Hepatobiliary morbidities detected by ultrasonography in Opisthorchis viverrini-infected patients before and after praziquantel treatment: a five-year follow up study

Abstract

Infection of the liver fluke, Opisthorchis viverrini (OV) is an important public health problem in northeast Thailand and adjacent countries, where people have a habit of eating raw or undercooked fish. A community case-control study was carried out with 8,936 participants from 89 villages, in Khon Kaen province, Thailand. There were 3,359 OV-infected participants all of whom underwent ultrasonography of upper abdomen for the evaluation of hepatobiliary morbidity. The participants with advanced periductal fibrosis (APF) by ultrasound (n = 785) were invited to undergo annual follow-up ultrasonography for five years after praziquantel treatment. The sonographer was blinded with respect to status of OV infection at each visit. The study findings revealed variability in the study population profile of the hepatobiliary morbidities before and after praziquantel treatment over the follow up interval. At the end of the study, 32 (30.8%) out of 104 participants showed no relapse of APF whereas, by contrast, 39 (37.5%) participants showed relapse or persistent APF since the outset of the study (≥ two consecutive visits). The APF in most follow-up visits was significantly associated with male sex, with intrahepatic duct stones, with the width of the gallbladder “pre” minus “post” fatty meal, and with the ratio of left lobe of the liver to aorta. Five cases of suspected cholangiocarcinoma were observed over the five years of follow-up. This long-term ultrasound follow-up study demonstrates a significant incidence of persistent APF in over one-third of opisthorchiasis cases after praziquantel treatment, findings that support the prospect of ongoing cholangiocarcinogenesis long after successful elimination of liver fluke infection among the population.

Graphical abstract

1. Introduction

The carcinogenic liver fluke, Opisthorchis viverrini (OV) remains an important public health problem in many parts of SE Asia, particularly in Northeast Thailand and the Mekong countries with over 10 million people infected (Sripa et al., 2010). Humans acquire the infection by eating raw or undercooked freshwater cyprinid fish harbouring infective parasite metacercariae. Once ingested, the worms excyst in the duodenum, migrate to extrahepatic and intrahepatic and bile ducts via the ampulla of Vater, mature, and reside mainly in the intrahepatic bile ducts for over 10 years (Sripa et al., 2010). The infection is associated with liver and biliary diseases including cholangitis, gallstones, periductal fibrosis and, a fatal bile duct cancer, cholangiocarcinoma (CCA) (Sripa et al., 2018). Northeast Thailand has reported to have the highest incidence region of this cancer in the world (Banales et al., 2016). Recently, co-infection with the carcinogenic bacterium Helicobacter pylori was shown to in synergistically enhance CCA development during chronic opisthorchiasis (Sripa et al., 2017). Given that CCA is a slow growing tumor, once it becomes clinically manifest, the overall one-year mortality rate is >80% (Chaiteerakij et al., 2017). Therefore, in-depth research on the etiology, pathogenesis, and early detection is needed. Until now, ultrasonography is the most practical method to screen for biliary tract morbidity in liver fluke endemic areas (Chamadol et al., 2019; Mairiang et al., 2012).

Our previous community-based, cross-sectional studies in OV-endemic regions of Khon Kaen and neighboring provinces in Northeast Thailand showed a wide range of hepatobiliary morbidities including periductal fibrosis, cholecystitis, gallstones and suspected CCA using abdominal ultrasonography (Elkins et al., 1996; Haswell-Elkins et al., 1994; Mairiang et al., 1992). The periductal fibrosis detected as enhanced periportal echoes in the liver parenchyma by ultrasound was histologically confirmed in chronic opisthorchiasis (Chamadol et al., 2014) and over 24% of advanced periductal fibrosis (grade 2 & 3) has been reported during chronic opisthorchiasis (Mairiang et al., 2012; Thinkhamrop et al., 2020). Parasite-specific interleukin 6 (IL-6), among other inflammatory and anti-inflammatory cytokines, may play an important role in the pathogenesis of advanced periductal fibrosis (APF) in opisthorchiasis (Sripa et al., 2009). Whereas the frequency of APF gradually decreased over 10 months after praziquantel treatment, APF persisted in many individuals, and, notably, two masses were observed in the right hepatic lobe of one individual who was previously considered normal (Mairiang et al., 1993). Given that APF is considered a precursor lesion of CCA (Chamadol et al., 2014; Mairiang, 2017), we have hypothesized that persistent APF is associated with increasing risk of CCA development (Sripa et al, 2009; Sripa et al. 2012; Brindley & Loukas, 2017). However, a long-term follow up study on hepatobiliary abnormalities post-praziquantel treatment to address the hypothesis has hitherto not be available. This study, therefore, aimed to monitor the liver fluke infection-associated hepatobiliary morbidities, particularly APF, in an ultrasonography based, five-year follow up study.

