• Open Access

Markers of hepatitis B virus infection and immunity in Victoria, Australia, 1995 to 2005

Authors


Correspondence to:
Dr Benjamin Cowie, Victorian Infectious Diseases Reference Laboratory, 10 Wreckyn Street, North Melbourne Victoria 3051. Fax: (03) 9342 2666; e-mail: benjamin.cowie@mh.org.au

Abstract

Objective: Estimating the prevalence of chronic hepatitis B virus (HBV) infection in generally low-prevalence populations containing communities with a higher disease burden is difficult. This study was conducted to estimate the prevalence of serological markers of infection with, and immunity to, HBV in the Victorian population and to analyse trends in these estimates over time.

Methods: A serological survey of 3,212 samples of convenience collected in the years 1995, 2000 and 2005 was conducted using a selection procedure designed to reduce selection bias. All samples were tested for hepatitis B surface and core antibodies; all core antibody positive samples (indicating previous infection) were then tested for the presence of hepatitis B surface antigen (HBsAg).

Results: HBsAg prevalence was 1.1% (95%CI 0.8-1.6%) with significant differences observed by area of residence, age, gender and test year. Serological evidence of immunisation in infants and adolescents were lower than established estimates following the introduction of universal vaccination for these groups.

Conclusions: This study emphasises the significant and growing problem of chronic HBV infection in Victoria and suggests lower than expected population immunity deriving from universal vaccination programs.

Implications: Greater efforts are needed to formulate a comprehensive public health response to address this relatively neglected blood borne viral infection, the burden of which is very significant in some marginalised sections of our community. Increased attention to improving the universality of our immunisation programs is also needed.

The epidemiology of hepatitis B virus (HBV) infection in Australia is complex and the burden of disease is poorly defined. Australia is a multicultural nation with more than one quarter of all Australians having been born overseas,1 many in regions with intermediate (2-8%) and high (>8%) population hepatitis B surface antigen (HBsAg) prevalence, countries that together account for nearly 90% of the global population.2–4 It has long been recognised that migrants from these areas have a higher prevalence of chronic HBV than do people born in Australia5–11 with the important exception of Indigenous Australians who also bear a disproportionate burden of chronic HBV infection and its complications.12,13

Current estimates of the population prevalence of HBV infection in Australia are the product of a limited number of studies.6,11,14 A clearer understanding of the burden of chronic HBV infection is required to inform public health interventions including targeted screening and vaccination programs. These data are also vital to allow planning for the significant resources required for the treatment of those chronically infected, without which many will develop cirrhosis and/or hepatocellular carcinoma or require liver transplantation.2,15

The Victorian HBV Serosurvey 1995-2005 was undertaken to investigate the prevalence of markers of infection with, and immunity to, HBV in an age-structured randomised sample of convenience drawn from archived residual serum following diagnostic testing. The serum samples were held at the Victorian Infectious Diseases Reference Laboratory (VIDRL), a state reference laboratory and public health virology research facility in Melbourne, Victoria. These samples were originally submitted for a variety of serological tests, predominantly by private and public laboratories throughout Victoria, and also directly from a number of primary care clinics in Melbourne. We aimed to estimate the prevalence of chronic HBV infection in Victoria, to analyse secular trends in estimates of infection, and to assess the serological evidence of immunity following the introduction of universal vaccination programs in Victoria.

Methods

The study was conducted in 2006/07 using sera collected in 1995, 2000 and 2005. A unique software interface was used to select samples for inclusion in the serosurvey from the stored serum archive at VIDRL. Any samples from a patient ever referred for serologic testing for HBV, hepatitis C virus (HCV) or human immunodeficiency virus (HIV) at VIDRL were excluded, with remaining candidate samples randomly selected in proportion to the age and gender distribution of the Victorian population for each test year.

