Hepatocellular carcinoma screening practices among patients with chronic hepatitis B by Canadian gastroenterologists and hepatologists: An online survey

Abstract

BACKGROUND

xpert guidelines recommend hepatocellular carcinoma (HCC) surveillance among patients with high-risk chronic hepatitis B (CHB); however, physician screening practices are often variable.

METHODS

An online survey of HCC screening practice was distributed to members of the Canadian Association for the Study of the Liver. Data were analyzed using appropriate statistical tests with p < .05 significance.

RESULTS

Of 71 respondents, 86% (n = 61) were gastroenterologists or hepatologists, and 72% (n = 51) reported having been in clinical practice for more than 5 years. A significant number of survey respondents performed HCC screening without consideration of concomitant non-alcoholic fatty liver disease (50.7%); non-Asian, non-African ethnicity (46.4%); and family history of HCC (28.6%). Most (67.6%) performed screening with ultrasound (US) at the time of specialty clinic visits, 28.2% had an automatic recall system, and only 2.8% referred back to primary care physicians to organize screening. More than half (54.9%) included alpha-fetoprotein in screening. Obstacles to screening included lack of an automatic recall system (42.9%), patient non-compliance (30.0%), and limited US/MRI access (17.1%).

CONCLUSIONS

HCC screening practices with hepatitis B patients vary widely among Canadian specialists, especially in unique populations with limited data to inform screening recommendations. Implementation of an automatic recall system could potentially increase HCC surveillance.

Introduction

Chronic hepatitis B (CHB) can progress to cirrhosis and complications of end-stage liver disease in as many as 20%–25% of individuals (1,2). The annual incidence of hepatocellular carcinoma (HCC) is estimated at less than 1% in non-cirrhotic individuals and at 2%–3% in cirrhotic individuals (3–5). Given the significant morbidity and mortality associated with chronic hepatitis B virus (HBV) infection, it is important for physicians to be familiar with screening recommendations for HCC. Expert guidelines recommend HCC surveillance with ultrasound (US) alone every 6 months for high-risk HBV patients, which include Asian men aged older than 40 years, Asian women aged older than 50 years, African Blacks aged older than 20 years, those with a family history of HCC, and those with cirrhosis (6). However, no specific recommendations are made regarding HCC surveillance for Caucasians and hepatitis B patients with non-alcoholic fatty liver disease (NAFLD). A recent Canadian consensus on the management of HCC highlighted gaps in knowledge and general underutilization of HCC surveillance (7).

A retrospective cohort study of cirrhotic patients diagnosed with HCC found a lack of surveillance in 75.7%, which was associated with a lower rate of early tumour detection (OR = 0.51) and worse overall survival (HR = 0.79) (8). Similarly, a meta-analysis of nine studies reported a pooled surveillance rate of only 18.4% for HCC among cirrhotic patients in the United States (9). The low rate of surveillance uptake was most commonly attributed to a lack of physician recommendation (10). Therefore, identifying reasons for the underutilization of HCC surveillance in clinical practice is important in designing interventions to increase HCC surveillance rates. In this study, we conducted an online survey of members of the Canadian Association for the Study of the Liver (CASL) and the Canadian Association of Gastroenterology (CAG) that included internists, gastroenterologists, hepatologists, and infectious disease specialists to examine specialist screening practices, mechanisms, and obstacles to screening for HCC among persons living with chronic HBV.

Methods

The survey questions and design were developed on the basis of expert input from members of the Canadian Hepatitis B Network and were based on recommended qualitative methods for survey design in health care epidemiology research (11,12). The survey was designed to characterize the survey respondents by self-reported clinical expertise and patient population, predominant screening practices and mechanisms, and major obstacles to recommended screening. The survey was pilot tested among a group of internal medicine and gastroenterology resident physicians before it was administered online. It was distributed to CAG and CASL members by means of an electronic mailing list. Data were collected, and statistical analysis using χ² and Fisher’s exact test was performed to determine whether there were any differences in screening practices, mechanisms, and obstacles based on medical specialty, years of practice, and number of HBV patients followed per year. We considered p < .05 statistically significant. The study was approved by the University of Calgary Conjoint Health Research Ethics Board (REB16-0041).

