Summary

Objective: To review the current hepatocellular carcinoma screening tests in hepatitis B virus (HBV) carriers in primary care, aiming at standardising and improving patient management, as well as increasing cost-effectiveness.

Design: A retrospective non-interventional study of screening methods in HBV carriers.

Subjects: Patients attending the Family Medicine Training Centre of Prince of Wales Hospital between 1 July 2003 and 3 October 2003 for the follow-up of hepatitis B.

Main outcome measures: Follow up frequencies, intervention frequencies (AFP, ALT, USG livers), presence or absence of HCC/cirrhosis.

Results: 282 patients were included in the study. The mean frequencies for interventions were: (1) follow up consultation: once every 7 months; (2) AFP: once every 12 months; (3) ALT: once every 11 months; (4) USG: once every 22 months. Five new occurrences of cirrhosis were detected by USG over a four-year period. No associated AFP/ALT changes were noted. The 14 cases with raised ALT and AFP had normal ultrasound findings. Cirrhosis was an important determinant in increasing intervention frequency. The sensitivity and specificity of detecting cirrhosis with AFP and ALT were both 0.08 and 0.97 respectively.

Conclusion: Data from this analysis support the use of USG alone in monitoring HBV carriers in the primary care setting. Biochemical investigations play a minimal role. Further studies will help to establish the feasibility of screening by USG alone

Keywords: Hepatitis B, cirrhosis, hepatocellular carcinoma, screening, Hong Kong

摘要

目的：就目前基層醫生對乙型肝炎病毒（HBV）攜帶者的肝細胞性肝癌篩查試驗進行回顧，旨在改善管理，使其標準化，並提高成本效益性。

設計：對HBV攜帶者的篩查方法的回顧性、非干預性研究。

對象：2003年7月1日到2003年10月3日期間到威爾斯親王醫院家庭醫學培訓中心覆診的乙型肝炎的病人。

測量內容：覆診頻率，干預頻率（AFP、ALT、肝臟超聲波，是否出現肝細胞性肝癌或肝硬化。結果：本研究包括了282名病人。進行干預的平均頻率為：(1)覆診：每7個月1次；(2)AFP：每12個月1次；(3)ALT：每11個月1次；(4)USG：每22個月1次。在4年之內通過USG發現了5例新發肝硬化。未發現AFP/ALT有相關改變。14例ALT和AFP升高的病人超聲波檢查正常。肝硬化是干預頻率增加的重要決定因素。用AFP和ALT檢測肝硬化的敏感度和特異度分別為0.08和0.97。

結論：本研究支持在基層醫療機構單獨採用超聲波（USG）監測乙型肝炎病毒攜帶者。生化檢查的作用很小。進一步的研究可以確定單獨用超聲波篩查的可行性。

主要詞彙：乙型肝炎，肝硬化，肝細胞性肝癌，篩查，香港

Introduction

Cancer is the leading cause of death in Hong Kong. Among various causes of cancer deaths, liver cancer ranked second in male (15.7%) and fourth in female (8%). The outcome of liver cancer remains poor with high mortality, the incidence ratio being around 0.9. In the year 2000, there were 1584 new cases of liver cancer and 1424 deaths from liver cancer.¹

Hepatitis B virus (HBV) infection is an important aetiological factor in hepatocellular carcinoma (HCC). Studies have shown that 85.3% to 91.6% of symptomatic HCC cases had evidence of previous HBV infection in Hong Kong.^2,4 Furthermore, it has been recognised that the estimated risk of HBV carriers for developing HCC is 100 times higher than non-HBV carriers.

There were attempts to achieve early detection of HCC, with the hope of improving its management outcome. In an alpha-fetoprotein (AFP) screening programme conducted in Hong Kong, 1.34% of Chinese patients with chronic HBV infection seen over a 5-year period developed HCC.³ All cases were picked up under the programme. The figures were similar in some Taiwan studies.

