
Nayak et al. [28]	Heart	2.5- and 5.0-MHz probes	Not indicated
Singh et al. [29]	Liver	Not indicated	Not indicated
Corkill et al. [36]	Spleen	3.5-MHz probe	Within 24 h from birth
Ray et al. [30]	Heart	Not indicated	Not indicated
Hati et al. [37]	Liver	Not indicated	Not indicated
Shah et al. [31]	Spleen, liver	Not indicated	Not indicated
Leopold et al. [13]	Lungs	5-MHz probe	On days 0, 1, 2, and 3
Brien el al. [14]	Cerebral vasculature	Not indicated	Repeated daily from admission to day 8

Examined organs and measurements

The spleen was assessed in ten papers (45%) [18, 19, 21–23, 25, 26, 31, 33, 36]. Each paper used slightly different formulas to document spleen size. One study considered the maximum longitudinal dimension of the organ [18]. Another considered the spleen volume resulting from the formula 0.524 × length (cm) × width (cm) × depth (cm) [19]. Two others used the geometrical formula for an ellipsoid [22, 33]. Another study calculated the dimension of the spleen using Dittrich’s formula while also measuring length, width, and depth [26]. Another study simply reported the presence of an enlarged spleen detected by ultrasound [36]. All but one found a correlation between the grade of malaria and the presence of splenomegaly; in that one, malaria severity was not assessed [37]. Interestingly, all the studies found that children with severe malaria had more pronounced splenomegaly than those with uncomplicated malaria or those who were healthy controls. However, one study found that children with CM and pronounced splenomegaly had a higher chance of recovering from coma [33]. One paper found no correlation between spleen size and hemoglobin levels [19]. Spleen size was found to decrease in non-African patients with splenomegaly after treatment for malaria, but that was not the case in African patients [22].

The heart was assessed in 6 (27%) articles [21, 24, 27, 28, 30, 35]. One article assessed the septal flattening, the ratio of right ventricular to left-ventricular diameter, peak tricuspid regurgitation jet velocity, and qualitative right and left-ventricular contractility [21]. The left-ventricular end-diastolic diameter (LVEDD) and left-ventricular end-systolic dimensions (LVESD) were measured in three articles [24, 28, 35]. One paper also included the left-ventricular myocardial performance index (LMPI), representing the “global” cardiac function along with a measurement of the inferior vena cava (IVC) collapsibility index (IVCCI) [35]. In one paper, the interventricular septal thickness (end-diastolic DIVST), left-ventricular posterior wall thickness (end-diastolic DLVPWT), left-ventricular ejection fraction (EF%), peak velocity of early filling phase (Ei), peak velocity of atrial filling phase (Ai), and Ei/Ai ratio were assessed [28]. One paper only reported a decrease in EF% and the presence or absence of diastolic disfunction, pericardial effusion, cardiac thrombi, or valvular dysfunctions [30]. One study found that the left-ventricular function was conserved in all assessed patients [21], while, in another, 86.3% of patients showed no pathological finding at echocardiography [24]. In another study, no patients showed the evidence of pericardial effusion [28], while in another 3.7% of patients did [30]. However, it should be noted that in the examined studies, malaria patients were found to be subject to cardiovascular problems in the course of a normal US examination: 17% of the patients with normal ultrasound exams had cardiovascular complications, and 11% had circulatory failure [28]. Finally, one paper showed that an increase in the cardiac index (CI) and stroke index (SI), as well as a reduction of ejection fraction (EF) and fraction shortening (FS) can be present during the course of the disease [27].

Cerebral circulation was assessed in three (14%) studies using Transcranial Doppler (TCD) [14, 21, 32]. All studies assessed cerebral circulation in children. Murphy et al. compared cerebral circulation among children with cerebral, severe, and uncomplicated malaria [21], while Newton et al. just compared children with cerebral malaria with case controls [32]. Brien et al. compared children with the signs of retinopathy and cerebral malaria with case controls [14]. Murphy et al. required a minimum of three recordings of the middle cerebral artery (MCA) on each side, calculating peak systolic flow velocity, mean flow velocity (average peak time), end-diastolic flow velocity, and pulsatility index (PI). Results were averaged for each side and the combined average was calculated [21]. The same measurements were calculated by Brien et al., who further divided findings into the following diagnostic categories: microvascular obstruction, hyperemia, cerebral vasospasm, low flow, isolated posterior hyperemia, and terminal intracranial hypertension (ICH) [14]. Newton et al. scanned the basilar artery (BA) and calculated mean peak-flow velocities (Vm) and Gosling’s pulsatility index (PI). The left-to-right ratios of the cerebral blood-flow velocity (CBFV) and PI were calculated [14]. Murphy et al. reported that 19.4% patients had mean blood-flow velocity > 2 SD below mean value for their age, while 9.7% had a value > 2 SD above the mean value for their age. No regional variation in cerebral blood flow was found [21]. Newton et al. supported these results, reporting irrelevant differences in the CBFV and PI of the right and left MCA and BA compared with those parameters in conscious children [32]. When comparing survivals and deaths, the same study found a maximum MCA PI significantly higher in children who died, but no significant difference in the other TCD parameters [32]. However, Brien et al. found that 98% of children who are clinically diagnosed with cerebral malaria could be categorized into one of several phenotypes compared to the control group. Children with cerebral malaria and the vasospasm phenotype have the highest incidence of neurologic sequelae (45%), while the low-flow phenotype is the most closely associated with death (32% compared to mortality of 22% for microvascular obstruction, 28% for hyperemia, and 18% for vasospasm) [14].

