Cost-effectiveness of oral treatments for chronic hepatitis B☆

Current treatment of hepatitis B has been facilitated with the advent of oral nucleoside analogues. These agents, at least in the short to medium term are generally safe drugs, given once daily, orally, usually with minimal toxicity. Chronic hepatitis B is a prolonged illness however, and the long-term efficacy and toxicity of oral nucleoside analogues remains unknown. Buti and co-authors in an analysis published in this issue of the Journal [1], review the cost-effectiveness of current nucleoside analogue therapy. The analysis is the result of collaboration between hepatologists, a health economist and Gilead Sciences. Their model has not included PEG interferon treatment, mainly because of reliance in the model on the assumption that the level of HBV DNA used to define virological response is higher than for oral antivirals [2]. This assumption would be disputed by some [3].

Lamivudine followed by adefovir after the onset of resistance has been the most widely applied nucleoside treatment option; this strategy has been shown to be cost-effective, even though it is now no longer considered clinically appropriate in many regions to begin lamivudine monotherapy for patients with high levels of HBV replication given the risk of multi-drug resistant hepatitis B and that fact that more potent agents with lower rates of resistance are available [4]. Recent guidelines have attempted to simplify and importantly optimise nucleoside analogue use. However these guidelines, including the EASL guidelines, seldom take cost-effectiveness or cost benefit into account [5]. This analysis usefully evaluates the cost-effectiveness of tenofovir, entecavir and telbivudine, all relatively recently approved potent nucleoside reverse transcriptase inhibitors active against hepatitis B virus.

In their analysis, the authors employed a Markov model used to project lifetime complications and costs in a cohort of 40 year old chronic hepatitis B patients. Both HBeAg positive and negative patients were evaluated, adjudging the annual probability of moving from one disease state, such as chronic hepatitis to cirrhosis. Short-term virological response rates have been established in controlled clinical trials, and these were utilized in this model. The authors used a “virological response” i.e. suppression of HBV DNA to less than 300 or 400copies/ml serum as an index of disease modification. It is noteworthy that a somewhat absolute assumption was utilized. It is likely that a gradation of histological improvement (itself used as a surrogate for modification of the natural history of the disease) occurs with lesser degrees of HBV suppression [6], [7], [8]. However more profound viral suppression is associated with less risk of resistance, other factors being equal. In the model, treatment was discontinued after HBeAg seroconversion, although the durability of this alteration has not been fully established in different ethnic groups or for different HBV genotypes. The model did not take into account loss of HBsAg, the ideal endpoint of nucleoside (or interferon) treatment [9], [10]. Treatment was applied indefinitely in HBeAg-negative patients.

For modeling several assumptions are required. Progression rates were assumed based on available natural history data. All models are likely to be sensitive to these assumptions. The authors have not, for example, taken into account the impact of the rate of progression of HBeAg-negative disease to cirrhosis – which may be crucial. The authors used cost data derived from the Spanish Health Care System.

Six different strategies were compared, including no treatment, or treatment with lamivudine, adefovir, entecavir, telbivudine, and tenofovir. ‘De novo’ combination treatments, particularly with lamivudine and adefovir were not compared. For resistance, a rescue therapy was recommended as resistance likely compromises the outcome of treatment. Non-responders and patients with resistance were treated according to the current EASL recommendations. In particular two combinations were considered for the management of resistance: a rescue therapy with lamivudine and adefovir, or alternatively with tenofovir combined with entecavir. The authors did not consider the addition of tenofovir for lamivudine resistance. This may have been revealing, as the efficacy of adefovir and lamivudine salvage therapy for lamivudine or adefovir resistance is partly dependent upon the concentration of HBV DNA at the time of additional therapy, and on the presence of specific substitutions – for example the addition of lamivudine may not be as effective if A181/VT substitutions compared to N236T substitutions are present. Thus this combination therapy is not infallible as implied in the model.

The results are expressed in costs of treatment over a lifetime, in life years saved and the standardized quality adjusted life year (QALY) saved. The model identifies the most efficacious treatment i.e. that which resulted in the highest number of life years saved and QALYs gained. Some sensitivity analyses are performed to examine the effect of different parameters on cost-effectiveness.

In both HBeAg-positive and HBeAg-negative groups the least expensive strategy was no treatment. The lifetime cost of tenofovir treatment in HBeAg-positive patients was €87,615 (which rose to €112,585 and €123,446 depending upon the salvage strategy), and €105,889 for HBeAg-negative patients. A different ranking for the costs of treatment was observed for tenofovir versus adefovir in HBeAg-negative versus positive patients. For HBeAg-positive patients, the incremental cost per QALY was €2426 and €632 versus lamivudine and no treatment. For HBeAg-negative patients the incremental cost per QALY with tenofovir was €954 with respect to adefovir dipivoxil, €3949 in relation to no treatment and €5211 in relation to lamivudine, respectively. These values are well below the reference threshold of €20,000–30,000 per QALY recognized by many public health care systems as the threshold at which society is willing to pay. Although treatment for resistance favored tenofovir and entecavir rescue treatment, it is worth pointing out that we do not have wide experience with this regimen, and its potential long-term toxicity. Longer term monitoring of osteopaenia in cirrhosis, renal impairment and carcinogenesis will be required.

Hepatitis B remains a numerically important disease in both endemic and non-endemic areas. In many parts of Europe, industrialized Asia and the Americas, reimbursement for optimal treatment is available. This analysis of cost-effectiveness validates the current EASL treatment guidelines, and the role that effective antiviral treatment has in controlling the disease. The situation is different in many endemic and resource constrained regions. Treatment is frequently only dispensed to those who can afford it – a very small minority in many endemic areas. The financial burden of treatment remains unaffordable for most.

This health economic model has demonstrated the cost-effectiveness of tenofovir for the treatment of hepatitis B. It is said that the cost of drugs is based on a premise of what the market will bear, (now tailored to cost-effectiveness thresholds) rather than being pegged to recoup research and development costs. It is fortunate that the cost of tenofovir for hepatitis B has probably been predetermined by the relatively low costs that were set for the treatment of HIV infection. Tenofovir is being distributed as a generic drug at relatively low cost and little profit, for the treatment of HIV infection in many African countries. However, unfortunately the anomaly exists that due to complex marketing, distribution and sub-licensing agreements, tenofovir is not currently available for the treatment of hepatitis B infection in large regions of sub-Saharan Africa, China and other regions. This situation is to be decried and will hopefully be urgently corrected by the responsible industrial partners and government agencies in these regions.