Guidelines, which guidelines?

HTU editor Gus Cairns investigates HIV treatment guidelines.

Last issue we reported on new treatment guidelines published by the European AIDS Clinical Society (EACS). This succinct 80-page booklet contains guidance not only on treatment for HIV, but also how to treat TB, hepatitis, cancers, high blood pressure, diabetes and other conditions in people with HIV.

Earlier in the year, the British HIV Association (BHIVA), distributed updated versions of six of its guidelines, running to 292 A4 pages. Areas covered included treating HIV, vaccines, pregnancy, common cancers and sexual health.

Comparing these guidelines reveals subtle differences of emphasis on some common and important decisions, such as when to start treatment.

It’s worth noting that the panel writing the recent US guidelines, published by the Department of Health and Human Services (DHHS), could not agree on this point, with half the doctors recommending treatment for virtually all patients, including those with CD4 counts over 500, and 45% recommending most patients wait until their CD4 count is below 350 cells/mm³. This is because a US study and a European study came to different conclusions on the benefits of HIV treatment in patients with high CD4 counts.^1,2

I interviewed the chairs of the panels that write the EACS and BHIVA guidelines. These are, respectively, Professor Jens Lundgren of the University of Copenhagen³ and Professor Brian Gazzard of London’s Chelsea and Westminster Hospital.

HTU: How prescriptive are your guidelines? Do you want physicians to regard them as gospel?

Jens Lundgren (JL): It’s understood that there is room for discussion and reflection. Ideally, I’d like people to add in their own stuff: I’d be delighted to see a copy with notes in saying “I don’t agree with this recommendation.”

Brian Gazzard (BG): They are couched in terms of ‘this is guidance’, not ‘this is what you need to do’. Guidelines do say, first, that here are some standards below which you should not fall.

However they also review the literature in cases where the evidence is not clear. Take protease inhibitor monotherapy – the evidence for its efficacy and who it might work best for is incomplete. So it’s good to summarise the literature and go into the pros and cons in a fair amount of detail.

Having said that, my feeling is that the main BHIVA guidelines are too long. I’d like to see them shortened, but with reference to appendices that explain why we’ve said what we’ve said.

HTU: What if it’s the patients who take them as gospel? If they read them and walk into their next appointment saying “Here, you’re not doing this”?

JL: Guidelines are advice, not rules. I’d feel terrible if patients used them to accuse their doctors of doing a bad job. They’re only as good as the knowledge people had when they were written.

More than 40 countries use our guidelines and what drugs are available and what they cost varies widely. Professor Jens Lundgren, University of Copenhagen

I think in the long run the single, simple recommendations may disappear. We have enough drugs now to provide quite a wide choice and the job of guidelines in that field should be to specify which populations you should not use specific drugs for.

BG: For some patients it’s good to know what should be done because there are still a lot of doctors around not practising good medicine. I don’t mind if they’re directive if the evidence is clear: for instance we now say “efavirenz should be considered as first line in all patients” and “boosted protease inhibitors (PIs) should ordinarily be reserved for specific groups of patients”. Note we say NNRTIs [non-nucleosides, including efavirenz] should come first, not that using PIs is wrong. The EACS guidelines can’t be as specific as this because they cover countries where boosted PIs have been the favoured first-line treatment. All sorts of politically sensitive decisions go into guidelines. What CD4 count to start at is one of those.

HTU: Do you think we’ll get to the point where CD4 count becomes less important and we’ll just say “treat everyone as soon as they’re ready”, though obviously with more urgency for low counts?

JL: I think CD4 count still matters. I don’t think it’s yet settled if HIV drugs have more benefit than harm in asymptomatic patients with CD4 counts over 350. If we start patients on treatment above this level it’s not to prevent AIDS, it’s to prevent serious non-AIDS-related conditions such as certain cancers; although we know treatment may reduce the rate at which these occur, we don’t know if it will. We also don’t know the really long-term effects of staying on antiretrovirals and if someone has a CD4 count of 800, they may be able to stay off them for a decade or more.

The START trial will address this. It will compare treatment outcomes in 2000 patients who start treatment at CD4 counts over 500 with 2000 who start below 350. It won’t report till 2015, if completed as intended, but I think it’d be really sad if we missed the window to recruit enough people to it. If a patient asks if they should start treatment early, refer them to the START trial.

BG: Yes, START is a very valid study. It’s still not clear what the benefits of early treatment are and whether the weariness of taking tablets for even five to six unnecessary years will outweigh them. The next set of BHIVA guidelines will have a more in-depth discussion on when to start.

HTU: To what extent is cost something you have to take into account? I’m thinking of the fact that boosted PIs are more expensive than NNRTIs, as well as the cost of starting early.

JL: At the moment not at all. More than 40 countries use our guidelines and what drugs are available and what they cost varies widely. We will have to continue to think about this, though: what happens when a reasonable choice of cheaper generic drugs appears?

BG: Here we’ve had to steer a course between pure clinical evidence with no mention of cost and some acknowledgement of it. NICE (the National Institute for Health and Clinical Excellence) authorises non-HIV treatments on cost-effectiveness grounds and acknowledging cost is an important part of making sure NICE continue to look elsewhere. I think if NICE got involved in guidelines you’d find regimen choice very constrained. 

HTU: What about the guidance on other conditions? Are you trying to get HIV doctors to treat conditions they should be leaving to other specialists?

JL: In some ways it’s there to remind HIV physicians what they learned in medical school, but it’s not about getting them to be experts on everything. It’s about what to think about when you encounter a particular problem in a patient and it’s about referral. This has to be a collaborative process. On the one hand the HIV physician is recognising the limits of their expertise and saying “Can you please take on the management of this patient’s diabetes” or whatever it is. On the other hand they know that in many conditions there is a component specifically caused by HIV, or its treatment, which the other specialist may not know about. It would be great if the guidelines could be disseminated through other specialities as well as through HIV medicine.

BG: The model of care for patients with these co-morbidities [concurrent diseases] is quite complex. I’d refer a patient with kidney failure to a nephrologist, but I wouldn’t let go of clinical responsibility. At the Chelsea and Westminster we’ve had a joint dermatology clinic for years and we’re getting together a clinic for ageing patients, which will include everything from bone specialists for people with osteoporosis to social workers and volunteers for people with social isolation.

HTU: So what should patients do if they come across different recommendations between guidelines?

BG: Guidelines need to come with the warning that as soon as we deviate from proven fact readers need to beware, but at the same time that’s where you’ll find the most valuable bits of evidence. You don’t need guidelines for what everyone knows.

JL: The EACS guidelines certainly aren’t intended to supersede national guidelines. A lot of it’s about semantic difference. Where one guideline says abacavir is an ‘alternative’ rather than ‘recommended’ regimen while another says it’s recommended but calls for ‘caution’, you’re essentially saying the same thing. You should expect difference: guidelines will reflect the time they are written.

Guidelines

BHIVA guidelines: www.bhiva.org/TreatmentofHIV1_2008.aspx

EACS guidelines: www.europeanaidsclinicalsociety.org/guidelines.asp

US DHHS guidelines: http://aidsinfo.nih.gov/contentfiles/AdultandAdolescentGL.pdf