One paper (4.5%) assessed kidney volume in malaria patients by measuring length, width, anteroposterior diameter, and cortical thickness. The volume was estimated using the ellipsoid formula [20]. No correlation was found between parasite counts and kidney sizes. No difference in renal volume was found between children who died and healthy controls [20].

One paper (4.5%) assessed the lungs in patients with malaria and sepsis by examining lung regions and defining aeration patterns, from normal aeration to severe loss of lung aeration and lung consolidation [13]. However, only 31 malaria patients were included in this study, and less than half (13) had severe malaria. No correlation between lung ultrasound findings and mortality was found in the subset of patients with malaria [13]. Interestingly, patients with uncomplicated malaria were more likely to demonstrate lung pathology in ultrasound than those with severe malaria (51% vs. 39%). Given the small numbers in the study, no definitive conclusions can be drawn, but lung findings in patients with malaria did not help with risk stratification in this case.

The optic nerve sheath diameter (ONSD) was assessed in three studies (13%) [18, 21, 34]. Zha et al.’s measurements were obtained 3 mm posterior to the retina, and an ROC curve was generated [18]. A similar approach was also implemented by Murphy et al., who averaged results for each eye [21]. Both studies adopted previously published criteria to determine if the measures were normal or altered [38, 39]. Assessments of the ONSD compared patients with uncomplicated and severe malaria, plus a group of healthy controls [18]. The authors could not find any correlation between ONSD measurements and malarial severity. However, a study of CM patients showed that 100% of them had increased ONSD measurements [21]. In another CM-focused study, 49% of patients had an ONSD of 4.3 mm or more [34]. An increased ONSD did not correlate with an increase in mortality, although there was a greater prevalence of neurological sequelae in the group with increased ONSD [34].

The liver was assessed in seven articles (32%) [18, 22, 23, 25, 29, 31, 37]. Zha et al. measured its size along the right mid-clavicular line in the sagittal plane. In the ROC curve test, the threshold of > 15.1 cm was found to define hepatosplenomegaly [18]. Richter et al. defined hepatomegaly as an index of liver size exceeding the upper 95% percentile of a height-adjusted value in a healthy reference population [22]. Kochar et al.’s study, size and echo-texture of the liver, gall bladder, intrahepatic or extrahepatic bile duct dilatation, and signs of portal hypertension were assessed in all the patients having serum bilirubin > 10 mg % [23]. Results of liver measurements were contradictory. Zha et al. reported a sensitivity of 58.3% and a specificity of 75.0% for patients with severe malaria [18]. A high prevalence of hepatomegaly was reported in two other studies, with prevalence rates of 24/29 (82.7%) [23] and of 25/29 (86.2%) [25]. However, a fourth study reported only a 2.6% prevalence of hepatomegaly [22]. This variance may be explained by the fact that studies conducted by Kochat et al. and Kachawaha et al. measured the liver only in patients with bilirubin > 10 mg % [23, 25], while in the study with a low percentage of hepatomegaly, more than half the population was European and not living in a tropical area [22]. A further study considered the liver diameter along the mid-clavicular line and found that 58.5% of patients with malaria had hepatomegaly. The echogenicity of the liver was also evaluated, and the authors found that there was a positive correlation with the elevation of liver enzymes in patients with hepatomegaly and alterations in liver echogenicity [29]. Another study also examined the liver in the context of P. vivax malaria cases. The authors tried to correlate the presence of a P. vivax relapse with an increase in liver diameter as measured on the mid-clavicular line, and observed a positive correlation. However, the number of patients enrolled in this study was small and no information was given about the presence or absence of other conditions which could cause liver enlargement [37].

The parameters assessed for each organ investigated are summarized in Table 3.

Table 3.

