Evaluation of antigen detection and polymerase chain reaction for diagnosis of amoebic liver abscess in patients on anti-amoebic treatment

Background Diagnosis of amoebic liver abscess (ALA) in patients on anti-amoebic drugs is difficult. There is scanty data on this issue using Entamoeba histolytica (E. histolytica) lectin antigen and polymerase chain reaction (PCR). We studied utility of lectin antigen, PCR, and IgG antibody in diagnosis of liver abscess in patients on anti-amoebic treatment. Liver aspirate of 200 patients, of which 170 had anti-amoebic drug prior to drainage, was tested for E. histolytica lectin antigen by (ELISA), PCR, bacterial culture, and serum IgG antibody by (ELISA). Classification of abscesses was based on result of anti-amoebic IgG antibody and bacterial culture, E. histolytica PCR and bacterial culture, and E. histolytica lectin antigen and bacterial culture. Findings Using anti-amoebic IgG antibody and bacterial culture, 136/200 (68.0%) were classified as ALA, 12/200 (6.0%) as pyogenic liver abscess (PLA), 29/200 (14.5%) as mixed infection, and 23/200 (11.5%) remained unclassified. Using amoebic PCR and bacterial culture 151/200 (75.5%) were classified as ALA, 25/200 (12.5%) as PLA, 16/200 (8.0%) as mixed infection, and 8/200 (4.0%) remained unclassified. With E. histolytica lectin antigen and bacterial culture, 22/200 (11.0%) patients were classified as ALA, 39/200 (19.5%) as PLA, 2/200 (1.0%) as mixed infection, and 137/200 (68.5%) remained unclassified. Conclusions E. histolytica lectin antigen was not suitable for classification of ALA patients who had prior anti-amoebic treatment. However, PCR may be used as alternative test to anti-amoebic antibody in diagnosis of ALA.


Introduction
Amoebic liver abscess (ALA) is caused by protozoan parasite Entamoeba histolytica (E. histolytica), a common parasitic infection in tropical countries [1][2][3]. Approximately 50 million people are infected with E. histolytica annually world-wide, with mortality ranging from 40,000 to 1,000,00 [4]. Most of the mortality due to amoebiasis results from hepatic rather than intestinal infection. Clinical and radiological features of ALA are often somewhat similar with pyogenic liver abscess (PLA). Hence, necessitating laboratory investigation for differentiation between ALA and PLA are required [5]. Currently ALA is distinguished from PLA by microscopic examination, anti-amoebic IgG serology, and culture of aspirate for pyogenic organisms.
Detection of trophozoites on microscopic examination in liver aspirate, though confirmatory of ALA, is quite insensitive [6,7]. Diagnosis of ALA is most frequently made using serum anti-amoebic IgG antibody [8,9]. However, this assay may not distinguish past from current infection, especially in endemic regions [10]. Thus, anti-amoebic IgG antibody may also be detected in a proportion of healthy people [10,11].
Sensitivity and specificity of E. histolytica lectin antigen have been reported to be as high as 80-90% in stool and serum specimens for diagnosis of amoebiasis [12][13][14]. Detection of E. histolytica lectin antigen in liver aspirate may also be useful for diagnosis of liver abscesses. Unfortunately, scanty data available on lectin antigen detection in liver aspirate are often contradictory [14,15]. Several PCR types like conventional, nested, touchdown and real time have been developed that detect E. histolytica DNA in stool samples and liver aspirate [7,12,[16][17][18][19][20][21]. Sensitivity and specificity of conventional PCR is comparable with touchdown as well as real time PCR [20,22]. Recently, amoebic DNA has been reported in saliva and urine specimens with low sensitivity [23,24]. E. histolytica DNA detection in liver aspirate has been found to be promising for diagnosis of ALA in a few recent studies on a small number of patients [14,19,25,26]. Since the administration of antibiotics and anti-amoebic drugs is common in developing countries like India where amoebiasis is endemic, it is essential to evaluate various diagnostic tests in these patients. Accordingly, we aimed to evaluate the anti-amoebic IgG antibody test, E. histolytica lectin antigen assay and conventional PCR in association with bacterial culture for the diagnosis of patients with liver abscess most of whom were on antiamoebic treatment.

