Treatment of amoebic liver abscess with metronidazole alone or in combination with ultrasound-guided needle aspiration: a comparative, prospective and randomized study


Dr Joerg Blessmann, Prof. Egbert Tannich and Professor Gerd Burchard (corresponding author), Bernhard Nocht Institute for Tropical Medicine, Bernhard Nocht Str. 74, 20359 Hamburg, Germany. Tel.: +49-40-42 818 390, +49-40-42 818 394;
Dr Ho Duy Binh and Dr Duong Manh Hung, Department of Abdominal Surgery, Central Hospital, 16 Le Loi Street, Hué, Vietnam. E-mail:


Thirty-nine patients with amoebic liver abscess (ALA), admitted to the Central Hospital of Hué (Vietnam), were evaluated in a comparative, prospective and randomized study for the treatment of ALA. Adult patients with an abscess located in the right liver lobe and an abscess diameter of 6 to 10 cm were included. Bacterial abscesses were excluded by microbiological examination of abscess fluid in all patients. Nineteen patients were treated with metronidazole for 10 days alone and 20 patients were punctured under ultrasound guidance with aspiration of abscess fluid in addition to drug administration. The clinical symptoms fever, pain in right upper abdomen and liver tenderness, and the laboratory parameters erythrocyte sedimentation rate, white blood cells, haemoglobin and C-reactive protein and the abscess size were determined on the day of admission and followed during an observation period of 38 days. Improvement of liver tenderness was significantly faster in the aspiration group during the first 3 days (P < 0.001), whereas all the other parameters showed no differences between the two groups. This minor benefit is obviously not sufficient to justify routine needle aspiration and advocates drug treatment alone for uncomplicated amoebic liver abscesses with a diameter up to 10 cm located in the right liver lobe.


The cornerstones of amoebic liver abscess treatment are nitroimidazole antibiotics such as metronidazole or tinidazole. In combination with drug treatment, three invasive therapeutic procedures, ultrasound guided needle aspiration, percutaneous catheter drainage and open surgical abscess drainage are effective (Adams & Macleod 1977; Van Sonnenberg et al. 1985). Open surgical drainage is indicated, if at all, only in some complicated cases of amoebic liver abscess (ALA) with abscess perforation. Some studies on the treatment of perforated ALA suggest that in most of these cases catheter drainage is equally effective with less morbidity and mortality (Greaney & Donovan 1985; Ken et al. 1989; Puspok et al. 2000).

The treatment of ALA with or without aspiration in addition to amoebicidal drugs is discussed controversially and therefore the therapeutic approach differs widely. Some authors advocate rare use of aspiration, indicating that less than 10% of all abscesses require this additional invasive treatment (Ralls et al. 1987; Sharma & Ahuja 2001). Others report good results with needle aspiration in addition to amoebicidal drugs and recommend treating about 80% of amoebic liver abscesses by this approach (Pham Van et al. 1996).

There are only a few randomized, prospective studies comparing metronidazole treatment alone with the combined treatment of needle or catheter aspiration and amoebicidal drugs. Sharma and co-workers performed such a study in Delhi (India) with 39 ALA patients and did not find any difference between the two groups (Sharma et al. 1989). However, in this study, sizes of abscesses were significantly smaller in the aspiration group. In a similar study performed in Varanasi (India), patients in the combined treatment group became afebrile more rapidly and pain and abdominal tenderness subsided earlier, whereas biochemical parameters improved similarly in both groups (Tandom et al. 1997). In a third study comprising 41 patients, which was performed in a non-endemic setting, no significant differences were reported (Van Allan et al. 1992). Thus, further studies are required to assess which therapeutic regimen is better suited for the treatment of amoebic liver abscess.

We report on a prospective and randomized study comparing metronidazole treatment alone with the combined treatment of metronidazole and ultrasound-guided needle aspiration. The study was performed in Hué (Vietnam), an area known for its high incidence of invasive amoebiasis (Pham Van et al. 1996; Blessmann et al. 2002).