Results

Characteristics of survey respondents

A total of 71 physicians responded to the survey. The majority of respondents were hepatologists (n = 37; 52.1%) or gastroenterologists (n = 24; 33.8%), followed by infectious disease specialists (n = 4; 5.6%), general internists (n = 3; 4.2%), other specialists (n = 2; 2.8%), and family physicians (n = 1; 1.4%) (Figure 1A). The number of years respondents reported having been in clinical practice were 5 or fewer (n = 20; 28.2%), 6–10 (n = 12; 16.9%), 11–15 (n = 8; 11.3%), 16–20 (n = 7; 9.9%), and 21 or more (n = 24; 33.8%) (Figure 1B). The number of HBV patients encountered per year were fewer than 10 (n = 16; 22.5%), 10–99 (n = 27; 38.0%), 100–199 (n = 15; 21.1%), 200–499 (n = 4; 5.6%), and 500 or more (n = 9; 12.7%) (Figure 1C).

Figure 1:

Characteristics of survey respondents (N = 71): identified medical specialty (A), number of years in practice (B), and approximate number of HBV patients seen annually (C)

HBV = hepatitis B virus

Almost all physicians, 90.1% (n = 64), reported following their own HBV patients, with the exception of one family physician, one hepatologist, and five gastroenterologists. Hepatologists tend to treat and continue to follow their HBV patients (p = .077), but physicians who see more HBV patients per year are more likely to treat and follow their HBV patients long term (p < .001; Table 1).

Table 1:

Survey on HCC screening practices in chronic HBV patients among Canadian physicians