2. Materials and methods

2.1. Study participants

This study is a part of a community-based, case control study in the pathogenesis of liver fluke induced cancer in Thailand (Sripa et al., 2009) and serves as the continuation of the previously published study (Mairiang et al., 2012). The study was approved by Khon Kaen University Ethic Committee (reference number HE4880528). Recruitment of volunteers as study participants was carried out based on a group of villages under the responsible Health Promoting Hospital (Primary Care Unit) at a time, starting from baseline screening and then followed up annually for five years. In total, the study duration was eight years, from August 2007 to September 2015. At large, the study included 8,936 participants from 89 villages of four districts of Khon Kaen province, aged 20 to 60 years. Stool samples were obtained from each participant at baseline screening of the first year and at one month prior to ultrasonography in each follow-up visit. The stool samples were examined for OV infection status using the formalin ethyl acetate concentration technique and quantified as eggs per gram of feces (EPG) (Elkins et al., 1990).

2.2. Ultrasonography

A total of 785 participants underwent ultrasonography before and after treatment with praziquantel, 40 mg/kg. Originally, the study was designed to follow up the participants for two years. However, at the end of the second-year follow-up, the annual follow-up was extended for five years after the screening as the second phase of the study. Ultrasonography of the upper abdomen was performed using a mobile high-resolution ultrasound machine (LOGIQ E book, GE Healthcare, Chicago, Illinois, USA). Hepatobiliary abnormality (HBA) including periductal fibrosis in the liver parenchyma, gallbladder wall, gallbladder size (width, length and cross section), sludge, suspected CCA (dilated intra or extra hepatic bile duct and/or liver mass) were recorded as described (Mairiang et al. 1993; Mairiang et al., 2012). For the study of gallbladder function (contractility), the gallbladder was measured 30 minutes after consumption of a fatty meal (Mairiang et al., 1993). A single radiologist performed all the examinations without knowledge of the liver fluke infection status of the participant.

2.3. Statistical analysis

Demographic and baseline variables were described. Percent distributions were presented for all categorical variables. For continuous variables, mean (± standard deviation) or median (range) were used. The progression of advanced periductal fibrosis (APF) of each participant was plotted as a horizontal line plot to visualize longitudinal categorical data (Tueller et al., 2016). Relationships among informative demographic factors and various ultrasonographic findings on APF were investigated. Logistic regression was used to obtain odds ratios (ORs) and their 95% confident intervals. Adjusted odds ratios were estimated for effect of sex and age groups with multiple logistic regression. All analyses were performed using R version 3.4.3 (R Foundation for Statistical Computing, Vienna, Austria). All statistical tests were two-sided and a significance level was set as 0.05.

4. Results

4.1. Liver fluke infection status, demographics, baseline and follow up characteristics

This study followed the progression of hepatobiliary abnormalities in the cohort of participants diagnosed with APF, as reported previously (Mairiang et al., 2012). Of 8,936 participants from 89 villages of four liver fluke endemic districts of Khon Kaen in this study, 3,640 participants were stool positive for OV eggs (41.5%), displaying a range of intensity of liver fluke infection from low to high. Among the egg-positive individuals, 3,359 participants consented to ultrasonographic examination of the upper abdomen which, in turn, revealed that 785 participants had developed severe periductal fibrosis. Details of the infection status, demographics and ultrasonographic findings are provided in Table 1. The individuals in the study cohort with severe periductal fibrosis were recruited for this follow-up study. During each follow-up visit, study participants were provided with praziquantel treatment as well as health education aiming for behavioral change away from eating raw or undercooked fish diets, resulting in more than 80% parasite eradication over the course of the study (Table 2). Fewer than 10% of the participants who returned for each follow-up visit were diagnosed with infection with OV. Almost all the participants (99.0%) with liver fluke infection had light OV burden, <500 EPG, at the first follow-up visit 12 months after the enrollment. Thus, the interventions provided to participants adequately alleviated the parasite burden. Nonetheless, the decreased parasite burden was not, in parallel, translated to resolution and elimination of APF (Figure 1).

Table 1.