There is as yet no consensus on the optimal screening programme for HCC in HBV carriers. Information given to the public often cites "regular alpha-fetoprotein and ultrasonography (USG) screening, though this is not 100% reliable".⁵ USG and AFP can detect asymptomatic HCC,^6-8 but the latest systematic review has exposed the fact that there are still no adequate data to support or refute screening in HBV carriers.⁹ In those with cirrhosis, however, it has been recommended that screening be performed every 6 months with AFP and USG,¹¹ since 80% of HCCs arise in individuals with a cirrhotic liver. The sensitivity and specificity of these two investigations at 6-montly intervals have already been analysed. AFP (>20mg/L) had a sensitivity of 41-65% and a specificity of 80-94%,¹⁵ while USG had values of 79% and 94% respectively.¹⁶ A recent cost-benefit analysis of screening Hong Kong recommended only USG for HCC detection, as AFP (six-monthly or yearly) is more expensive and less effective.¹⁰

The latest Asia-Pacific consensus¹⁷ on the management of chronic hepatitis B (with negative HBeAg) recommended the followings:

• 6 monthly follow up
• USG and AFP every 3-6 months in high risk patients (male, >40yrs of age, cirrhotics, positive family history of serious liver disease)
• Treatment if there is active HBV replication and raised ALT
• Those with persistently normal ALT need adequate follow up and HCC surveillance every 3-6 months

In our locality, it has been a common practice to routinely check patient's AFP levels, alanine transaminase levels (ALT), and regular USG for cirrhosis/HCC detection.

Purpose

This study aims to review the current HCC screening tests ordered by primary care doctors for HBV carriers, with a view to standardise and improve patient management, in addition to enhancing cost-effectiveness.

Methods

This was a retrospective study looking at case notes of patients who had attended the family medicine clinic (Prince of Wales Hospital, Shatin, Hong Kong) over the three-month period from 1st July 2003 to 3rd October 2003 for the follow up of hepatitis B. These patients have either been referred from the hepatology clinic of the same hospital for "continuation of management", or identified during follow up for other diseases, such as diabetes, hypertension. Their status as a HBV carrier has previously been confirmed with a positive HbsAg for six months. All patients are HBeAg -ve and -ve for hepatitis C.

Information starting from the first consultation (dating back to 1999 - the year when clinic was set-up) for hepatitis B at our clinic were gathered including:

1. Age
2. Gender
3. Follow up frequency
4. Frequency of blood taking for AFP (the cut-off value for abnormal AFP at our laboratory is >7mgL)
5. Frequency of blood taking for ALT
6. Frequency of USG screening
7. Presence or absence of cirrhosis/HCC prior to clinic attendance8. Development of cirrhosis/HCC during follow up

The exclusion criteria included:

1. Patients who had defaulted for any intervention/FU.
2. Those who were HBV carriers but not followed up at our clinic for hepatitis.
3. Those who were first seen during the three months inclusion time period.
4. Patients who required more regular FUs or liver function testing (LFT) because of other coexisting diseases.

Results

A total of 351 patients with the diagnosis of HBV carrier status had attended the clinic within the three-month period. Sixty-nine of these were either defaulters, not attending our clinic for following up of their hepatitis status, with co-morbidities or were seen as a new case during this period (see Figure 1). The follow up and intervention frequencies as well as their standard deviations are shown in Tables 1 and 2.

Of the 282 patients included in the study, all five new occurrences of cirrhosis were detected by USG. There were no associated changes in AFP/ALT values in these newly detected cirrhotics. Two cases with a raised AFP (>7mcg/L) and two cases with raised ALT already had cirrhosis prior to first attendance. There were seven other cases with raised AFP (>7mcg/L) and seven with raised ALT, none of whom had ultrasonographic evidence of abnormal liver changes (Tables 3 and 4). No HCC were detected amongst the 282 patients. The relationship between different modalities and cirrhotic outcome are illustrated in Figures 2-5.

There were no significant differences (P=0.38) in follow up frequencies in those with and without cirrhosis. There were, however, significant differences between usage frequency of the three screening modalities and the presence/absence of cirrhosis (AFP, P=0.0076; ALT, P=0.024; USG, P=0.022). The sensitivity, specificity, positive and negative predictive values of detecting cirrhosis with AFP and ALT are shown in Table 5 (diagnosis of cirrhosis from USG findings).

Discussion

Variations in current practice

The above results show that the presence of cirrhosis was associated with a significantly more frequent investigation rate. Follow up frequency was, however, not affected. There was a wide variability in the FU and test intervals, ranging from 1 to 14 months for FU (one month FU was in fact given to a patient who was over anxious about her condition), 2 to 36 months for blood tests, and 6 to 60 months for USG. This practice differs significantly from the recent consensus of HBV management.¹⁷ Doctor's perception of the problem, patient's expectations and demands, and patient's perception of disease severity may all contribute to this variability.