Materials and methods

Study design

The study was a prospective and randomized trial, performed in the department of surgery of Hué Central Hospital (Vietnam) between May 1998 and October 1999. A total of 39 patients with ALA were enrolled and randomized into two groups. Patients randomized into group A were treated with metronidazole, 30 mg/kg daily in three divided doses for 10 days. A puncture with a fine needle (0.95 mm × 120 mm) was performed in order to obtain a small amount of abscess fluid (0.5–1.0 ml) for microbiological examination. Patients randomized into group B received the same metronidazole treatment as subjects in group A but, in addition, as much abscess fluid as possible was removed by ultrasound-guided needle aspiration (needle size: 1.5–2 mm × 120 mm) within the first 24 h after hospital admission.

Diagnosis of ALA

The diagnosis of ALA was based on sonographic findings (focal round or oval hypoechoic liver lesions), the presence of fever, pain or tenderness in the right upper abdomen, elevated values for sedimentation rate (ESR), leucocyte counts or C-reactive protein and the presence of serum antibodies against Entamoeba histolytica antigen.

Inclusion criteria

Patients included into the study had to fulfill the following criteria: abscess with a diameter of 6–10 cm; abscess localization in the right liver lobe, except the caudate lobe; age >18 years. Exclusion criteria comprised of: treatment with amoebicidal drugs before hospital admission, pregnancy, or presence of aerobic or anaerobic bacteria in the abscess fluid.

Clinical, laboratory and ultrasound examinations on the day of admission

On the day of admission, body temperature, pain and tenderness in the right upper abdomen were documented as part of a thorough clinical examination. Blood examinations were performed for white blood cells (WBC), haemoglobin, erythrocyte sedimentation rate (ESR), C-reactive protein (Fa. Dade Behring, Marburg, Germany; normal value: <5 mg/dl) and antibodies against E. histolytica antigen using a membrane-latex agglutination test (M-LA) and an enzyme-linked immunosorbent assay (ELISA), based on a recombinant E. histolytica surface antigen, as previously described (Lotter et al. 1992; Lotter et al. 1993). Abscess fluids were examined for aerobic and anaerobic bacteria by inoculation on Columbia and Schaedler agar plates as well as by culturing in a thioglycolate-enriched bouillon. Ultrasound examinations were performed using a 3.5 or 4.5 MH sector transducer and the abscess diameter was measured in two planes and abscess volume was estimated using the formula: 4/3π(a/2 × b/2)3.

Clinical, laboratory and ultrasound examinations during follow-up

Patients were followed as inpatients during therapy for 10 days and re-examined 28 days after discharge (day 38). Body temperature, pain and tenderness of the right upper abdomen were followed daily from day 1 to 10 and on day 38. WBC and haemoglobin were measured on day 1, 5, 10 and 38; ESR on day 1, 10 and 38; CRP and antibodies against E. histolytica antigen on day 1 and 38.

Statistical evaluation

The outcome of eight parameters, including body temperature, pain and tenderness in right upper abdomen, WBC, ESR, haemoglobin, CRP and abscess volume was evaluated to reject the hypothesis that patients benefit from needle aspiration in addition to drug therapy. The Mann–Whitney U-test was used to compare differences in body temperature, WBC, ESR, haemoglobin, CRP and abscess size. A graph was created for each parameter and patient, with time (days) on the x-axis and the parameter with the appropriate scale on the y-axis. The area under the curves (AUC) was calculated and used for comparison between the two groups. Fisher's exact test was used to compare pain and liver tenderness rates during follow-up on day 3, 5, 10 and 38. The Bonferroni correction for multiple testing was used and the threshold P-value for statistical significance according to eight parameters was calculated.P-values <0.006 were considered to be significant.

Ethical clearance

The study was approved by the Scientific Council of Education, Training and Ethics of Hué Medical School.