Demographics and baseline characteristics of the study participants (Mairiang et al., 2012).

Characteristics	n = 785
Sex
- Male	469 (59.7%)
- Female	316 (40.3%)
Age groups (years)
- 20 to < 30	51 (6.5%)
- 30 to < 40	142 (18.1%)
- 40 to < 50	299 (38.1%)
- ≥50	293 (37.3%)
- Mean ± SD	45.4 ± 9.0
- Median (range)	46.0 (20.0 - 60.0)
Periductal fibrosis status
- APF+2	738 (94%)
- APF+3	47 (6%)
Intensity of liver fluke infection
- Light (< 500 EPG)	647 (82.4%)
- Medium (500 to < 1,000 EPG)	58 (7.4%)
- Heavy (1,000 to < 2,000 EPG)	47 (6%)
- Very heavy (≥ 2,000 EPG)	33 (4.2%)
- Mean ± SD	404.4 ± 1,084.9
- Median (range)	96 (2 −17,523)
Ratio of left lobe of the liver to aorta
- Mean ± SD	2.95 ± 0.67
- Median (range)	2.89 (1.35 - 5.89)
Gall bladder dimensions “Pre” minus “Post” fatty meal (cm)
- Length, mean ± SD	1.52 ± 1.17
- Width, mean ± SD	0.67 ± 0.56
- Cross-section, mean ± SD	0.54 ± 0.51
Kidney
- Normal	760 (96.8%)
- Cortical cyst	4 (0.5%)
- Stone	19 (2.4%)
Other ultrasonography findings
- Gall stone	46 (5.9%)
- Intra hepatic duct stone	6 (0.8%)
- Renal stone	18 (2.3%)

Table 2.

Opisthorchis viverrini infection in study participants over the five-year follow-up period.

Intensity of Infection	Time after the screening/baseline visit
	12 months (n = 679)	24 months (n = 617)	36 months (n = 59)	48 months (n = 127)	60 months (n = 123)
Negative (0 EPG)	564 (83.1%)	532 (86.2%)	52 (88.1%)	121 (95.3%)	107 (87%)
Light (<500 EPG)	108 (15.9%)	83 (13.5%)	7 (11.9%)	5 (3.9%)	15 (12.2%)
Medium (500 to < 1,000 EPG)	5 (0.7%)	2 (0.3%)	0 (0%)	1 (0.8%)	1 (0.8%)
Heavy (1000 to < 2,000 EPG)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Very heavy (≥ 2,000 EPG)	2 (0.3%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Participants with infection (> 0 EPG)	115 (16.9%)	85 (13.8%)	7 (11.9%)	6 (4.7%)	16 (13%)
Persistent and increasing	19 (2.8%)	8 (1.3%)	1 (1.7%)	0 (0%)	2 (1.6%)
Persistent but decreasing	96 (14.1%)	14 (2.3%)	0 (0%)	0 (0%)	0 (0%)
Relapse	0 (0%)	52 (8.4%)	4 (6.8%)	0 (0%)	11 (8.9%)
Missing EPG at the previous visit	0 (0%)	11 (1.8%)	2 (3.4%)	6 (4.7%)	3 (2.4%)

^*EPG = eggs per gram

Figure 1.

Horizontal line plot showing Opisthorchis viverrini fecal eggs per gram (EPG) of participants over the course of the study, stratified by the last visit.

Figure 2 presents study participation and advanced periductal fibrosis progression of each participant over the course of the study. During the first phase of the study, 687 (87.5%) and 625 (79.6%) participants returned for the first (12 months after screening) and the second visits (24 months after screening), respectively. At the first follow-up visit, 487 (70.9%) participants returned with APF grade 1 or were negative for APF, which showed that the APF had improved or had been cured. At the second follow-up, 479 (76.6%) participants returned with APF grade 1 or negative, and 146 participants (23.4%) had APF grade 2 or 3. Of the aforementioned 146 patients, 60 patients (41.1%) had APF grade 1 or 0 (negative for APF) at the first visit showing the relapse of APF.

Figure 2.

Horizontal line plot showing advanced periductal fibrosis (APF) of each participant over the course of the study stratified by the last visit. The severities of APF are represented by colors, with red as APF grade 3, yellow as APF grade 2, green as APF grade 1 or negative, and white as lost to follow-up. Only one participant had the last visit at 36 months after screening so that person is included in the “24 and 36” stratum.