	Specialty				Years of practice				Approximate no. of HBV patients seen per year
Question	All	HEP	GI, ID	p	≤ 5	6–20	≥ 21	p	< 10	10–99	≥ 100	p
Do you treat and follow your HBV patients?				.077				.783				< .001
No	7 (9.9)	1 (2.7)	5 (17.9)		1 (5.0)	3 (11.1)	3 (12.5)		6 (37.5)	1 (3.7)	0 (0.0)
Yes	64 (90.1)	36 (97.3)	23 (82.1)		19 (95.0)	24 (88.9)	21 (87.5)		10 (62.5)	26 (96.3)	28 (100.0)
Do you have access to transient elastography (Fibrocan)?				.030				.121				.057
No	6 (8.5)	0 (0.0)	4 (14.3)		4 (20.0)	1 (3.7)	1 (4.2)		3 (18.8)	3 (11.1)	0 (0.0)
Yes	65 (91.5)	37 (100.0)	24 (85.7)		16 (80.0)	26 (96.3)	23 (95.8)		13 (81.3)	24 (88.9)	28 (100.0)
Which would best describe your HCC screening practice?				.087				.937				.244
Automatic recall of patients for US every 6 mo	20 (28.2)	13 (35.1)	5 (17.9)		7 (35.0)	7 (25.9)	6 (25.0)		4 (25.0)	9 (33.3)	7 (25.0)
Follow patients and order US during clinic visits every 6 mo	48 (67.6)	24 (64.9)	21 (75.0)		13 (65.0)	19 (70.4)	16 (66.7)		10 (62.5)	17 (63.0)	21 (75.0)
Refer back to primary care physician to take over HCC screening	2 (2.8)	0 (0.0)	2 (7.1)		0 (0.0)	1 (3.7)	1 (4.2)		2 (12.5)	0 (0.0)	0 (0.0)
None of the above	1 (1.4)	0 (0.0)	0 (0.0)		0 (0.0)	0 (0.0)	1 (4.2)		0 (0.0)	1 (3.7)	0 (0.0)
Do you include AFP with US screening?				.969				.327				.088
No	32 (45.1)	17 (45.9)	13 (46.4)		11 (55.0)	13 (48.1)	8 (33.3)		4 (25.0)	16 (59.3)	12 (42.9)
Yes	39 (54.9)	20 (54.1)	15 (53.6)		9 (45.0)	14 (51.9)	16 (66.7)		12 (75.0)	11 (40.7)	16 (57.1)
How do you screen non-Asian non-African HBV patients?				.179				.107				.910
Missing	2	0	1		1	1	0		1	1	0
Same as Asian population	32 (46.4)	16 (43.2)	13 (48.1)		7 (36.8)	9 (34.6)	16 (66.7)		7 (46.7)	11 (42.3)	14 (50.0)
Screen only cirrhotic patients	12 (17.4)	10 (27.0)	2 (7.4)		4 (21.1)	6 (23.1)	2 (8.3)		2 (13.3)	5 (19.2)	5 (17.9)
US every 12 mo, start at older age	2 (2.9)	0 (0.0)	2 (7.4)		1 (5.3)	0 (0.0)	1 (4.2)		0 (0.0)	0 (0.0)	2 (7.1)
US every 12 mo, start at same age	6 (8.7)	3 (8.1)	3 (11.1)		0 (0.0)	4 (15.4)	2 (8.3)		1 (6.7)	3 (11.5)	2 (7.1)
US every 6 mo, start at older age	17 (24.6)	8 (21.6)	7 (25.9)		7 (36.8)	7 (26.9)	3 (12.5)		5 (33.3)	7 (26.9)	5 (17.9)
How do you screen young (age < 40) non-African HBV patients with a family history of HCC?				.660				.227				.092
Missing	1	0	1		1	0	0		1	0	0
Same as without a family history of HCC	20 (28.6)	10 (27.0)	6 (22.2)		7 (36.8)	6 (22.2)	7 (29.2)		6 (40.0)	10 (37.0)	4 (14.3)
US and AFP every 6 mo, regardless of age	17 (24.3)	11 (29.7)	5 (18.5)		3 (15.8)	9 (33.3)	5 (20.8)		2 (13.3)	4 (14.8)	11 (39.3)
US every 6 mo, regardless of age	24 (34.3)	13 (35.1)	11 (40.7)		6 (31.6)	8 (29.6)	10 (41.7)		5 (33.3)	11 (40.7)	8 (28.6)
US and AFP every 12 mo, regardless of age	4 (5.7)	1 (2.7)	2 (7.4)		3 (15.8)	0 (0.0)	1 (4.2)		2 (13.3)	0 (0.0)	2 (7.1)
US every 12 mo, regardless of age	5 (7.1)	2 (5.4)	3 (11.1)		0 (0.0)	4 (14.8)	1 (4.2)		0 (0.0)	2 (7.4)	3 (10.7)
How do you screen HBV patients with concomitant NAFLD?				.360				.602				.871
Same as other chronic HBV patients	36 (50.7)	19 (51.4)	15 (53.6)		8 (40.0)	14 (51.9)	14 (58.3)		9 (56.3)	12 (44.4)	15 (53.6)
US every 6 mo, regardless of age if have advanced hepatic fibrosis	24 (33.8)	15 (40.5)	7 (25.0)		10 (50.0)	7 (25.9)	7 (29.2)		5 (31.3)	9 (33.3)	10 (35.7)
US every 6 mo, regardless of age and stage of hepatic fibrosis	8 (11.3)	2 (5.4)	4 (14.3)		2 (10.0)	4 (14.8)	2 (8.3)		1 (6.3)	5 (18.5)	2 (7.1)
MRI every 6 mo, regardless of age if have advanced hepatic fibrosis	3 (4.2)	1 (2.7)	2 (7.1)		0 (0.0)	2 (7.4)	1 (4.2)		1 (6.3)	1 (3.7)	1 (3.6)
Which is the main obstacle to HCC screening in your practice?				.172				.617				.296
Missing	1	0	1		1	0	0		1	0	0
Lack of automatic recall system	30 (42.9)	12 (32.4)	16 (59.3)		10 (52.6)	9 (33.3)	11 (45.8)		10 (66.7)	12 (44.4)	8 (28.6)
Limited accessibility to US, CT, or MRI at recommended intervals	12 (17.1)	8 (21.6)	3 (11.1)		1 (5.3)	7 (25.9)	4 (16.7)		2 (13.3)	3 (11.1)	7 (25.0)
Patient non-compliance	21 (30.0)	12 (32.4)	7 (25.9)		6 (31.6)	9 (33.3)	6 (25.0)		3 (20.0)	9 (33.3)	9 (32.1)
Other	7 (10.0)	5 (13.5)	1 (3.7)		2 (10.5)	2 (7.4)	3 (12.5)		0 (0.0)	3 (11.1)	4 (14.3)

* p value was based on χ ² or Fisher’s exact test as appropriate.

