>>what do colleagues think of the meta-analysis in annals on internal
medicine 20th June, 2006?  

The  concern about increased risk of asthma death with LABA has never
really been dispelled, and this puts it fimrly on an evidence based
footing. 

may be fortunate for the NHS that inhaled steroids are mostly off patent
and coming down in price.
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>


Hi Olive, Douglas and Group

This isn't that new or easy - see e.g. Lurie and Wolf. Lancet
2005;366:1261-62 for a critique of the SMART RCT which features in the
Salpeter MA. 

Apologies for the length of this posting.

There's a (in my view) balanced commentary in the EBM Journal 13th Feb
07 responding to the Salpeter MA. 

My $0.02 would be that in the clinical situation of someone with
problematic asthma despite a short acting beta 2 plus low dose inhaled
corticosteroid (ICS) there are two choices:-
1. Add in a long acting beta 2 (LABA) - with an apparent very small
absolute increase really bad things happening like asthma related death
or life threatening asthma exacerbations (based on very small numbers of
events in the Salpeter MA) OR
2. Increase the dose of inhaled corticosteroids - with an apparent very
small absolute increase in pretty bad things happening like osteoporosis
and cataract but based on observational data with the risk of oral
corticosteroid courses confounding(see Drug and Therapeutics Bulletin
2000;38:5-8). There seems no doubt that hoareness, dysphonia and candida
are all dose-related ICS side effects. (Powell H, Gibson PG. Med J Aust
2003; 178: 223-225). And we know about high doses of ICS rarely causing
adrenal suppression.

There's actually a MA of RCTs where salmeterol was added to moderate
dose ICS (<200 microg FP) vs. use of high dose ICS (>200 microg FP
daily). 12 studies, 4,576 subjects. Withdrawals due to asthma was higher
in high dose ICS group OR 1.58 (95%CI 1.12-2.24). As was the rate of at
least one moderate/severe exacerbation OR 1.35 (95%CI 1.1-1.66). Other
outcomes - FEV1, morning and evening PF, and beta-agonist use were all
significantly favoured by "adding-in" the LABA. (Masoli, Weatherall, et
al. Thorax 2005;60:730-734)

Lots of professional and lay people are adverse to the name
"corticosteroids " and "increased dose" being used in the same sentence
in consultations, hence at least some of the growth in LABA use. 

But there seems to be some protection from very bad things happening if
people take at least some inhaled corticosteroid - so it seems important
that if the LABA option is taken every effort is made to reinforce the
importance of not discontiuning prematurely the ICS. One might assume
that this may be a particular risk if symptoms improve with the LABA.

I've tapped to the group before about weighing all this up using the
four domains of effectiveness, safety, cost and patient factors /
preferences. It would seem that:- 

Effectiveness - just favours the LABA (e.g Masoli), 
Safety - might just about favour the ICS now if we accept the Salpeter
data and trade off mostly short term harms from the ICS vs very small
absolute risks of very bad things happening with the LABA or ICS,
Cost - well, neither LABAs and moderate / high dose ICS are low cost in
the NHS
Patient factors / preference - might this dominate here? - especially if
the patient has a trial of one of the 2 options and does well. If they
don't then there's always the option of the alternate. 

Do you agree? 

In the UK, SIGN and the British Thoracic Society resolved their debate
(pre-Salpeter) by going for the LABA option in their appraoch with the
caveat that at follow up:-
- good response to LABA - continue LABA
- benefit from LABA but control still inadequate - continue LABA and
increase inhaled steroid dose to 800mcg/day (if not already on this
dose)
- no response to LABA - stop LABA and increase inhaled steroid to
800mcg/day .  
- If control still inadequate, institute trial of other therapies.

The question of course is whether the Salpeter data would substantially
tip the policy back towards doing the ICS step up before adding the
LABA. My view is even if that was decided, there's enough anti-steroid
mentality prevalent in clinical decision (and guideline?) making that it
might be difficult to 'sell' that concept. 

Ian Scott's brief but wonderful essay on the history of EBM,
implementation and a look towards the factors influencing decision
making in consultations in the same EBM Journal (13 Feb 07) was
especially persuasive that clinicians and their patients are human and
not, therefore, entirely cold and calculating when it comes to making
decisions. Even when they have all the knowledge (debatable given the
volume), the skills to summarise even trusted summaries of evidence into
acccurate, fair and balanced personal 'mindlines' (a long way off given
that's not the approach taken to continuing professional development),
and can translate the benefits and risks into terms patients can
understand (lots of work needed given front line clinician's current
facility with even absolute and relative risk)- it seems likely that
decisions will continue to be made at least partly on the basis of
emotions as well as facts. Quite right too.

There are harder rocks and harder places in consultations than this
dilemma, but it just shows that this medicine thing is not easy. Indeed
more evidence sometimes seems to make it harder to come off the fence
and make a clear recommendation, as here. But perhaps that's what
patient-centered medicine is about?

Best wishes to all

Neal

Neal Maskrey
Director of Evidence Based Therapeutics
National Prescribing Centre
Liverpool
UK