Sample size calculations for precision were performed with an estimated HBsAg prevalence of 0.5% derived from a recent Australian serosurvey.6 Using 95% confidence intervals for the estimate of +/− 0.4% for each of the three years, an initial target of 1,200 samples from each year was determined. Using this sample size, corresponding 95% confidence intervals for precision around hepatitis B surface antibody (anti-HBs) and hepatitis B core antibody (anti-HBc) estimates derived from the same study were 28.7%+/− 2.6% and 6.9%+/− 1.4% in each test year.

Stored frozen serum samples were thawed and aliquots transferred into new specimen tubes identified only with a unique study number. All identifying information other than test year, age at time of test, gender and postcode (where available) was permanently delinked prior to testing. Missing or inadequate volume samples were replaced by age- and gender-matched samples wherever possible, using the same retrieval methodology.

All samples were tested for the presence of anti-HBs and anti-HBc by enzyme immunoassay (EIA) using commercial kits (Murex Biotech/Abbott Diagnostics Division, Temple Hill, Dartford UK). Any sample with a positive, borderline, indeterminate or otherwise unknown anti-HBc result was tested for HBsAg, again using commercial Murex Biotech/Abbott Diagnostics kits. The EIA reactions were performed using a Triturus EIA Analyzer (Grifols S.A., Barcelona Spain). Fewer than 1% of all samples tested gave inconclusive or incomplete results for any parameter, usually due to insufficient volume to complete testing.

The pattern of serological markers enabled assignment of HBV status to individual samples (Table 1). Victorian postcodes were individually mapped to Australian Bureau of Statistics (ABS) divisions16 to enable analysis by geographic area. Information on population characteristics by statistical division for the 2001 and 2006 Australian censuses was obtained from resources available on the ABS website (http://www.abs.gov.au/websitedbs/d3310114.nsf/home/Census+data). Vaccine coverage data was obtained from the Australian Childhood Immunisation Register (ACIR) reports.17

Table 1.  Serological profiles used to define HBV infection status.
 Anti-HBsAnti-HBcHBsAg
Susceptible
Immunised+
Infected±++
Resolved++
Isolated anti-HBc+
Otherwhere any parameter missing or indeterminate

Statistical analysis was performed using Stata v8.2 (StataCorp, College Station, Texas US). Confidence intervals of 95% were calculated around seroprevalence estimates using the exact binomial method; p-values for differences in antibody prevalence were calculated by z test, and for differences in HBV status between groups were calculated by chi-square test. Direct standardisation of seroprevalence estimates relative to the Victorian or Melburnian populations was performed for age group, gender and geographic area to assess for the impact on prevalence estimates of differences between the sample and the source population. Ethics approval for the study was obtained from the Melbourne Health Human Research Ethics Committee (HREC 2005.096).

Results

Estimates of HBV prevalence

Complete results of the serosurvey by test year and age-group are presented in Tables 2a and 2b. Of the 3,212 samples tested over the 10 year period, 30.3% (28.7-31.9) were positive for anti-HBs, 9.4% (8.4-10.5) positive for anti-HBc, and 1.1% (0.8-1.6) positive for HBsAg. A steady increase from 1995 to 2005 was observed in both anti-HBs (17.4% to 40.5%) and anti-HBc (6.5% to 11.3%) (p<0.001 for both comparisons). HBsAg prevalence peaked at 1.9% in 2000 and dropped significantly in 2005 to 0.6% (p=0.005).