HCC = hepatocellular carcinoma; HBV = hepatitis B virus; HEP = hepatology; GI = gastroenterology; ID = infectious disease; US = ultrasound; AFP = alpha–fetoprotein; NAFLD, non-alcoholic fatty liver disease; CT = computed tomography

Summary of screening practices

For CHB patients with concomitant NAFLD (Figure 2A), 50.7% (n = 36) of physicians screen them the same way as patients without NAFLD (i.e., on the basis of age cut-offs, ethnicity, and presence of cirrhosis), and 11.3% (n = 8) screen with US every 6 months regardless of age or hepatic fibrosis stage. For CHB patients with concomitant NAFLD and advanced fibrosis (at Metavir Stage F3 or higher), 33.8% (n = 24) of physicians screen with US every 6 months, and 4.2% (n = 3) screen with MRI every 6 months, regardless of age.

Figure 2:

Summary of HCC screening practice by survey respondents in special patient groups in which the evidence to base screening decisions is limited: HBV and concomitant NAFLD (regardless of age) (A), family history of HCC (regardless of age) (B), and non-Asian, non-African HBV patients (C)

HCC = hepatocellular carcinoma; HBV = hepatitis B virus; NAFLD = non-alcoholic fatty liver disease; F3 = advanced fibrosis; Q6M = every 6 months; Q12M = every 12 months

For non-African CHB patients with a family history of HCC (Figure 2B), 28.6% (n = 20) of physicians screen them the same way as patients without a family history of HCC. In addition, 58.6% (n = 41) screen with US every 6 months, either with (n = 17) or without (n = 24) the addition of alpha-fetoprotein (AFP), regardless of age. Moreover, a significant number only do annual US screening for all ages (n = 9; 12.8%), either with (n = 4) or without (n = 5) addition of AFP. There were only 70 respondents; 1 respondent did not respond to the survey question. Physicians who see more HBV patients per year tend to take into account the family history of HCC as a risk factor when informing HCC screening practice (p = .092; Table 1).

For non-Asian, non-African CHB patients (Figure 2C), 46.4% (n = 32) of survey respondents use the same screening recommendations as for Asian CHB patients, 24.6% (n = 17) screen with US every 6 months but starting at an older age, 11.6% (n = 8) screen with annual US, and 17.4% (n = 12) screen only cirrhotic patients. There were only 69 respondents; 2 respondents did not respond to this survey question. Physicians who have been in practice longer tended to screen all patients similarly regardless of ethnicity (p = .107; Table 1).

Summary of screening modalities, access to transient elastography (Fibroscan), and screening obstacles among survey respondents

With regard to fibrosis assessment and modality of screening, all respondents used US for screening, but 54.9% (n = 39) continued to include AFP along with US. Almost all physicians, 91.5% (n = 65), have access to the non-invasive method of fibrosis assessment (i.e., stiffness measurement by transient elastography, or Fibroscan). As expected, hepatologists and other specialist physicians who see more HBV patients per year tend to have more ready access to transient elastography (ps = 0.03 and .057, respectively; Table 1).

Regarding HCC screening mechanisms (Figure 3A), the majority of respondents (n = 48; 67.6%) would order US at the time of scheduled clinic visits, whereas 28.2% (n = 20) have an automatic recall system, and 2.8% (n = 2) refer their patients back to their primary care physician to coordinate screening. More hepatologists (35.1%) used an automatic recall system (i.e., US every 6 mo) compared with 17.9% of gastroenterologists and infectious disease specialists (p = .087; Table 1).

Figure 3:

Summary of HCC screening strategies in patients with hepatitis B reported by members of Canadian Association for the Study of the Liver (A) and identified obstacles to following HCC screening guidelines (B)

HCC = hepatocellular carcinoma

With respect to potential obstacles to HCC screening (Figure 3B), 42.9% (n = 30) of respondents included the lack of an automatic recall system, 30.0% (n = 21) reported patient non-compliance, and 17.1% (n = 12) had limited access to US or MRI. Other unspecified barriers were reported by 10.0% (n = 7) of respondents. There were only 70 respondents; 1 respondent did not respond to this survey question.

Discussion

The Canadian data on HCC screening practices and potential barriers for optimal screening are limited. Our survey of 71 physician members of CASL and CAG highlights significant variability in screening and inconsistency in following published guidelines, even among specialists. Across Canada, barriers to effective screening included access to diagnostic imaging and lack of an automated screening program. Increased awareness of and education about screening guidelines may be needed, especially for physicians who tend to see fewer hepatitis B patients in clinical practice.

