The Natural History of Asymptomatic Hepatitis B Surface Antigen Carriers

Methods

Inclusion Criteria

Patients were eligible for the study if they had serum HBsAg positivity lasting at least 6 months; no previous history of acute hepatitis; no clinical signs or symptoms of liver disease; normal alanine transaminase, aspartate transaminase, and alkaline phosphatase levels on at least two occasions 6 months apart; and no associated nonviral liver disease.

Enrollment and Follow-Up

Between November 1976 and July 1981, blood donors found to be HBsAg positive at the Transplantation Immunology and Blood Transfusion Service of the Ospedale Policlinico of Milano were referred to our institution for further evaluation. Patients meeting the inclusion criteria were asked to participate in a long-term follow-up study, which, according to the protocol, included liver biopsy and subsequent regular clinical and biochemical evaluations.

Ninety-two patients gave informed consent and were enrolled in the study. The study group comprised 75 men and 17 women (mean age, 31.5 years; range, 19 to 56 years). Patients were found to be HBsAg positive for the first time 6 to 60 months before enrollment in the study. In most cases, this finding coincided with the introduction of the radioimmunoassay for HBsAg in our hospital and with the application of this test in the mass screening of blood donors. When we began our study, we recalled all donors who had been found positive during the previous years and enrolled newly discovered carriers when they met the inclusion criteria. In 71 patients, the interval between a positive serologic test result and enrollment was less than 2 years. Baseline histologic and biochemical data for the first 68 patients enrolled have been previously reported [11].

At entry, patients were given a questionnaire aimed at collecting demographic data as well as information on alcohol use, hepatotoxic drug use, and possible sources of the infection. Blood samples were collected, assayed for the presence of HBV markers (HBsAg, antibody to HBsAg [anti-HBs], anti-hepatitis B core antigen (anti-HBc), hepatitis B “e” antigen (HBeAg), and antibody to HBeAg [anti-HBe]) and assessed for liver enzyme levels; aliquots of serum were stored at −20 °C. In the following years, when the assays became available, serum specimens were retrospectively tested for the presence of HBV DNA, anti-hepatitis delta antibody (anti-HD), and anti-hepatitis C virus antibody (anti-HCV). Assays for these three markers were done retrospectively in 60, 92, and 74 frozen serum samples, respectively.

Patients then underwent percutaneous liver biopsy, which was done by the standard technique, using a Tru-Cut needle (Baxter, Valencia, California). Patients were invited to return to the clinic twice a year for clinical and biochemical evaluation. During these follow-up visits, serum samples were collected for the evaluation of serum viral markers and assessment of liver enzyme and serum α-fetoprotein levels; again, aliquots of serum were stored at −20°C in most cases.

Patients showing increased serum aminotransferase levels had more extensive biochemical and serologic testing, which included assays for serum IgM anti-HBc, HBV DNA, and anti-HD. These patients also had liver ultrasound scans, and an assay for serum anti-HCV was done in the frozen serum specimen as soon as possible. If a more than twofold increase in serum alanine aminotransferase levels persisted for more than 6 months, repeated liver biopsy was proposed. Patients completing 10 years of follow-up were asked to have repeat liver biopsy, even when no biochemical abnormalities were present; at this time, all patients had liver ultrasound scans and a complete serologic and biochemical evaluations.

Patients refusing to come to the clinic were regularly interviewed by telephone, when possible. Patients for whom telephone numbers were unavailable were traced during the last 2 years through registry offices.

Laboratory Studies

Commercial radioimmunologic assays for serum HBV and HDV markers and α-fetoprotein (Abbott Laboratories, North Chicago, Illinois, and Sorin Biomedica, Saluggia, Italy) were used. Serum HBV DNA was assayed using dot-blot hybridization; positive signals were graded from ± (weakly positive) to ++++ (strongly positive), according to the size and the intensity of spots. Serum anti-HCV was assessed using a first-generation enzyme-linked immunosorbent assay (ELISA) (Ortho Diagnostic Systems, Raritan, New Jersey); positive samples underwent confirmatory testing with a first-generation recombinant immunoblot assay (RIBA) (Ortho Diagnostic Systems). Standard immunohistochemical and immunofluorescence methods were used for the detection of hepatitis B core antigen (HBcAg) and hepatitis delta antigen (HDAg), respectively, in all tissue specimens collected at repeated liver biopsy in patients with normal and abnormal liver chemistries.