Use of AFP as a screening tool for cirrhosis/HCC

The specificity of AFP (at a cut-off value of >7mg/L) is comparable to that from other studies.¹⁵ Population bias may be the key in the marked difference in sensitivity in this study, since most of our cases are stable HBV carriers without complications. However, it must be noted that in the primary care setting most cases will be stable HBV carriers. Those with high risks (see introduction) or complications may already have been channeled to appropriate specialist care. Therefore if this study population reflects the actual population in the primary care setting, then AFP screening (with a cut-off value of >7mg/L) would be futile in view of such a low sensitivity. Furthermore, if we take the standard cut-off value of >20mg/L, the sensitivity will be even lower, further arguing against the use of AFP.

Role of ALT

ALT has not been recommended for the screening of cirrhosis or HCC in HBV carriers. However, since a raised ALT is one of the essential criteria for medical treatment, its use in the management of HBV carriers remains crucial. The recent Asia-Pacific consensus stated that regular surveillance for HCC should be performed in those with persistently normal ALT. There were no guidelines on the use of ALT in this "surveillance" strategy. Should ALT be measured regularly? Data from this analysis show that ALT measurement had a low sensitivity, making this test unsuitable for detection of HCC/cirrhosis in a general population basis.

Cost-effectiveness

The costs of the individual investigations are listed in Table 6.¹² From our current practice (taking the mean investigation frequency from above), the average cost per patient per year for the routine monitoring (ALT, AFP and USG) of HBV carriers without cirrhosis amounts to HK$2300 [$200 + ($400 x 12/13) + ($1000 x 12/7)]. This has not taken into account the additional costs of running an outpatient department and its associated staff, time taken off by the patient for tests and psychological burdens. From this study, 52.6 persons need to be tested before detecting one cirrhosis over 4 years (5 new cirrhotics in 263 cases in 4 years), thus translating to a cost of HK$480,000 per person per year ($2300 x 263 x 4/5). The cost will no doubt rise in order to detect HCC, since this complication is much less common.

If USG is used as the sole monitoring investigation, the cost per person per year per cirrhosis detected comes to HK$360,000 [$1000 x 12/7 x 263 x 4/5], a 25% reduction. If investigations are performed according to the recent Asia-Pacific consensus (6 monthly USG and AFP), the cost comes to a staggering HK$590,000 [{($400 x 2) + ($1000 x 2)} x 263 x 4/5] (Table 7).

Finally, this study was limited by low patient numbers, short reviewing period and lack of new HCC cases. One further drawback is the definition of cirrhosis using USG findings. This will inherently bias results towards USG screening. However, it would be impossible and unethical to perform biopsies on all subjects for the detection of cirrhosis/HCC.

Conclusion

Half-yearly and yearly USG alone have already been shown to be both cost-effective and comparable to other cancer screening programmes such as cervical cancer¹³ and breast cancer.¹⁴ Data from this analysis support the use of USG alone in screening HBV carriers in the primary care setting, both in terms of cost and effectiveness. Furthermore it illustrates and emphasises that haematological investigations play a minimal role in the monitoring of HBV carriers in the primary care setting, especially in those without cirrhosis. Although our management in part is influenced by patient's ideas, concerns and expectations, adequate explanation and reassurance will help to correct their misunderstandings and allay their fears. To provide better patient care in modern medicine, one must rationally utilise our limited resources. Further studies involving a larger number of patients and more primary care centres over a longer period will help to establish the feasibility of USG alone as a screening test and to determine different management strategies in different centres, with a view to provide a more consistent and reliable care for this common and potentially life-threatening disease.

Key message

1. There is as yet no consensus on the optimal screening programme for HBV carriers.
2. A large variation in frequency of haematological investigations, ultrasound imaging and follow ups were noted in our practice.
3. Presence of cirrhosis was an important determinant in increasing intervention frequency.
4. Both ALT and AFP testing showed a low sensitivity in detecting cirrhosis, arguing against their use in screening HBV carriers.
5. Half-yearly to yearly USG alone appears to be the most cost-effective method of screening.