Additional studies are needed regarding use of AFP and limitations of US screening in patients with concomitant NAFLD. The current guidelines recommend HCC surveillance with US alone every 6 months because both US alone and US and AFP have led to similar rates of curative treatment with no appreciable statistical differences between the two strategies (13). However, some studies have shown that combining AFP with US has additional benefits. One randomized controlled study in China showed that screening with AFP and US every 6 months reduced HCC mortality by 37% among CHB patients (14). Despite the guideline recommendation, it is interesting to see that slightly more than half of the physicians in this study included AFP with their every-6-month US screening. This could be due to the recognition by specialists that US is operator dependent, and diagnostic accuracy may be limited by obesity, ascites, and liver nodularity (15). According to one meta-analysis, the sensitivity of US was only 47% for early-stage HCC and 84% for any-stage HCC (16). In addition, the effectiveness of HCC surveillance among patients with NAFLD, especially CHB patients with NAFLD, remains unclear. First, the incidence of HCC among patients with NAFLD in the presence of cirrhosis is unknown (17). Second, HCC may develop in patients with NAFLD in the absence of cirrhosis (18,19). Third, knowledge of the pathways and mechanisms of hepatocarcinogenesis among patients with NAFLD is limited (19). Finally, clinical tools for risk stratification of HCC in patients with NAFLD are lacking (19). The current hepatitis B guidelines do not mention any specific considerations in HCC surveillance among CHB patients with coexisting NAFLD (16,20). This uncertainty and the lack of evidence in the literature are reflected by the variable HCC screening practices among our survey respondents for CHB patients with concomitant NAFLD. Future studies are needed to better assess the efficacy of HCC surveillance among CHB patients with NAFLD.

A family history of HCC is a well-known risk factor for an increased risk of developing HCC (21). At the time of this study, the guidelines recommended screening this group every 6 months with US, but they did not specify at what age to start screening. Without specific recommendations, the responses in this study were quite variable; about half of the physicians would start screening with US (with or without AFP) every 6 months, regardless of age. Recently, the most updated Canadian hepatitis B guidelines recommended screening CHB patients with a positive family history of HCC starting at age 40 years, regardless of gender (20).

An automatic recall system was cited as a mechanism used for HCC screening by 28% of physicians in this study, and 42% said the lack of an automatic recall system was an obstacle to HCC screening. Although not specifically studied in HCC surveillance protocols, an automated follow-up reminder system has been shown to improve patients’ adherence to surveillance colonoscopy examinations (22). Implementation of such a recall system could potentially increase the uptake rate of HCC surveillance among physicians. However, a retrospective cohort study in which only 20% of patients with HCC had undergone surveillance found that the low rate was mainly attributable to provider factors, with failure of physician to recognize liver disease or cirrhosis in 39% and failure of physician to order surveillance in 38%, rather than system and patient factors (10). In fact, studies have demonstrated high levels of patient acceptance for HCC surveillance (23). However, in this study, it is interesting that 30% of survey respondents perceived patient non-compliance as an obstacle to HCC screening.

Limitations of this study include the small sample size, which may have reduced the statistical power of our data analysis. A de novo survey had to be developed because there is no previously validated or published questionnaire addressing HCC screening practices for patients with chronic HBV. In addition, the specific survey questions were necessarily brief and anonymous to maximize survey response and completion rate. Compared with investigator-administered face-to-face (i.e., interviewer) surveys, online questionnaires limit investigators’ ability to clarify data and to collect detailed, high-quality, and systematic data (24). Moreover, the majority of our survey respondents were either gastroenterologists or hepatologists; hence, our study findings might not generalize to the primary care setting in which most of HCC surveillance takes place. Overall, with more than 70 respondents the study is likely a representative sample of CASL members and specialists who care for patients with hepatitis B in Canada.

Conclusion

HCC screening practices among patients with hepatitis B vary widely among Canadian CASL specialist members (gastroenterologists and hepatologists). Additional research on HCC risk is needed to guide screening recommendations for specific groups (i.e., hepatitis B patients with concomitant NAFLD, family history of HCC, or non-Asian and non-African ethnic background). Implementation of an automatic recall system could potentially improve patient adherence and compliance with HCC surveillance protocols.

Ethics Approval:

The study protocol was approved by an ethics committee and the ethics certificate information is available from the authors upon request.

Informed Consent:

Informed consent was obtained from the patient(s).

Registry and Registration No. of The Study/Trial:

N/A

Funding:

None to declare.

Disclosures:

Dr Coffin reports grants from Gilead Sciences, grants from GSK, and personal fees and other from Springbank, outside the submitted work.

Peer Review:

This article has been peer reviewed.