Table 2a.  Serosurvey results by age-group for test years 1995 and 2000.
Age group19952000
 No. tested% positive anti-HBs (95% CI)% positive anti-HBc (95% CI)% positive HBsAg (95% CI)No. tested% positive anti-HBs (95% CI)% positive anti-HBc (95% CI)% positive HBsAg (95% CI)
0 to 42835.7006446.93.11.6
  (18.6-55.9)(0-12.3)(0-12.3) (34.3-59.8)(0.4-10.8)(0-8.4)
5 to 9319.7006833.830
  (2.0-25.8)(0-11.6)(0-11.2) (22.8-46.3)(0.4-10.4)(0-5.3)
10 to 143713.5007157.72.90
  (4.5-28.8)(0-9.5)(0-9.5) (45.4-69.4)(0.3-9.9)(0-5.1)
15 to 197212.54.21.47925.38.92.5
  (5.9-22.4)(0.9-11.9)(0-7.6) (16.2-36.4)(3.6-17.4)(0.3-8.7)
20 to 248732.23.41.18031.311.32.5
  (22.6-43.1)(0.7-9.7)(0-6.2) (21.3-42.6)(5.3-20.3)(0.3-8.7)
25 to 298218.53.71.2883381.1
  (10.8-28.7)(0.7-10.4)(0-6.6) (23.3-43.8)(3.3-15.7)(0-6.2)
30 to 348325.3121.28027.510.12.5
  (16.4-36.0)(5.9-21)(0-6.5) (18.1-38.6)(4.5-19.0)(0.3-8.7)
35 to 398114.84.92.59330.1124.3
  (7.9-24.4)(1.4-12.2)(0.3-8.6) (21.0-40.5)(6.1-20.4)(1.2-10.8)
40 to 447415.19.51.48830.714.82.3
  (7.8-25.4)(3.9-18.5)(0-7.3) (21.3-41.4)(8.1-23.9)(0.3-8.0)
45 to 497317.89.608129.611.31.2
  (9.8-28.5)(3.9-18.8)(0-4.9) (20.0-40.8)(5.3-20.3)(0-6.7)
50 to 545619.69.107723.46.52.6
  (10.2-32.4)(3.0-20.0)(0-6.4) (14.5-34.4)(2.1-14.5)(0.3-9.1)
55 to 594511.18.92.2583110.50
  (3.7-24.1))(2.5-21.2)(0.1-11.8) (19.5-44.5)(4.0-21.5)(0-6.2)
60 to 644814.68.304714.912.86.4
  (6.1-27.8)(2.3-20.0)(0-7.4) (6.2-28.3)(4.8-25.7)(1.3-17.5)
65 to 69449.36.80348.85.92.9
  (2.6-22.1)(1.4-18.7)(0-8.0) (1.9-23.7)(0.7-19.7)(0.1-15.3)
70 to 743813.210.803917.915.40
  (4.4-28.1)(3.0-25.4)(0-9.3) (7.5-33.5)(5.9-30.5)(0-9.0)
75 to 79224.54.502814.314.30
  (0.1-22.8)(0.1-22.8)(0-15.4) (4.0-32.7)(4.0-32.7)(0-12.3)
80 and over254803525.725.70
  (0.1-20.4)(0.1-26.0)(0-13.7) (12.5-43.3)(12.5-43.3)(0-10.0)
Missing (% total) 351 061
  0.320.540.11 00.540.09
All ages92617.46.50.9111030.29.81.9
  (15.0-20.0)(5.0-8.3)(0.4-1.7) (27.5-33.0)(8.1-11.7)(1.2-2.9)
Table 2b.  Serosurvey results by age-group for test year 2005, and all test years combined.
Age group2005All years
 No. tested% positive anti-HBs (95% CI)% positive anti-HBc (95% CI)% positive HBsAg (95% CI)No. tested% positive anti-HBs (95% CI)% positive anti-HBc (95% CI)% positive HBsAg (95% CI)
0 to 47271.81.4016455.81.90.1
  (59.9-81.9)(0-7.7)(0-5.0) (47.9-63.6)(0.4-5.3)(0-3.4)
5 to 97946.82.6017835.42.30
  (35.5-58.4)(0.3-9.0)(0-4.6) (28.4-42.9)(0.6-5.7)(0-2.1)
10 to 14815810018949.25.30
  (46.5-68.9)(4.4-18.8)(0-4.6) (41.9-56.7)(2.6-9.6)(0-1.9)
15 to 197657.913.31.322732.28.91.8
  (46.0-69.1)(6.6-23.2)(0-7.1) (26.1-38.7)(5.5-13.4)(0.5-4.5)
20 to 248242.74.9024935.36.41.2
  (31.8-54.1)(1.3-12.0)(0-4.4) (29.4-41.6)(3.7-10.2)(0.2-3.5)
25 to 298144.47.41.3251326.41.2
  (33.4-55.9)(2.8-15.4)(0-6.8) (26.3-38.2)(3.7-10.2)(0.2-3.5)
30 to 348748.313.82.325034122
  (37.4-59.2)(7.3-22.9)(0.3-8.1) (28.1-40.2)(8.3-16.8)(0.7-4.6)
35 to 398645.318.62.326030.4123.1
  (34.6-56.5)(11.0-28.4)(0.3-6.4) (24.9-36.4)(8.3-16.6)(1.3-6.0)
40 to 448537.616.51.224728.513.81.6
  (27.4-48.2)(9.3-26.1)(0-6.4) (22.9-34.5)(9.7-18.7)(0.5-4.1)
45 to 498433.39.5023827.310.10.4
  (23.4-44.5)(4.2-17.9)(0-4.3) (21.8-33.4)(6.6-14.7)(0-2.3)
50 to 547638.211.8020927.89.11
  (27.2-50.0)(5.6-21.3)(0-4.8) (21.8-34.3)(6.6-14.7)(0.1-3.4)
55 to 597220.86.9017521.78.60.6
  (12.2-32.0)(2.3-15.5)(0-5.0) (15.8-28.6)(4.9-13.8)(0-3.1)
60 to 64522515.4014718.412.20.2
  (14.0-38.9)(6.9-28.1)(0-6.8) (12.5-25.6)(7.4-18.7)(0.4-5.8)
65 to 694621.717.4012413.810.50.8
  (10.9-36.4)(7.8-31.4)(0-7.7) (8.3-12.2)(5.7-17.3)(0-4.4)
70 to 743813.213.5011514.813.30
  (4.4-28.1)(4.5-28.8)(0-9.3) (8.9-22.6)(7.6-20.9)(0-3.2)
75 to 79348.814.70849.511.90
  (1.9-23.7)(5.0-31.1)(0-10.3) (4.2-17.9)(5.9-20.8)(0-4.3)
80 and over4522.224.4010519210
  (11.2-37.1)(12.9-39.5)(0-7.9) (12.0-27.9)(13.6-30.0)(0-3.5)
Missing (% total) 162 4174
  0.090.510.17 0.120.530.12
All ages117640.511.40.6321230.39.41.1
  (37.7-43.4)(9.5-13.2)(0.2-1.2) (28.7-31.9)(8.4-10.5)(0.8-1.6)