After the second follow-up visit, we decided to extend the annual follow-up to five years (as detailed above; Section 2.2). However, delays with protocol amendment and reconsent procedures resulted in significant loss to follow-up with 57 (7.3%), 131 (16.7%) and 104 (13.2%) participants returned for the third (36 months), the fourth (48 months) and the final follow-up visits (60 months), respectively. At the third follow-up visit (36 months), eight (50.0%) out of 16 participants with APF grade 2 or 3 were a result of relapse of APF. At the fourth visit (48 months), 40 (75.5%) out of 53 participants with APF grade 2 or 3 were a result of the relapse of APF. At the final follow-up visit (60 months), 29 (87.9%) out of 33 participants with APF grade 2 or 3 were a result of the relapse of APF. At the end of the study, 32 (30.8%) out of 104 participants showed no relapse of APF over the period of five years while 39 (37.5%) participants showed some relapse or persistent APF since the outset of the study (≥ two consecutive visits). Since the majority of the participants (92.7% of enrolled cases, 60.3% of those returned for the fourth visit and 58.7% of those returned for the final visit) missed the third follow-up visit; the data from this visit were not used for further analyses.

4.2. Follow up findings

4.2.1. First year follow up

At the first follow-up visit, 687 participants returned of which 200 (29.1%) still had APF (grade 2 or 3). Factors that were statistically significantly associated with APF included sex (P < 0.001), ratio of left lobe of the liver to aorta at baseline (P = 0.022), ratio of left lobe of the liver to aorta at the first follow-up visit (P = 0.013), cross-section of the gallbladder dimensions “pre” minus “post” fatty meal at the first follow-up visit (P = 0.006), the presence of intrahepatic stones at baseline (P = 0.022), and the presence of intrahepatic stones at the first follow-up visit (P = 0.030) (Table 3 and Supplementary Table 1). The presence of intrahepatic duct stones positively associated with APF, while other factors including female sex negatively associated with APF. Since 90.7% of participants were cured of OV infection or had only light liver fluke infection (<500 eggs) at the first follow-up visit, and no significant association between baseline EPG and persistent APF was observed, EPG was no longer used to adjust odds ratios. When odds ratios were adjusted with sex and age groups, all factors associated with APF remained significant, except for the presence of intrahepatic stones likely due to an inadequate sample size for multivariate analysis). It was notable that the width of the gallbladder dimensions “pre” minus “post” fatty meal at the first follow-up visit becomes significant after the adjustment (P = 0.044).

Table 3.

Relationships between demographic, previous clinical findings and current clinical findings on advanced periductal fibrosis status as determined by ultrasonography at the first follow-up visit.

			Unadjusted				Adjusted
Characteristics	Number (n = 687)	APF (%)	OR	Lower OR	Upper OR	p-value	OR*	Lower OR	Upper OR	P-value
Sex
Male	406	34.7	1				1
Female	281	21	0.5	0.35	0.71	<0.001	0.49	0.34	0.7	<0.001
Age (years)
20–29	36	38.9	1				1
30-39	114	27.2	0.59	0.27	1.31	0.184	0.59	0.27	1.33	0.199
40-49	271	27.1	0.6	0.29	1.26	0.167	0.54	0.26	1.14	0.099
50+	266	30.1	0.68	0.33	1.42	0.286	0.59	0.28	1.24	0.154
Ratio of left lobe of the liver to aorta at baseline (every 1 unit)	NA	NA	0.74	0.57	0.96	0.022	0.66	0.5	0.86	0.003
Ratio of left lobe of the liver to aorta at 12 months (every 1 unit)	NA	NA	0.7	0.53	0.92	0.013	0.59	00.43	0.8	<0.001
Gall bladder dimensions “Pre” minus “Post” fatty meal at 12 months
a) Length (every 1 cm)	NA	NA	0.96	0.82	1.12	0.575	0.93	0.8	1.09	0.398
b) Width (every 1 cm)	NA	NA	0.77	0.56	1.06	0.115	0.71	0.5	0.98	0.044
c) Cross-section (every 1 cm)	NA	NA	0.6	0.42	0.86	0.006	0.59	0.41	0.85	0.005

EPG = eggs per gram; APF = advanced periductal fibrosis; OR = odds ratio; CI = confidence interval; NA = not applicable;

^*Odds ratio adjusted for age group and sex.