Parameters assessed for each organ in the included papers

	Parameters assessed
Spleen	Longitudinal diameter [18, 21, 31]	Volume (ellipsoid formula) [19, 22, 33, 36]	Volume (Dittrich’s formula) [26]
Liver	Longitudinal diameter [18, 29, 31, 37]	Volume via three measurements (parasternal, mid-clavicular, and ant. axillary [22]
Cerebral vessels	Measurements on MCA: Peak systolic flow, Mean flow, End-dias. flow, Pulsatility Index [14, 21]	Measurements on MCA and BA: mean peak-flow velocities Pulsatility Index—left-to-right ratios of the cerebral blood-flow velocities (CBFV) and PI [32]
ONSD	Bilateral ONSD measurements [18]	Multiple measurements averaged [21]	Measurements 3 mm behind the posterior sclera surface [34]
Heart	Septal flattening; Right-to-left-ventricular diameter ratio; Peak tricuspid regurgitation jet velocity; qualitative right- and left-ventricular contractility [21]	Left atrial and left-ventricular dimensions; fractional shortening; ejection fraction (%EF); end-systolic and end-diastolic volumes; stroke volume; cardiac output [27, 30]	LVEDD, LVESD. interventricular septal thickness (end-diastolic). left-ventricular post wall thickness (end-diastolic). %EF; peak velocity of early filling phase (Ei); peak velocity of atrial filling phase (Ai) and Ei/Ai ratio [24, 28]	LVEDD and LVESD dimensions, left-ventricular posterior wall thickness (end-diastolic); septum thickness; fractional shortening; ejection fraction; mitral and aortic valve flow; left-ventricular myocardial performance index; stroke volume; inferior vena cava collapsibility index [35]
Kidney	Volume (ellipsoid formula) [20]
Lungs	12 lung regions assessed for aeration patterns [13]

Discussion

This study summarizes a heterogeneous collection of articles that attempted to identify ultrasound findings that would help risk-stratify patients with malaria. Included papers assessed the use of ultrasound in multiple organs, with different methods and standard reference values [18, 22, 27]. Some studies gave no information on the type of probe used, while the other studies evaluated the same organ using different probes. Most of the studies involved a small number of patients [19, 21]. Some were done only in adults [18, 29, 30, 37] and others only in children [19–21]. Two studies did not define age groups at all [22, 34]. The largest studies did not stratify patients by severity [22, 34]. No common criteria were used when malarial severity was defined [18, 19]. Only five studies included a healthy control group [14, 18, 20, 32, 33]. Small sample sizes likely contributed to inconsistent results between studies. It is also possible that malaria’s effect on organ systems varies between patient populations. As a result, drawing definitive conclusions about the utility of ultrasound in the care of patients with malaria is difficult.

Evaluation of the spleen did show a good specificity for severe malaria in one study (90.9%) [18], but with low sensitivity, and striking differences in prevalence across studies, ranging from 24% to 82.7% [21–23]. There may be differences in immune and non-immune populations, with reductions in spleen size documented after treatment in non-African patients, but not in the African cohort [22]. Kotlyar et al. suggested that an increase in spleen volume during an episode of severe malaria may have a protective effect in relation to CM and mortality in children. The authors also reported that the absence of acute splenic enlargement in acute P. falciparum malaria may be associated with increased mortality [33]. This could be explained by the fact that the sequestration of parasites in the spleen could diminish the number of parasites invading the brain. However, these results need to be tested in larger studies, especially considering that African patients often manifest splenomegaly as a consequence of the chronic exposure to a multitude of tropical diseases [40, 41]. Results of the assessment of the heart demonstrate the limited role of cardiac ultrasound in the diagnosis of malaria. Abnormalities found by these studies seem nonspecific, and changes found are explained by the pathophysiologic mechanisms set in motion by the presence of an infection. Evaluations of the lung did not demonstrate that pulmonary involvement in malaria patients could help with risk stratification. In both severe and uncomplicated malaria, lung findings were present on ultrasound, which were not seen on CXR. Larger studies are needed in this population to see if there is any clinical utility in these findings, especially given the higher percentage of lung findings in the uncomplicated malaria patients. Studies evaluating cerebral circulation have inconclusive results, with low sensitivity for the detection of severe malaria in two studies [21, 32]. One study did report a significant difference in cerebral circulation findings between children with known cerebral malaria and a comparison group of those without cerebral malaria. However, both case and control populations were carefully selected. Cerebral malaria patients who were included had to have the signs of retinopathy, which is well established as marker of increased severity in cerebral malaria. In this carefully selected subset of cerebral malaria patients, there was a suggestion of prognostic value in specific patterns seen on transcranial Doppler (TCD). This study differed from the other studies on cerebral circulation both in patient selection and in its more specifically defined diagnostic categories, as seen on TCD.

Examination of the optic nerve sheath has the most promise in the evaluation of malaria. However, these measurements are highly operator-dependent and available studies have a small number of patients. Studies focusing on the liver have yielded contrasting results. The evaluation of either the dimensions or the echogenicity of the liver can show alterations in malaria patients, but these are nonspecific and have a little value as prognostic or diagnostic tools.

Conclusions

Our review found that the current literature on the use of bedside ultrasound in the care of patients with malaria is heterogeneous. This is due to differences in study populations and study methodology.

Based on the current evidence, bedside ultrasound performed by a clinician is minimally useful in risk-stratifying patients with malaria. Additional research is needed as ultrasound becomes more widely used in low-resource settings, as there is suggestion of utility in certain patient populations. Measurement of cerebral perfusion and retinal perfusion could be a potentially effective tool, but is far from having proven utility. Measurements of the spleen in children suffering from coma from cerebral malaria could also serve as a prognostic tool. Measurement of the optic nerve sheath diameter in the context of cerebral malaria cases may hold promise as an adjunct to the clinical evaluation of such patients.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Ethical approval

This article does not contain any studies with human participants or animals performed by any of the authors.