Males were significantly more likely to have been infected with HBV than females. Of males, 11.1% (9.6-12.7) were anti-HBc positive compared to 7.5% of females (6.2-8.8) (p<0.001). Males also appeared more likely to be chronically infected, with 1.3% being HBsAg positive (0.8-1.9) compared to 0.9% (0.4-1.4) for females but this difference was not significant (p=0.24).

HBV infection status by location for the 2,598 Victorian samples with complete postcode information available is shown in Table 3. Although the proportion of susceptible samples from outside Melbourne was significantly higher (p<0.001), the vaccination rates were not different (22.6 vs. 23.9%, p=0.44). The proportion of samples from patients ever infected with HBV was higher in Melbourne than in the rest of Victoria (12% vs. 2.8%, p<0.001), as was the proportion of HBsAg positive samples (1.5% vs. 0.3%, p=0.013).

Table 3.  HBV infection status by location for Victorian samples with complete postcodes.
 Metropolitan Melbourne (n = 1,944)Non-metropolitan Victoria (n = 654)
Susceptible63.4%74.5%
Immunised23.9%22.6%
Infected1.5%0.3%
Resolved7.6%1.2%
Isolated anti-HBc2.4%1.2%
Other1.2%0.2%

Marked diversity in anti-HBc and HBsAg prevalence across different areas of Melbourne was observed. Figure 1 shows the ABS statistical subdivisions within Melbourne that are coloured relative to the proportion of people born overseas in that area as determined in the 2006 Australian census.18 The numbers show the HBsAg prevalence for that area over all three test years. Although the association is not absolute, a striking example is that of the 56 samples from the region with highest proportion of people born overseas (56% on the 2006 census), 7.2% were HBsAg positive and 30.9% were anti-HBc positive.