4.2.2. Second year follow up

At the second follow-up visit, 625 participants returned, of whom 146 (23.4%) had APF (grade 2 or 3). Factors that were statistically significantly associated with APF included sex (P < 0.001), APF at the first follow-up visit (P < 0.001), ratio of left lobe of the liver to aorta at the first follow-up visit (P = 0.038), ratio of left lobe of the liver to aorta at the second follow-up visit (P = 0.001) (Table 4 and Supplementary Table 2). Persistent APF at the first follow up visit strongly and positively associated with APF at the second visit. Other factors found to significantly associate with APF appeared to have negative relationships. When odds ratios were adjusted with sex and age groups, all factors associated with APF remained significant. Notably, ratio of left lobe of the liver to aorta at the baseline and the presence of gall stone at the first follow-up visit became significant after the adjustment (P = 0.042 and P = 0.039, respectively).

Table 4.

Relationships between demographic, previous clinical findings and current clinical findings on advanced periductal fibrosis status as determined by ultrasonography at the second follow-up visit.

			Unadjusted				Adjusted
Characteristics	Number (n = 687)	APF (%)	OR	Lower OR	Upper OR	P-value	OR*	Lower OR	Upper OR	P-value
Sex
Male	406	34.7	1				1
Female	281	21	0.5	0.35	0.71	<0.001	0.49	0.34	0.7	<0.001
Age (years)
20–29	36	38.9	1				1
30-39	114	27.2	0.59	0.27	1.31	0.184	0.59	0.27	1.33	0.199
40-49	271	27.1	0.6	0.29	1.26	0.167	0.54	0.26	1.14	0.099
50+	266	30.1	0.68	0.33	1.42	0.286	0.59	0.28	1.24	0.154
Ratio of left lobe of the liver to aorta at baseline (every 1 unit)	NA	NA	0.74	0.57	0.96	0.022	0.66	0.5	0.86	0.003
Ratio of left lobe of the liver to aorta at 12 months (every 1 unit)	NA	NA	0.7	0.53	0.92	0.013	0.59	00.43	0.8	<0.001
Gall bladder dimensions “Pre” minus “Post” fatty meal at 12 months
a) Length (every 1 cm)	NA	NA	0.96	0.82	1.12	0.575	0.93	0.8	1.09	0.398
b) Width (every 1 cm)	NA	NA	0.77	0.56	1.06	0.115	0.71	0.5	0.98	0.044
c) Cross-section (every 1 cm)	NA	NA	0.6	0.42	0.86	0.006	0.59	0.41	0.85	0.005

EPG = eggs per gram; APF = advanced periductal fibrosis; OR = odds ratio; CI = confidence interval; NA = not applicable;

^*Odds ratio adjusted for age group and sex.

4.2.3. Fourth year follow up

At the fourth follow-up visit, 131 participants returned of which 53 (40.5%) participants had APF (grade 2 or 3). The only factor that remained statistically significantly associated with APF was ratio of left lobe of the liver to aorta at this visit (P = 0.016 and P = 0.022 for adjusted odds ratio). The ratio appeared to negatively associate with APF (Table 5 and Supplementary Table 3).

Table 5.

Relationships between demographic, previous clinical findings and current clinical findings on advanced periductal fibrosis status as determined by ultrasonography at the fourth follow-up visit.

			Unadjusted				Adjusted
Characteristics	Number (n = 131)	APF (%)	OR	Lower OR	Upper OR	P-value	OR*	Lower OR	Upper OR	P-value
Sex
Male	92	42.4	1				1
Female	39	35.9	0.76	0.34	1.64	0.489	0.83	0.36	1.84	0.643
Age (years)
20–29	3	33.3	1				1
30-39	27	37	1.18	0.1	27.23	0.900	1.22	0.1	28.34	0.878
40-49	51	35.3	1.09	0.1	24.43	0.945	1.08	0.1	24.28	0.950
50+	50	48	1.85	0.17	41.24	0.626	1.8	0.16	40.2	0.642
Ratio of left lobe of the liver to aorta at baseline (every 1 unit)	NA	NA	0.91	0.53	1.52	0.722	0.98	0.56	1.69	0.944
Ratio of left lobe of the liver to aorta at 12 months (every 1 unit)	NA	NA	0.57	0.28	1.1	0.102	0.59	0.28	1.19	0.150
Ratio of left lobe of the liver to aorta at 24 months (every 1 unit)	NA	NA	0.79	0.38	1.64	0.530	0.7	0.32	1.52	0.377
Ratio of left lobe of the liver to aorta at 48 months (every 1 unit)	NA	NA	0.44	0.21	0.83	0.016	0.43	0.2	0.86	0.022

EPG = eggs per gram; APF = advanced periductal fibrosis; OR = odds ratio; CI = confidence interval; NA = not applicable;

^*Odds ratio adjusted for age group and sex.