Thirty-nine patients were included into the study according to the inclusion criteria (Table 1). Nineteen patients, randomized in group A, were treated by oral application of metronidazole for 10 days, whereas the remaining 20 patients, randomized in group B, were punctured with evacuation of abscess fluid, in addition to a 10 days metronidazole therapy. The male/female ratio was 15/4 in group A and 16/4 in group B. Despite randomization, the median age was significantly lower in group A (36 years, range 22–64) compared with group B (46.5 years, range 18–80) (P = 0.02). Two patients (5.1%) reported a previous history of amoebic liver abscess, which could be confirmed by the hospital files.

Table 1.  Clinical features and laboratory results in the two treatment groups at admission
ParameterGroup A (n = 19)Group B (n = 20)
  1. * Median (range).

  2. † Mean value (standard deviation).

Gender M/F15/416/4
Age (years)36 (22–64)*46.5 (18–80)*
Fever (°C)†38.0 (0.8)†37.9 (0.7)†
Pain in right upper abdomen9/19 (47%)12/20 (60%)
Liver tenderness19/19 (100%)17/20 (85%)
ESR (mm after 1 h)89 (24)†98 (22)†
WBC (number/μl)14 132 (4389)†14 865 (5136)†
Haemoglobin (g/dl)11.3 (1.5)†11.3 (1.4)†
C-reactive protein (mg/l)210 (139)†224 (145)†
Abscess volume (ccm)169 (90)†161 (49)†

The most frequent clinical findings on the day of admission were liver tenderness (92%), fever (69%) and pain in right upper abdomen (54%). Elevated values for ESR, C-reactive protein and leukocyte counts were present in 100%, 97% and 75% of the patients. In addition, 62% of the study subjects had reduced haemoglobin values of less than 12 g/dl. Serum antibodies against E. histolytica were detected in all 39 patients using an agglutination test and in 92% of the patients using a highly specific but less sensitive ELISA. There were no differences for the various clinical findings and laboratory parameters between the two groups.

Ultrasound examinations revealed a total of 45 abscesses in the 39 patients examined, as there were two patients in group A and four patients in group B with two abscesses. The mean abscess volume was 169 cm3 and 161 cm3, respectively (Table 1), corresponding to a mean abscess diameter of 74 mm in both groups.

Follow-up after initiation of therapy revealed normalization of body temperature in all patients within 6 days, whether belonging to treatment group A or B. Likewise, abdominal pain disappeared in 95% of all patients within 5 days. In contrast, follow-up of liver tenderness revealed a significant difference between the two groups. During the first 3 days only one patient from group A while nine patients from group B reported disappearence of pain during liver palpation (P < 0.001). Within the following observation period, however, all patients cleared for liver tenderness, and on day 5 and further on there were no differences between the two groups (Table 2). On day 38, none of the patients examined had fever, abdominal pain, liver tenderness or reported any other complaints. Likewise, the laboratory parameters improved in all patients during follow-up and showed normal values in most of the study subjects on day 38 (Table 3). There were again no differences between the two treatment groups.

Table 2.  Follow-up and comparison of right upper abdominal pain, liver tenderness and abscess volume
 Pain in right upper abdomenLiver tendernessAbscess volume
No./total no. (%) No./total no. (%) Millilitre 
Group AGroup BP-value*Group AGroup BP-value*Group AGroup BP-value†
  1. P-values were determined by Fisher's exact test.

  2. † P-value was determined by Mann–Whitney U-test.

  3. P-values <0.006 were considered significant.

  4. n.d., not determined.

Day 19/19 (47)12/20 (60)0.52719/19 (100)17/20 (85)0.2311691610.028
Day 34/19 (21)4/20 (20)1.00018/19 (95)8/20 (40)<0.00115987
Day 51/19 (5)1/20 (5)1.00012/19 (63)7/20 (35)0.113n.d.n.d.
Day 101/19 (5)0/20 (0)0.4877/19 (37)4/20 (20)0.30110964
Day 380/17 (0)0/16 (0)1.0000/17 (0)0/16 (0)1.0003324
Table 3.  Follow-up and comparison of laboratory results
  Day 1Day 38P-value*
  1. P-values were determined by Mann–Whitney U-test.