Figure 1.

Percentage of serosurvey samples positive for HBsAg and proportion of residents born overseas in the 2006 Census by statistical subdivision, Melbourne. Map adapted from reference (18).

Table 4 shows that the proportion of the samples from subjects susceptible to infection reduced from nearly 80% in 1995 to 55% in 2005 – largely explained by the increase in the immunised proportion from 13.6 to 33% over this period, but also by the significant increase in patients with natural immunity following HBV infection, rising from 6.5% in 1995 to 11.3% in 2005 (p<0.001 for all comparisons). The proportion of samples with isolated anti-HBc was 2.2%.

Table 4.  HBV infection status by test year for age groups targeted in infant and adolescent vaccination programs and for all samples.
 0–4 years old (n = 164)10–14 years old (n = 189)All samples (n = 3,212)
 199520002005199520002005199520002005
Susceptible64.3%51.6%27.8%86.5%40.9%39.5%79.1%66.6%55.2%
Immunised35.7%45.3%66.7%13.5%54.9%49.4%13.6%23.2%33.0%
Infected0%1.6%0%0%0%0%0.9%1.9%0.6%
Resolved0%1.6%1.4%0%1.4%8.6%3.4%6.7%6.9%
Isolated anti-HBc0%0%0%0%1.4%1.2%1.9%1.1%3.3%
Other0%0%4.2%0%1.4%1.2%1.2%0.6%1.0%

Estimates of immunisation

Universal vaccination programs targeting 12-13-year-old adolescents were implemented in Victoria in 1998, and universal infant vaccination commenced throughout Australia in May 200017,19 in line with WHO recommendations since 1991.20 These programs are reflected in significant changes in anti-HBs prevalence in the relevant age-groups. As seen in Table 2, anti-HBs prevalence in 0-4 year olds rose from 35.7% (18.6-55.9) in 1995 to 71.8% (59.9-81.9) in 2005 (p=0.001). For children aged 10-14, anti-HBs positivity rose from 13.5% (4.5-28.8) in 1995 to 57.7% (45.4-69.4) by 2000 (p<0.001), with no further increase in 2005.

The relative contribution of vaccination and resolved infection to estimates of population immunity can only be assessed by examining the complete serological profile of each sample. HBV status for the samples drawn from age-groups targeted by universal vaccination programs are shown by test year in Table 4. The proportion of 0-4 year olds that were susceptible to HBV infection fell from 64.3% in 1995 to 27.8% in 2005 (p<0.001). Although this is largely explained by an increase in the immunised proportion (anti-HBs positive only) from 35.7% to 66.7% over this time period (p<0.001), there was also one HBsAg positive three year-old in the 2000 test group, and one child with resolved infection in both 2000 and 2005.

From 1995 to 2000 the proportion of 10-14 year olds susceptible to HBV infection fell significantly from 86.5% to 40.9% (p<0.001), with a further non-significant reduction to 39.5% in 2005 (p=0.81). Although, again, mostly due to immunisation with isolated anti-HBs positivity rising from 13.5 to 54.9% in 2000 (p<0.001), the immunised proportion dropped non-significantly in 2005 (p=0.32). Immunity due to resolved infection was unexpectedly high with 10% of samples anti-HBc positive in 2005, compared with 2.9% in 2000 (p=0.08) and none in 1995 (p=0.046). No samples in this age-group were positive for HBsAg.

Most notified acute HBV infections in Australia occur in people between the ages of 20 and 45 years.21,22 From 1995 to 2005, vaccination coverage increased from 18.4% to 36.1% across these age groups, with the proportion of samples coming from patients remaining susceptible to infection falling from 74.2% to 51.5% (p<0.001 for both comparisons).