4.2.4. Fifth year follow up

At the final follow-up visit, 104 participants returned of which 33 (31.7%) participants had APF (grade 2 or 3). Factors that were statistically significantly associated with APF included width of the gallbladder dimensions “pre” minus “post” fatty meal at baseline (P = 0.011) and width of the gallbladder dimensions “pre” minus “post” fatty meal at the first follow-up visit (P = 0.025) (Table 6 and Supplementary Table 4). When odds ratios were adjusted with sex and age groups, all factors associated with APF remained significant with an addition of APF at the second follow-up visit (P = 0.036). The width of the gallbladder dimensions “pre” minus “post” fatty meal at baseline and the first follow-up negatively associated with APF at the final visit, while APF status at the second follow-up visit positively associated with APF at the final visit.

Table 6.

Relationships between demographic, previous clinical findings and current clinical findings on advanced periductal fibrosis status as determined by ultrasonography at the final follow-up visit.

			Unadjusted				Adjusted
Characteristics	Number (n = 104)	APF (%)	OR	Lower OR	Upper OR	P-value	OR*	Lower OR	Upper OR	P-value
Sex
Male	71	35.2	1
Female	33	24.2	0.59	0.22	1.46	0.266	0.63	0.23	1.65	0.360
Age (years)
20–29	3	0	NA	NA	NA	NA
30-39	21	26.8	1				1
40-49	40	30	1.07	0.34	3.6	0.908	1	0.31	3.41	0.998
50+	40	37.5	1.5	0.49	4.97	0.487	1.27	0.39	4.41	0.693
Hepatobiliary abnormalities at 12 months
APF−	71	33.8	1				1
APF+2/APF+3	33	27.3	0.73	0.28	1.79	0.506	0.59	0.22	1.5	0.283
Hepatobiliary abnormalities at 24 months
APF−	67	25.4	1				1
APF+2/APF+3	37	43.2	2.24	0.96	5.3	0.063	2.58	1.07	6.34	0.036
Hepatobiliary abnormalities at 48 months
APF−	61	24.6	1				1
APF+2/APF+3	43	41.9	2.21	0.96	5.18	0.065	2.19	0.94	5.22	0.072
Gall bladder dimensions “pre” minus “post” fatty meal at baseline
a) Length (every 1 cm)	NA	NA	1.14	0.8	1.62	0.477	1.08	0.75	1.57	0.669
b) Width (every 1 cm)	NA	NA	0.35	0.15	0.75	0.011	0.35	0.14	0.77	0.014
c) Cross-section (every1 cm)	NA	NA	0.45	0.2	1.18	0.123	0.5	0.2	1.2	0.131
Gall bladder dimensions “pre” minus “post” fatty meal at baseline 12 months
a) Length (every 1 cm)	NA	NA	1.07	0.75	1.55	0.697	1.02	0.7	1.49	0.900
b) Width (every 1 cm)	NA	NA	0.36	0.14	0.85	0.025	0.37	0.14	0.92	0.039
c) Cross-section (every1 cm)	NA	NA	0.45	0.18	1.04	0.07	0.49	0.2	1.14	0.109
Gall bladder dimensions “pre” minus “post” fatty meal at baseline 24 months
a) Length (every 1 cm)	NA	NA	1.42	0.93	2.22	0.113	1.33	0.86	2.12	0.210
b) Width (every 1 cm)	NA	NA	0.77	0.31	1.92	0.577	0.73	0.28	1.85	0.508
c) Cross-section (every 1 cm)	NA	NA	0.71	0.29	1.61	0.417	0.63	0.25	1.49	0.308

EPG = eggs per gram; APF = advanced periductal fibrosis; OR = odds ratio; CI = confidence interval; NA = not applicable;

^*Odds ratio adjusted for age group and sex.

4.2.5. Suspected cholangiocarcinoma

Nine participants (1.1%) developed a liver mass during the course of the five-year follow-up (Table 7). Four participants were found to already have a liver mass at the screening/baseline visit. Thus, although the data from only five patients (0.6%) could be used for identifying predictive factors for liver mass, the number of participants was insufficient for statistical analysis.

Table 7.

Participants with liver mass detected by ultrasound.

