A presentation at this week's virtual HIV Glasgow conference on the cost-effectiveness of the new injectable formulation of the antiretroviral drugs cabotegravir and rilpivirine finds that, if it improves adherence relative to oral drugs, it should save the UK NHS money, at least over a patient's lifetime. It would also add from 2-15 weeks to the time people with HIV who take it can expect to live in good health, and would avert some HIV transmissions.
The analysis, which was conducted by manufacturers ViiV Healthcare, compared the projected adherence rates to injectable cabotegravir/rilpivirine, which is given every two months, to the projected adherence rates to oral therapies. Adherence rates were modelled on data from clinical studies.
It was assumed that there was no difference in the biological efficacy of injectable and oral treatment, and that the annual cost of the injectable therapy would be the same as the average cost of the three most widely used oral therapies in the UK.
The adherence rates for injectables seen in clinical trials were defined as optimal, although this does not mean that adherence to injectable therapies will be literally 100%. For example, in the FLAIR study, which compared the efficacy of cabotegravir/rilpivirine injections with the efficacy of oral therapy in patients new to HIV treatment, the adherence rate to the injectable therapy in the first 48 weeks was 98%, and estimated adherence to daily oral therapy was 90%. (The 98% is the proportion of patients who attended their next appointment no more than seven days before or after the two-month period since the previous injection.)
The cost-effectiveness calculations were based on modelling the assumption that adherence to oral therapies in clinical settings is usually lower than that seen in scientific trials used to get the drugs approved. Calculations were made for various levels of adherence, ranging from the same as injectables (100%) to 75% (25% poorer adherence).
The model assumed that a proportion of patients with poor adherence will develop detectable viral loads and low CD4 counts, will eventually need more expensive fourth-line therapies, and that some may become ill and die earlier. These inputs were based on the UK HIV epidemic.
The researchers also then added into this basic cost calculation a further assumption that a few people whose HIV became detectable would go on to infect others, with further consequent costs to the health system.
An example of how they did the calculation was given, using the assumption that adherence to oral therapies would be 5% lower than adherence to cabotegravir/rilpivirine injections. This finds that the lifetime cost of first-line oral therapy would be over £4500 cheaper per patient than the injectables; this is because patients would spend less time on it before needing to switch.
But the cost of second-line and subsequent therapies they switch to would eventually add nearly £11,000 to the overall lifetime extra drug costs of patients taking oral treatment, compared with those on cabotegravir/rilpivirine injections. This would lead to an overall lifetime cost saving of around £4200 per patient on the injectable therapy.
This implies that for a lifetime’s oral therapy for every 1000 patients, the excess cost to the NHS of providing therapy with 95% adherence relative to injectable therapy would be £4.2 million, rising to over £20 million if adherence was only 75% of adherence to injectable therapy.
If patients taking oral therapy were 95% as adherent, this would also mean that by using injectables instead, they would have gained extra time in good health amounting to 43 years per 1000 patients, which averages out to just over two weeks per patient, rising to over eleven weeks per patient if those on oral therapy were only 75% adherent.
Adding in the extra assumption of transmissions due to unsuppressed viral loads finds that over a lifetime, there would be two extra HIV transmissions per 1000 patients if adherence to oral therapies was 95%, rising to eleven if it was 75%.
This would increase the lifetime extra cost per 1000 patients to over £5 million if adherence to oral therapies was 95% of adherence to injectable therapy, and over £26 million at 75% adherence. It would also increase the average extra time spent in good health by three and 15 weeks respectively.
Health economics researcher Ian Jacob, who co-authored the paper, said: “Our model shows that cabotegravir/rilpivirine injections could be considered cost-effective in the UK setting, according to our analysis and model assumptions.”
He noted that the model did not include the cost of treating opportunistic infections and other sickness in non-adherent patients in their model, nor adherence support and mental health support. These could in theory, if the underlying model assumptions reflect reality, make cabotegravir/rilpivirine injections even more cost-effective.
"The gains would be wiped out if adherence to cabotegravir/ rilpivirine injections turns out to be worse than, or no better than, to oral therapies."
As with all cost-effectiveness models, this one makes some highly simplifying assumptions in order to come out with its findings, and should be taken as a statement of potential cost savings rather than a forecast.
In particular, the gains would be wiped out if adherence to cabotegravir/rilpivirine injections turns out to be worse than, or no better than, to oral therapies. This would happen if people dropped out of receiving their injectable therapy, missed appointments, or if adherence to oral therapy was as good as to the injections. The 96-week data from FLAIR showed that the percentage of people who developed a viral load of over 50 copies was the same in both arms, at 3.2%. This could mean that relatively small differences in adherence might not make much difference to effectiveness in first-line therapy.
However Ian Jacob told aidsmap.com that, at least in the FLAIR study, none of the 2% of patients on injectable therapy who attended appointments more than seven days late experienced a rebound of their HIV viral load. There were also no viral load rebounds seen in patients on cabotegravir/rilpivirine injections in the second year of FLAIR. And another presentation at Glasgow found that since the the majority of patients were keen to start injectable therapy once available, their motivation to take it is likely to be high.
Nonetheless, the model seems to assume that near-100% adherence to appointments can be achieved without the same level of logistical and social support provided in clinical trials.
The manufacturers of the cabotegravir/rilpivirine injectable formulation, ViiV Healthcare, submitted their licensing application to the European Medicines Agency last summer. They re-submitted their application to the US Food and Drug Administration in April 2020, and it will probably be licensed for use late this year or early the next.
Parker B et al. Cost-utility analysis of long-acting cabotegravir + rilpivirine for the treatment of HIV infection in the United Kingdom. HIV Glasgow 2020, poster presentation P030.