Representativeness of sample structure

The final sample of 3,212 sera (89% of the target) closely followed the age, gender and geographic distribution of the Victorian population.23,24 However, samples from children under 15 years in 1995 were under-represented, males were slightly over-represented, and there was some disparity in sampling between statistical subdivisions within Melbourne.

To assess the effect of these differences on population prevalence estimates, direct standardisation with weighting by age group, gender and region was undertaken relative to the composition of the source population. The similarity between the sample and standardised prevalence estimates (Table 5) provides reassurance of the representativeness of the sample.

Table 5.  Raw sample and standardised prevalence estimates by gender, age group, and geographic regions within Victoria and Melbourne.
SamplesWeighting bySample prevalence estimatesStandardised prevalence estimates
  Anti-HBsAnti-HBcHBsAgAnti-HBsAnti-HBcHBsAg
AllGender30.3%9.3%1.1%30.3%9.3%1.1%
1995Age group17.4%6.5%0.9%17.7%5.7%0.8%
2000Age group30.2%9.7%1.9%30.3%9.6%1.9%
2005Age group40.5%11.2%0.6%40.7%11.2%0.6%
AllAge group30.3%9.3%1.1%31.0%9.1%1.1%
VictoriaStatistical Division30.1%9.6%1.2%29.8%9.5%1.2%
MelbourneStatistical Subdivision32.3%11.9%1.5%30.9%10.9%1.4%

Discussion

The prevalence estimates of 30.3% for anti-HBs, 9.4% for anti-HBc and 1.1% for HBsAg are all higher than the results from a previous national serosurvey of 2,476 sera gathered in 1996-1999 from 45 laboratories around Australia with corresponding estimates of 28.7%, 6.9% and 0.49/0.87%.6 A more recent serosurvey of samples collected in 2002 estimated anti-HBs prevalence at 32.3%, but both anti-HBc and HBsAg were again lower than in our study at 6.1% and 0.7/0.8%.14 Interestingly, a recently published global model from the Centers for Disease Control (CDC) also estimated HBsAg prevalence in Australia at 1.1%.25

One explanation of the higher prevalence of markers of infection in this study is that it was based predominantly on Victorian samples (95.7% of samples with complete postcodes), whereas the other seroprevalence surveys were national.6,14 Although the proportion of immigrant Victorians is similar to the rest of Australia, compared with the national average, a lower proportion of Victorians born overseas come from low HBV prevalence countries (35.8% versus 47.2%), and a greater proportion from intermediate prevalence countries (39.1% versus 27.0%).1,26 The proportion of migrant Victorians born in high HBV prevalence regions is comparable with the Australian average (25.1% versus 25.8%).

The observation that HBsAg prevalence peaked in 2000 and dropped significantly in 2005 may be related to imprecision arising from the relatively small number of HBsAg positive samples (n=36) in the serosurvey. However, a similar peak in 2000/01 with subsequent decline has been observed in unspecified (non-incident or chronic) HBV notifications to the National Notifiable Diseases Surveillance System (NNDSS),22 suggesting the serosurvey may have detected an underlying trend also reflected in surveillance data. We suggest that the trend is real and that both data sources reflect changes in migration patterns into Australia over the past few decades.

There were regional differences in markers of HBV infection, with 1.5% of the Melburnian samples positive for HBsAg, and with statistical subdivisions in Melbourne ranging from 0% to as high as 7.2% (2.0-17.6) HBsAg prevalence – the latter almost reaching the threshold characterising highly endemic countries.4 However the relatively small sample size within these geographic subdivisions means such estimates are subject to significant uncertainty (demonstrated by the broad confidence interval).