Methods
We examined 220 patients who were subjected to drainage of liver abscesses in the radiology department of a tertiary care center over a 3-yr-period (January 2006 to December 2009). Among 200 patients with well-defined liver abscess, 170 (85.0%) received an anti-amoebic drug prior to drainage. Liver abscesses were diagnosed by abdominal ultra-sonography. Patients not requiring aspiration, or drainage, or finally diagnosed to have tubercular or fungal abscesses were excluded from the study. Demographic and clinical parameters were recorded in a standard questionnaire. Informed consent was obtained from all the patients and the protocol was approved by the Institution's (Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow-India) Ethics Committee (PGI/DIR/RC/957/2007).
Five ml of liver aspirate obtained during drainage of abscesses were examined for bacteria using Gram staining and bacterial culture was also done using standard media [27]. Sera obtained from these patients were stored at -40°C till tested. Anti-amoebic IgG antibody was assayed using a commercially available kit (Nova Tec Immunodiagnostica GmbH, CITY, Germany) following the manufacturer's instructions. India is endemic for amoebiasis. Accordingly, 100 age and sex matched volunteers were used as healthy controls for determination of anti-amoebic IgG antibody titer in healthy population. Based on optical density (OD) results from healthy people, a cut-off OD value was determined using receiver operating characteristic (ROC) curve. Serum with an absorbance in excess of cut-off OD on 450 nm was considered positive. E. histolytica lectin antigen in liver aspirate was identified using commercially available kit (TechLab E. histolytica II, Blacksburg, Virginia) within 24 hr or stored at -40°C for no more than 7 days. DNA was isolated directly from the liver aspirate samples using CTAB (hexadecyltrimethylammonium bromide) method [28]. A 125 base pair region of extrachromosomal circular DNA of E. histolytica was amplified as previously described [19]. All PCR assays were repeated twice for the validation of the result. DNA extracts from E. histolytica strains HM-1:1MSS were used as positive controls and E. dispar CDC 0784 as a negative control in each PCR run.
Sensitivity, specificity, positive, and negative predictive values (PPV, NPV) were calculated using standard formulae [29]. Continuous and qualitative variables were analyzed using independent sample t-test, Pearson's Chisquare, and Fisher's exact test wherever appropriate. Two-tailed P-values <0.05 were considered as significant.

Results
Twenty patients were excluded from the final analysis as a definite diagnosis could not be made due to inadequate work-up (n = 12), or drainage not needed (n = 4), tubercular liver abscesses (n = 2), and fungal liver abscesses (n = 2). Demographic, clinical, imaging and laboratory data of 200 patients with liver abscess are shown in Additional file 1: Table S1. Patients with ALA were younger in age, more often male, and had history of alcoholism, diarrhea in the recent past, or blood in the stool (dysenteric). In contrast, patients with PLA were more often diabetic, and had gall stones. On laboratory investigation, patients with PLA had lower levels of serum albumin, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and higher hemoglobin than those with ALA. A solitary lesion in right lobe of liver was seen more frequently in ALA than PLA on imaging.
Bacteria Based on the OD from healthy controls, a cut-off OD value was designated. Best sensitivity (83%) and specificity (63%) of a commercially available kit was calculated using ROC curve at OD = 0.291. In total, 165 of 200 (82.5%) liver abscess patients produced an OD >0.291. However, 35 of 200 (17.5%) patients with liver abscesses had an OD < 0.291, while 141 of 170 (82.9%) patients had anti-amoebic IgG antibody in their sera, which were collected after initiation of the anti-amoebic drug. In contrast, 24 of 30 (80.0%) patients had anti-amoebic IgG antibody in their sera, which were collected before initiation of the anti-amoebic drug (P = ns).

Classification of liver abscesses using anti-amoebic IgG antibody and bacterial culture
In 136 of 200 (68%) patients with liver abscess classified as ALA as anti-amoebic IgG antibody was present in their sera, and aspirate was sterile on bacterial culture. Twelve of 200 (6.0%) patients were classified as pyogenic as their sera did not show anti-amoebic IgG antibody, while their aspirate culture revealed growth of a pyogenic organism. Twenty-nine of 200 (14.5%) were classified as mixed infection since their sera had anti-amoebic IgG antibody and aspirate culture revealed presence of bacteria; 23 of 200 (11.5%), however, remained unclassified as their sera did not have anti-amoebic IgG antibody and aspirate was also sterile on bacterial culture ( Table 1).

Classification of liver abscesses using E. histolytica PCR and bacterial culture
In 151 of 200 (75.5%) patients with liver abscess were classified as ALA as their aspirate had E. histolytica DNA and bacterial culture was sterile. Twenty-five of 200 (12.5%) were classified as pyogenic, since their aspirate did not show E. histolytica DNA but grew bacteria on culture. Sixteen of 200 (8.0%) patients were classified as having a mixed infection as their aspirate had both E. histolytica DNA and bacteria grew on culture, while 8 of 200 (4.0%) remained unclassified as their aspirate did not have E. histolytica DNA and bacterial culture was sterile ( Table 1).