  2. † Mean value (standard deviation).

ESR (mm 1 h)Group A89 (24)†21 (13)0.734
Group B98 (22)29 (16)
CRP (mg/l)Group A210 (139)1 (0.8)0.678
Group B224 (145)  1.8 (1.6)
WBC (number/μl)Group A14 132 (4389)7853 (1726)1.000
Group B14 865 (5136)7267 (1932)
Haemoglobin (g/dl)Group A11.3 (1.5)13.0 (1.6)0.292
Group B11.3 (1.4)12.9 (0.9)

The mean quantity of liver fluid that was aspirated in group B patients comprised of 117 ml (range 50–200 ml) resulting in an immediate decrease of the abscess volumes by about 50%, as determined by ultrasound on day 2. In contrast, there was no reduction in the size of abscesses in group A patients at this time point. However, during ultrasound follow-up the differences between the two groups subsequently decreased. Statistical evaluation revealed no significant difference in abscess resolution over the whole observation period from day 1 to day 38 (Table 2). In one patient from group B, the ultrasound control on day 2 revealed a narrow hypoechoic sickle under the hepatic capsule, as a sign of subcapsular bleeding. The follow-up in this case was uneventful.


In an attempt to compare the clinical outcome of the two most widely used regimens for the treatment of amoebic liver abscess, a prospective and randomized study was performed. This comprised ALA patients treated either only with metronidazole or with a combination of metronidazole and abscess puncture, in order to intially remove most of the abscess fluid. The most difficult differential diagnosis of bacterial abscess was excluded in all patients by microbiological examination of fine needle aspirated abscess fluid. At enrollment there were no differences between the two treatment groups with regard to laboratory or clinical findings as well as gender distribution. However, the mean age of the study subjects was significantly higher in the aspiration group. But it is unlikely that the age differences has influenced the results, as comparisons of the oldest and the youngest subjects within each group as well as between the two groups revealed identical outcomes.

Both therapeutic regimens were equally effective. All patients were cured during the observation period of 38 days and no major complication occured. Resolution patterns for the various laboratory and clinical parameters revealed no differences except for liver tenderness, which disappeared significantly faster in the aspiration group during the first 3 days (P < 0.001). However, the benefit from pain relief is likely not to be sufficient to justify needle aspiration, because of possible complications, which might be severe and even fatal (Piccinino et al. 1986). Symptomatic treatment of pain, if necessary, can be achieved probably equally effectively by administration of analgesics.

To our knowledge, only three similar prospective studies on the treatment of ALA have been reported. The results presented here are in full agreement with two of these studies performed in Delhi, India (Sharma et al. 1989) and Los Angeles, USA (Van Allan et al. 1992), both of which concluded to recommend drug treatment only. The third study performed in Varanasi (India), recommended aspiration of abscess fluid (Tandom et al. 1997). However, this recommendation was based solely on differences in normalization of body temperature and improvement of pain and tenderness in the right upper abdomen during the first 4 days of treatment. Moreover, the statistical evaluation of this study has to be called into question as the authors did not consider Bonferroni correction for multiple testing, although they compared eight different parameters.

Thus, in addition to the aforementioned studies, the results presented here confirm that puncture and removal of abscess fluid are not required for effective ALA treatment. Therefore, we suggest to treat at least those amoebic abscesses with a diameter of up to 10 cm and which are located in the right liver lobe only with amoebicidal drugs, such as metronidazole, as the first line therapy. The effect of metronidazole on ALA is striking, as body temperature returned to normal and pain disappeared in all patients within 5–6 days. This fast improvement of clinical symptoms after initiation of nitroimidazole therapy was also observed in other studies (Ralls et al. 1987; Guérin et al. 1995; Sharma & Ahuja 2001; Weinke et al. 2002). Thus, lack of clinical improvement within a few days after initiation of medical treatment should call into question the diagnosis of ALA.


The study was supported by the Volkswagen Foundation. This article is based on doctoral studies carried out by Joerg Blessmann at the Faculty of Medicine, University of Hamburg.