The anti-HBs prevalence of 71.8% was less than expected in the 0-4 age group when tested in 2005 – and one three-year-old was also anti-HBc positive, indicating immunity through resolved infection rather than vaccination. The anti-HBs estimate for one year-olds in the 2002 national serosurvey was substantially higher at 86%.14 These children should have been included in the universal vaccination program for infants starting in 2000 if born in Australia, and any born outside Australia would be eligible for catch-up vaccination.19 The ACIR estimate for the proportion of children fully immunised against HBV by 12 months of age has been consistently 94-95% since vaccination commenced.17 With the expectation of seroconversion to anti-HBs of over 95% in this age group27 anti-HBs positivity in this cohort should therefore be around 90%.

There are a number of possible explanations for this observation. Coverage or seroconversion, or both, may be overestimated. In the 2006 Australian Census, 8.5% of 0-4 year olds living in Victoria were born overseas, 54.5% from countries with intermediate or high HBV prevalence.2,26,28 If the source countries lacked universal hepatitis B vaccination, this would contribute a group of susceptibles not accounted for in the ACIR data (not being registered with Medicare by age 12 months). Conversely, those migrating from areas of high endemicity will be contributing to the pool of children with chronic or resolved HBV infection. For the 10-14 year old group, the proportion immunised was 5% less in 2005 than in 2000, but this could be due to vaccination of children already previously infected who would not therefore be included in the ‘Immunised’ status group (isolated anti-HBs positivity). Overall the estimate for adolescents was similar to that of a recent national serosurvey.14

This study has a number of limitations. Although previous research has suggested that seroprevalence estimates derived from convenience samples at VIDRL were comparable to estimates derived by random cluster sampling,29 this applied only to markers of vaccine-derived immunity in school-aged children. The use of samples of convenience introduces a selection bias which is impossible to quantify without a comparator, such as large-scale random population sampling. Such studies are rare30,31 expensive29,32 and remain susceptible to bias.29,32

Conclusions

Despite excluding samples from people ever tested for HBV, HCV and HIV, the Victorian HBV serosurvey 1995-2005 demonstrated a higher burden of chronic HBV infection than has previously been estimated, with 1.1% (0.8-1.6) of samples across the three test years positive for HBsAg. If this is extrapolated to the Victorian population this would represent over 54,000 (approximate 95% CI 39,000-79,000) chronically infected Victorians out of a population of 4.93 million in 2006.26

In the light of this and another recent large-scale convenience sample serosurvey14 it appears that Australia has a higher burden of chronic HBV infection than previously thought. This study also suggests lower population immunity from recently introduced universal vaccination programs than other research has suggested. There is a clear need for further investigations to confirm and explore these findings in other areas of Australia.

The design of this serosurvey allowed the detection of marked variability in prevalence of chronic HBV infection in different regions of Melbourne, primarily reflecting differences in proportions of people born overseas in HBV endemic countries. This information presents a significant opportunity for targeted public health interventions, geographically focused in high prevalence areas and engaging predominantly affected groups. Such interventions, from HBV education for affected communities and the primary care providers serving them, to more comprehensive screening, vaccination and referral programs are likely to be both more effective and more efficient than less targeted strategies.

HBV is under-represented in research and public health funding in Australia when compared with other blood-borne viruses.33 Unlike HIV or HCV, which have been the subjects of successive national strategies for years, Australia is only now developing the first National Hepatitis B Strategy to address the significant and growing burden of HBV infection in our population.34 This burden will clearly be felt most by those infected and their families and communities, but will also challenge the resources of the health care system as a whole.

Funding

BCC was supported by postgraduate scholarships from the Centre for Clinical Research Excellence in Infectious Diseases and the National Health and Medical Research Council. Abbott Diagnostics supported this research by providing the EIA kits at a discounted price. The Division of Epidemiology, VIDRL purchased the kits and provided other resources used in this research. The Serology Laboratory, VIDRL provided laboratory facilities and assisted with testing. All authors declare they have no conflict of interest.

Acknowledgments

BCC acknowledges the members of his PhD supervisory panel for their advice and contribution to this research; Prof. Graham Brown, Assoc. Prof. Margaret Hellard, Assoc. Prof. Heath Kelly and Prof. Sharon Lewin.

Ancillary