Classification of liver abscesses using E. histolytica lectin antigen and bacterial culture
Twenty-two of 200 (11.0%) patients with liver abscess were classified as amoebic since their aspirate had E. histolytica lectin antigen, but was sterile on bacterial culture. Thirty-nine of 200 (19.5%) were classified as pyogenic as their aspirate did not show E. histolytica lectin antigen, but bacteria grew on culture, and 2 of 200 (1.0%) were classified as mixed infections as their aspirate was positive both for E. histolytica lectin antigen and bacterial culture. However, 137 of 200 (68.5%) remained undiagnosed as their aspirate was negative for both E. histolytica lectin antigen and bacterial culture (Table 1).
Using a combination of various tests, patients not classified into any category was highest by E. histolytica lectin antigen and bacterial culture and lowest using E. histolytica PCR and bacterial culture [8/200 (4 Table 1).

Discussion
The current study showed that E. histolytica lectin antigen in association with bacterial culture of liver aspirate was less useful in classifying liver abscesses. In contrast, serum anti-amoebic IgG antibody or E. histolytica PCR in combination with bacterial culture classified most of the liver abscesses and confirmed the biological significance of these conditions to a reasonable degree [3].
Our study showed that TechLab E. histolytica lectin antigen test was not useful for diagnosis of ALA. This is in contrast to 2 previously reported results in which sensitivity of this test in liver aspirate sample were 40.7% and 50.0%, respectively [14,15]. Anti-amoebic drugs are known to cause false negative results using E. histolytica lectin antigen assay in an animal model [30]. In a study published recently from Malaysia, 29 of 32 (90.6%) patients with ALA [31], E. histolytica lectin antigen had become negative in sera within 48 hr of metronidazole administration. Another study showed, 3 of 7 (42.9%) patients were positive for E. histolytica lectin antigen prior to treatment while only 3/40 (7.5%) patients had lectin antigen during or after treatment [14]. The same study also reported that 9/11 (81.8%) serum samples of patients with ALA had become negative for E. histolytica lectin antigen after 1 wk of antiamoebic treatment. In developing countries where amoebiasis is endemic, anti-amoebic drugs are often used indiscriminately. This could explain the low sensitivity of E. histolytica lectin antigen test on liver aspirate in our study. Thus, E. histolytica lectin antigen test is not useful for diagnosis of ALA in patients receiving anti-amoebic treatment.
One sample was negative by PCR but positive by IgG antibody. The possible explanation to this could be presence of PCR inhibitors in liver aspirate as present in stool samples. Although, we used inhibitors tablet to optimize the PCR, some PCR inhibitors might be coextracted with the DNA, which inhibited the PCR.
One interesting finding of the present study is 2 patients with of ALA, which were diagnosed by PCR but not by IgG antibody. Presence of amoebic DNA in two liver aspirates specimens, establishes amoebic infection in them. The negative result of ELISA (IgG) shows that the host made an alternate mechanism for defense against the parasite. The above cases present an interesting area of research on defense mechanism in host infected with E. histolytica.
In the present study we have used conventional PCR for diagnosis of ALA. Using E. histolytica PCR with bacterial culture, most of patients with liver abscess were classified as ALA. E. histolytica PCR had greater sensitivity for diagnosis of ALA than anti-amoebic IgG antibody. Twenty-three patients with liver abscess in the present study could not be unclassified using antiamoebic antibody with bacterial culture. However, only 8 patients were remained unclassified by E. histolytica PCR with bacterial culture. Thus in country like India where real time PCR is not available in all laboratories, conventional PCR can be used for the diagnosis of ALA. It reduces cost of sample processing compared to in real time and nested PCR.

Conclusions
E. histolytica lectin antigen in combination with bacterial culture was least useful in classifying liver abscesses in patients who had received anti-amoebic drug prior to collection of the liver aspirate sample; in contrast, a combination of serum anti-amoebic antibody or E. histolytica PCR and bacterial culture classified most of the liver abscesses successfully.

Additional file
Additional file 1: Table S1. Demographic and clinical features of 200 liver abscess patients classify using anti-amoebic IgG antibody and bacterial culture.

Competing interests
The authors declare that they have no financial and non financial competing interests.
Authors' contribution UG-Additional Professor and UCG-Additional Professor conceived the study; BM-Professor and TND-Professor designed the protocol; SSB-Professor drained liver pus from patients; VJ-PhD student carried out microscopy, ELISA, PCR; VJ, UG and UCG analysed the data and drafted manuscript; UCG