Histologic Criteria

Chronic persistent hepatitis, chronic lobular hepatitis, and chronic active hepatitis were defined according to standard histologic criteria [16]. The term “minimal inflammatory changes” (“minimal hepatitis”) refers to mild lobular inflammatory changes without inflammation in the portal tracts [16].

Statistical Analysis

Because our follow-up study was mainly descriptive in nature, no statistical inferences had to be drawn. At entry, patients were grouped according to histologic findings in liver specimens and HBeAg status (Table 1). The crude rates of occurrence for some major events were recorded; for clinical end points, such as death and the development of hepatocellular carcinoma, the total number of patients followed was used as the denominator, and for end points that might be markers of clinical events, such as increased serum alanine aminotransferase levels, the number of patients who underwent regular biochemical follow-up was used as the denominator (Table 2).

Results

Baseline

Histologic evaluations of liver biopsy specimens at baseline showed normal findings in 7 patients, minimal inflammatory changes in 46, fatty liver in 16, chronic persistent hepatitis in 18, and mild chronic active hepatitis in 5. At the time of liver biopsy, 4 patients were HBeAg positive and the remaining 88 patients were anti-HBe positive (see Table 1).

All 92 serum samples collected at the time of liver biopsy were negative for anti-HD when tested at a later date. Of the 74 available specimens, 7 were positive for anti-HCV when tested using a first-generation ELISA, but the first-generation RIBA failed to confirm this positivity. We detected HBV DNA in 10 of 60 baseline serum specimens; two HBeAg-positive patients were strongly positive for HBV DNA, whereas 8 of 58 patients (13.8%) with anti-HBe had a weakly positive signal (± or +). None of these patients had histologic findings consistent with chronic active hepatitis, and only one patient had findings indicative of chronic persistent hepatitis.

Follow-Up

Of the 92 patients who entered the study, 68 underwent a more extensive clinical and biochemical follow-up (mean follow-up, 130 months; range, 70 to 170 months). Twenty-two other patients were either interviewed regularly by telephone or traced through the registry office and contacted by the end of 1990; all denied having liver disease, and most stated that they were having regular biochemical evaluations elsewhere. One patient died in 1988 of lung cancer, and the remaining patient could not be traced through the registry office because of insufficient demographic data.

In 10 patients, loss of serum HBsAg was observed 7 to 11 years after liver biopsy; only 2 of these 10 patients became anti-HBs positive thereafter. All 66 anti-HBe-positive patients who had clinical and biochemical follow-up remained anti-HBe positive throughout follow-up; both patients who were HBeAg positive had anti-HBe seroconversion, one at 15 months and one at 39 months after liver biopsy. The latter patient showed a sustained increase in serum transaminase levels at the time of seroconversion.

Serum HBV DNA positivity was not subsequently confirmed in any of the six patients who were positive at enrollment.

Liver enzyme levels were persistently normal in 58 of the 68 patients (85%) who had regular follow-up. Two patients showed mild and fluctuating elevations in aminotransferase levels; one of them refused repeat liver biopsy, and the other showed slight histologic improvement (from mild chronic active hepatitis to chronic persistent hepatitis).

Another eight patients (12%) had sustained increases in serum aminotransferase levels: The biochemical changes in five of these patients appeared to be unrelated to HBV or HCV infection, because they were persistently negative for serum HBV DNA, IgM anti-HBc, anti-HD, and anti-HCV, and tissue staining for both HBcAg and HDAg was negative at repeat liver biopsy. Four of these five patients showed no histologic deterioration, and serum aminotransferase levels returned to normal a few months after repeat liver biopsy. In the fifth patient, repeat liver biopsy showed features consistent with alcoholic liver cirrhosis; no evidence of liver decompensation or portal hypertension was present at the time of liver biopsy. Three years later, when interviewed by telephone, this patient denied any further ethanol ingestion as well as any symptoms.

In the other three patients with sustained aminotransferase increases, biochemical changes appeared to be related to HBV infection. One patient had four flare-ups of viral replication, which were accompanied by sharp elevations in serum transaminase levels; in addition, this patient's histologic evaluation showed deterioration (from minimal changes to chronic active hepatitis). Another patient had a sustained elevation in serum transaminase level at the time of anti-HBe seroconversion. This elevation lasted 42 months before levels returned to normal; no histologic deterioration was observed. A second, self-limited flare-up of viral replication was seen 2 years later. The third patient showed persistent increases in serum transaminase levels during the last 7 years; this patient was persistently negative for HBeAg and HBV DNA but persistently positive for IgM anti-HBc; no histologic deterioration was observed in this patient. By 31 December 1990, we had completed 10 years of clinical and biochemical follow-up in 51 of the 58 patients with repeatedly normal liver enzyme levels (in the remaining 7 patients, clinical follow-up ranged from 6 to 9 years, after which they were followed by telephone only). Blood samples were collected for HBV DNA, anti-HD, and anti-HCV assay, which were negative in all instances. Liver ultrasound was normal in all cases. Eight of these 51 patients had already shown clearance of serum HBsAg, and repeat liver biopsy was not proposed to them. Twenty-one other patients consented to have repeat liver biopsy, and 22 patients refused. All patients who had repeat liver biopsy showed no histologic changes, and tissue staining for HBcAg and HDAg was negative.