Participant ID	Details of the first liver mass findings	Month of visit when mass first detected	APF grade
HC16-0038	liver mass, size 11.61 x 6.17 cm.	0	3
KK03-0088	liver mass, size 2.32 x 1.69 cm.	0	2
NP01-0083	mass at right hepatic duct, size 1.41 cm (diameter).	0	0
TM11-0106	liver mass at segment 3, size 7.94 cm and segment 7, size 5.24 cm.	0	3
NP01-0134	Two liver masses at segment 7; sizes 1.87 cm and 1.68 cm.	12	2
DD09-0030	liver mass at segment 5, size 2.10 cm.	24	0
KN07-0098	liver mass at segment 7, size 1.35 cm and gallstone, size 0.62 cm.	24	0
HC08-0047	liver mass at segment 6, size 1.55 cm.	48	2
NH01-0104	liver mass at segment 6, size 1.11 cm.	60	3

5. Discussion

Liver fluke infection induces several hepatobiliary abnormalities including cholangiocarcinoma (Sripa et al., 2018). Treatment of liver fluke can reduce the infection and morbidities of the liver and bile ducts (Mairiang et al., 1993). However, findings during long-term follow-up after praziquantel treatment have not been reported. Our five-year follow-up findings reported here confirmed that praziquantel treatment of OV infection leads to improvement of parasite-associated hepatobiliary abnormalities. About 90% of participants were cured of liver fluke infection or exhibited only light infection (<500 EPG). Overall, 30.8% of participants showed no relapse of APF over five years whereas 37.5% showed persistent APF. APF positively correlated with male sex, ratio of left lobe of liver/aorta, and intrahepatic duct stone. However, APF did not correlate with baseline EPG of the participants in the study cohort.

Advanced periductal fibrosis represented by severe increased periportal echoes of the liver parenchyma, as detected by ultrasonography, is a prominent hallmark of chronic opisthorchiasis. Mairiang et al. (1993) reported that mild to moderate periportal echoes (fibrosis) persisted in 26.4% of opisthorchiasis cases at 10 months following treatment with praziquantel compared to 41.7% at baseline. We report here that >38% of the enrolled opisthorchiasis cases presented persistent APF during the five years following PZQ treatment, and that only 30.8% of the study participants did not exhibit relapse to APF. The presence of relapsing and persistent fibrosis in greater than two thirds of the participants implies certain (unidentified) pathological processes continued even after clearance of infection with O. viverrini. Inflammation and fibrosis are consequences of proliferation, pro-fibrotic and pro-inflammatory secretions released from activated cholangiocytes (Banales et al., 2019). Following clearance of the parasite, the biliary epithelium apparently remains activated and continues to release mediators of fibrosis. Activation of the bile duct epithelium may result from liver fluke or other infectious agents, cholestasis, ischemia, and/or be of idiopathic origin (Banales et al., 2019). During opisthorchiasis, colonization of the biliary tract by Helicobacter pylori may be a second activator, given that O. viverrini is a reservoir of the bacteria (Deenonpoe et al., 2015). A majority of O. viverrini infected residents in liver fluke endemic regions in northeastern Thailand are co-infected with H. pylori (Deenonpoe et al., 2017). Periductal fibrosis in these individuals was significantly associated with carriage of CagA+ H. pylori (Deenonpoe et al., 2017) and, moreover, co-infection with O. viverrini and H. pylori may orchestrate the pathogenesis of opisthorchiasis (Sripa et al., 2017).

Given that APF is a background characteristic of opisthorchiasis-associated CCA (Chamadol et al; 2014; Sinawat and Hemsrichart, 1991), persistent APF has been implicated as a risk factor for bile duct malignancy (Sripa et al., 2009; 2018). The lesion of persistent APF may include activated fibroblasts that continuously release cell growth mediators and cytokines (Van Linthout et al., 2014; Yoshida et al., 2019). Inflammation is one of the major factors that can induces fibroblast activation (Van Linthout et al., 2014). We recently reported that APF was associated with upregulated phagocytic and proteolytic activities together with elevated ROS production by macrophages (Salao et al., 2018) and neutrophils (Salao et al., 2020) in opisthorchiasis. In addition, a phenotype of highly activated neutrophil functions also was significantly associated with persistent APF but not with APF-negative livers in post-PZQ treated cases (Salao et al., manuscript in preparation). This indicates that inflammation remains active during persistent APF. The activated fibroblasts in cases with persistent APF may produce a spectrum of growth factors and inflammatory cytokines such as EGF, FGF, TGF-β that induce epithelial cell growth, anti-apoptosis, and epithelial to mesenchymal transition leading to initiation of malignancy (Hao et al., 2019; Sirica, 2011; Yoshida et al., 2019; Van Linthout et al., 2014;). This may display desmoplastic characteristics of cholangiocarcinoma (Sirica, 2011; Rizvi and Gores, 2014; Banales et al., 2019).

