Liver abscess (LA) is a suppurated collection in the hepatic parenchyma, caused by a pathogenic agent that may be bacterial, parasitic or fungal. It is relatively rare, with a difficult to assess, as it varies from one region to another. However, LA cases have increased over the last fifty years [1] [2] .

In our context, African, amebic etiology was the most common cause of liver abscess [3] [4] . Pyogenic liver abscesses appeared rarer, and caused mainly by Gram-negative bacilli and anaerobes; they are more common in Europe, where its prevalence varies from 0.29 to 1.47% [5] . The characteristic clinical picture of the disease remains Fontan’s triad (a combination of fever, right upper abdominal pain, and hepatomegaly). However, in most cases, the presentation remains subtle and non-specific.

Diagnosis of LA is facilitated, thanks to the progress made in recent decades in the field of biological and, above all, morphological investigations. Nevertheless, confirmation of the etiology remains difficult in Africa due to limited access to the necessary biological investigations [6] . The diagnosis of certainty is based on the demonstration of abscesses on liver imaging and aspiration of the abscess contents, with microbiological examination. PCR is considered the gold standard to confirm the amoebic nature of abscess for patients living in endemic areas. However, serology is still the most widely used in routine practice due to the unavailability of PCR in most of African countries. Moreover, the amoebic nature can be made simply based on epidemiological evidence, combined with the chocolate-coloured appearance of the pus, which is sterile on culture [7] [8] .

For pyogenic abscesses, positive culture of the pus confirms the diagnosis. The prognosis of LA depends on the patient’s condition, the speed with which the diagnosis is made, the rapid initiation of treatment, the occurrence of complications, and the etiology PLA/ALA and underlying conditions [9] [10] .

Diagnosis of amoebic or pyogenic is of utmost importance. Thus, we undertook this study to compare epidemiological characteristics, clinical features, biological and radiological findings, and outcomes of patients with pyogenic liver abscess (PLA) vs amebic liver abscess (ALA) in order to determine the potential factors that may help improve diagnosis and treatment for this disease in African care system.

During the study period, among 47,553 patients hospitalized during the same period, 67 patients were retrieved in the hospital database. 61 patients were finally diagnosed according to the study criteria, 32 (52.45%) ALA and 29 (47.54%) PLA, representing 0.12% of inpatients ( Figure 1 ). More than half of the patients (60%) were managed in surgical departments.

Amoebic liver abscess is the most common type of abscess in tropical areas. It is the main extra-intestinal complication of amoebiasis, occurring in 3 to 9% of cases [11] . Its frequency in hospital varies greatly from countries in Africa, and it accounted for 0.1 to 0.2% of hospital admissions in Dakar [9] [12] . In Europe, its prevalence is 0.7%, and mainly concerns migrants [1] . Pyogenic liver abscesses are rarer in tropical areas, but account for 80% of LA in Western countries [11] .

The median age of our patients was 39 years [28] -[50]. More than half of our patients (53.7%) were in the 20 - 40 age group, which is similar to the data found in the African literature [6] . Many reports have shown a preponderance of ALA in male patients, who also tend to be younger (with a median age of around 20 - 40 years) than patients with PLA. [13] - [15] Among the clinical signs, jaundice remains the most important discriminating sign [16] . It is very common in pyogenic abscesses. However, in our study, we found other signs such as right basi-thoracic pain, fever, chills.

The markers of the non-specific biological inflammatory syndrome were higher in cases of PLA, in our study. However, according to Sifri et al., laboratory tests, such as leukocytosis (predominantly neutrophils), raised inflammatory markers (e.g. C-reactive protein), increased Alkaline phosphatase and abnormal liver function tests are often present but they have no real value in differentiating PLA versus ALA. [17] .

Imaging techniques, such as ultrasonography and computed tomography (CT) scanning, are useful tools to demonstrate a space-occupying lesion and confirm presence or absence of a liver abscess. It may not reliably differentiate between PLA and ALA [18] . However, amoebic abscesses tend to be single and large in most cases [17] , whereas bacterial abscesses may be single or multiple, ranging in size from a few millimeters to several centimeters. The preferred site for both types of abscesses remains the right liver [8] .