In summary, during a mean follow-up of 130 months, only 3 out of 68 patients (4%) showed HBV-related biochemical changes, with only 1 of these 3 patients exhibiting histologic deterioration. Another patient developed alcoholic liver cirrhosis, but neither HDV superinfection nor HCV infection was observed, and no patient developed hepatocellular carcinoma or had an increased serum α-fetoprotein level. No liver-related deaths occurred (see Table 2).

Discussion

Although much morbidity and mortality have been found in Asia among HBsAg-positive patients, even in the absence of overt liver disease [15, 17], our data strongly indicate that in the Western world, the natural history of HBsAg carriers with persistently normal transaminase levels is, in most cases, uneventful, as has been suggested by previous studies [10, 12, 18]. In fact, during a mean follow-up of 130 months, only 1 of 91 patients developed severe liver disease (cirrhosis), and no liver-related deaths occurred. In the patient with cirrhosis, clinical, serologic, and histologic features strongly suggested that alcohol use was the cause of the disease. It has been reported that HBsAg carriers are more prone to develop alcohol-related liver damage than normal persons [19]; thus, it seems wise to recommend that such patients abstain from alcohol.

Further, only 3 of 68 patients who had regular clinical and biochemical follow-up showed sustained elevations in serum aminotransferase levels that could be considered HBV related, as judged by the positivity for serum HBV DNA, IgM anti-HBc, or both; only 1 of these patients showed histologic deterioration, progressing from minimal inflammatory changes to chronic active hepatitis. It seems, therefore, that spontaneous reactivation of viral replication is a rare event in HBsAg-positive patients in whom a previous replicative phase was never observed; this concept appears to be at variance with the fairly high rate of spontaneous reactivations seen in patients in whom previous seroconversion to anti-HBe has occurred [14, 20] and with the frequent, often devastating reactivations that have been described after discontinuation of immunosuppressive treatments [21, 22]. Even HDV superinfection appears to be an extremely rare event in Italian asymptomatic HBsAg carriers. We did not observe it in any of our patients over a mean follow-up of 11 years, which may seem surprising, given the high rate of HDV infection among Italian HBsAg carriers and the well-documented frequency of inapparent parenteral transmission [23]. However, although a rate of anti-HD positivity of up to 6% has been previously reported among Italian HBsAg carriers with normal liver function tests [24], other observations also suggest that HDV superinfection is most infrequent in HBsAg-positive patients with persistently normal liver chemistries [10]

. In our series, no patients developed hepatocellular carcinoma after 9 to 14 years of follow-up. Although this finding is reassuring and is consistent with findings in previous surveys [10, 12], we cannot estimate the risk for hepatocellular carcinoma in asymptomatic HBsAg carriers because a much longer follow-up and a much larger patient sample is required for this purpose. However, the risk appears to be much lower than that seen in Italian HBsAg-positive patients with liver cirrhosis [25]. This finding provides support for the concept that a promoting event, such as necroinflammatory activity, is the central factor leading to cancer in most HBV-infected patients [3].

Finally, over a mean follow-up of nearly 11 years, 10 of 68 patients lost HBsAg, and two of these patients subsequently acquired anti-HBs; it seems, therefore, that an extremely late stage of HBV infection exists, in which even HBsAg production by integrated virions ceases.

We therefore conclude that asymptomatic HBsAg carriers have an extremely low risk for developing severe liver disease over a 10-year period. Although regular follow-up assessments of liver enzymes and α-fetoprotein may be advisable, a more extensive diagnostic workup, including liver biopsy, is justified only if significant biochemical abnormalities occur.

Abbreviations

anti-HBc: antibody to hepatitis B core antigen

anti-HBe: antibody to hepatitis B “e” antigen

anti-HBs: antibody to hepatitis B surface antigen

HBV: hepatitis B virus