Beyond the factors noted above that likely were linked to relapse to periductal fibrosis, chronic hepatitis, alcohol-related liver disease, non-alcoholic fatty liver, cholestasis inducing conditions such as primary sclerosing cholangitis and autoimmune liver diseases may also underlie “de-novo” liver fibrosis (Patel & Sebastiani, 2020). Here, however, rates of fatty liver detected after treatment were <2% (follow-up Year 1, 10/688 cases, 1.5%; Year 2, 9/576, 1.3%) and all were negative for periductal fibrosis. We have not calculated the rates for the follow-up years 3 to 5 as there were few such cases (Year 3, 6/57; Year 4, 0/52; Year 5, 1/104) (data not shown).

We observed that APF positively correlated with male sex, ratio of the left lobe of liver/aorta, and intrahepatic duct stone. The persistence of APF preferentially in males correlates positively with the higher incidence of CCA in men than women (Vatanasapt et al., 1990). Elevated fibrosis, specifically persistent fibrosis, can impede the bile flow and enhance ascending cholangitis leading to cholangiocarcinogenesis in chronic opisthorchiasis (Sripa et al., 2018). Intrahepatic duct stones are prevalent in Asian countries and is associated with several factors including cholestasis, infection, and anatomic abnormity of bile duct and bile metabolic defects (Ran et al., 2017; Lorio et al., 2020). The intrahepatic duct stone formation reported here may be a result of O. viverrini infection with liver fluke eggs precipitated with calcium bilirubinate and mucus in similar fashion to those found in gallstones during opisthorchiasis (Sripa et al., 2004) and clonorchiasis (Ran et al., 2017). Persistent APF may enhance cholestasis inducing cholangitis and orchestrate the stone formation (Ran et al., 2017). Given that intrahepatic duct stone is a risk factor for intrahepatic CCA (Choi et al., 2004), persistent APF may enhance CCA development through the stone formation in another way. The deviation in the ratio of liver left lobe to aorta size implicates the shrinkage of the liver due to inflammation leading impeded bile flow (Mairiang et al., 1992).

Concerning the findings of the study dealing with liver mass, too few cases were seen to enable informed comments; five years may not be insufficient for development of CCA to be detected by abdominal ultrasonography (Mairiang et al., 2006). A retrospective case-control study designed based on the findings from this study should be suitable for the identification of factors predictive of liver mass. There are heterogeneities of the pathological outcomes of O. viverrini infection. About 25% of infected individuals develop APF and only 1-2% of these persons develop CCA (Mairiang et al., 2012). Some people with lifelong infection with O. viverrini fail to develop clinical symptoms of the malignancy (Sripa et al., 2012). Heterogeneous immune response and genetic mutations of oncogenic pathways have been proposed as an explanation on this heterogeneity in the appearance of APF and CCA following infection (Chan-On et al., 2013; Jiao et al., 2013; Onk et al., 2012; Jusakul et al., 2017).

Limitations of ultrasound alone in the detection of fibrosis, especially regarding sensitivity, user-dependent, intra-observer variability or other complicated liver and biliary diseases, such as intrahepatic duct stones. These limitations may be similar to those described in more specific imaging techniques for fibrosis scanning, i.e., imaging elastography (Patel & Sebastiani, 2020). However, ultrasound was the only feasible imaging method in the context of our study. Our study obviously obviated intra-observer variability given that only one of us, an experienced radiologist who has worked in this field for more than 30 years, performed the imaging.

Conclusion

This study provided an intervention to participants that alleviated the burden of opisthorchiasis. However, decreased liver fluke infection following treatment with praziquantel was not translated to elimination of the periductal fibrosis that had resulted from the parasite infection. Appropriate preventive measures to reduce the frequency of infection remain necessary and would include educating people to cook fish and enhanced hygienic toilet behavior (Tangkawattana & Sripa, 2018).

Highlights.

• Praziquantel treatment for opisthorchiasis alleviates parasite burden

• 30.8% of participants showed no relapse of hepatobiliary abnormality over five years

• 37.5% of participants showed persistent hepatobiliary abnormality