In high-income countries, cause of liver abscess is usually determined using multiple diagnostic strategies, including blood cultures, Entamoeba serology and liver abscess aspirate for culture. Each of these individual options is challenged in the low- and middle-income countries setting. In low-income countries, patients generally present after initial unsuccessful antibiotic therapy, imaging reveals an abscess most often in the suppurated stage due to the long delay of diagnosis, and the cause may remain undifferentiated. Our study confirms this situation.

The color of the aspirate fluid may provide some preliminary information on the cause of the liver abscess. Traditionally, ALA is odorless, chocolate brown and thick, and commonly referred to as anchovy paste, while PLA is usually purulent and foul smelling. Although this may be helpful, its role in differentiation for the purpose of diagnosis remains uncertain. A bacterium is detected in 70 to 95% of cases of pyogenic abscess, providing a guide to antibiotic therapy. In the case of amoebic abscesses, the diagnosis is sometimes based on a suggestive clinical history, combined with the chocolate appearance of the pus, which is sterile on culture. In case of doubt, amoebic serology is a strong argument for the diagnosis.

Positive amebic serology can confirm the amoebic nature of the abscess, but PCR is considered the gold standard for patients living in endemic areas. However, this test is not always widely available in our context. In addition, the diagnostic value of amoebic serology in areas of high prevalence should only be considered informative in the case of high positivity, with a threshold of 1/160 [18] . In fact, antibodies may remain detectable for 6 to 12 months, making moderate serology positivity attributable to a previous infection [7] [8] . The test can also be falsely negative in case of acute presentations, patient’s immune response, and the type of serologic test or the pathogen strain [19] .

Fine needle aspiration for culture is the gold standard for diagnosis of PLA. This is not the case for ALA as parasite culture is insensitive and not routinely available in clinical laboratories. Microscopy also lacks sensitivity as trophozoites are seen in < 25% of cases [17] .

It is recommended to perform a blood culture for any patient suspected of liver abscess on entry. They may provide helpful information in patients before they receive antimicrobials or aspiration of their abscess. Blood cultures are an important adjunct to the diagnosis of pyogenic abscess. But their yield is usually lower than pus aspirate of liver abscess, and the concordance between blood cultures and abscess aspiration cultures was less than 60% [2] [17] . This explains why it was not carried out in our study, as patients’ limited financial resources, mean that bacteriological examination of abscess aspiration fluid is preferred.

Gram-negative bacilli (GNB) are implicated in 40% - 60% of cases of PLA, particularly enterobacteria (Escherichia coli, Klebsiella pneumoniae) and Pseudomonas aeruginosa. Anaerobic bacteria account for 35% - 45%, in particular Bacteroides sp. Gram-positive cocci, most frequently Streptococcus sp., Staphylococcus sp. and Enterococcus sp [1] [20] [21] .

All of PLAs in our study were monomicrobial, and the most common bacterial isolates (found in roughly equal proportions) were E. coli, consistent with results from a number of other studies [22] [23] .

Amoebic liver abscess can be treated medically, while pyogenic liver abscess usually needs to be percutaneously drained and treated with effective antibiotics [24] .

In our context, antimicrobial guidelines generally recommend empiric therapy targeting both amoebic and pyogenic causes of liver abscess. ALA is managed medically, while PLA requires drainage and appropriate antimicrobial treatment. [25] [26] . Surgical drainage is usually reserved for complicated cases and has now been replaced by less invasive methods as the standard of care [26] . Despite improvements in diagnostic and therapeutic methods in recent years, the mortality rate of liver abscesses is still not negligible. It varies from 0 to 14% depending on the series [6] . Mortality rates from PLA have declined over time [27] [28] . Recent retrospective studies report mortalities between 0 and 13%, although length of follow-up varies and often only in-hospital mortality is reported [5] [10] [22] [29] [30] . In the literature, the mortality rate is higher in cases of pyogenic abscesses, especially those presenting complications [5] [29] . In our study, the occurrence of a complication is higher and the mortality was higher in cases of PLA.