A ‘meta-analysis’ of 54 clinical trials involving 3,000 patients has just been published which concluded that all current methods of mite control or elimination are ineffective. The leading author actually stated ‘we can conclude with confidence that there is no need to buy expensive vacuum cleaners or mattress covers or to use chemical methods against house dust- mites because these treatments do not work’. Adding up all the patients in 54 trials creates impressive numbers but exaggerates the significance of this analysis.
This bad news has been seized upon by many medical journalists and featured in the BBC News, most major newspapers and the internet. This survey will be believed without question because it was carried out by the Cochrane Collaboration, which is accepted worldwide as the last word. Remarkably, this is the third meta-analysis by the same investigators in the last ten years, each time producing a similar result, and well supported by other reports. Can we explain how they get it so wrong?
How meta-analyses work
How these research statisticians, who no doubt work in an office with computers and never see live patients, obtained this result requires an oversimplified explanation of how a meta-analysis is carried out.
The first step is to find all trials published in the last ten or more years, decide on the criteria for inclusion or rejection, and select article that conform.
All trials that are not double-blind, where neither doctor nor patient know if they are having treatment or placebo, are rejected, along with those too small, or not acceptable in some other way. This selection process is crucial, because it can completely distort the final result, especially as there are some very important factors that are often not considered.
Flaws in mite research trials and analyses
The discrepancy between the results of this meta-analysis and clinical experience may be explained by unique features of mite allergic patients, who vary enormously in their sensitivity to mites, just as the amount of mite allergen in their environment also varies widely.
The effects of essential drug treatments need also be taken into account.
The confusing effect of these three major variables could be minimised by measuring the sensitivity of the patient to mites and also the amount of mite allergen in their home environment, plus a survey of treatment requirements.
If this was done before
enrolment in any double blind placebo-controlled trial, random allocation of equally sensitive patients to placebo or active groups would become possible. If these three vitally important variables are unknown it is possible to put too many extremely sensitive or slightly
sensitive patients in active treatment or in the placebo group. This may be why so many trials of allergy treatments are inconclusive.
For these reasons statisticians could be blissfully unaware that they have cherry-picked the wrong studies, twisted the results, and may be regarding apples and oranges as identical without realising it.
Unfortunately meta-analyses are invariably accepted without question for publication in prestigious journals, presumably because a meta-analysis is considered to be evidence-based research which represents the final truth. But, in statistics as in computers, rubbish in equals rubbish out!
In my opinion allergic disease is a special case quite unsuitable for this sort of statistical treatment – because every patient is an individual with many variable factors contributing to the problem.
The best and easiest compromise, but one which is not, unfortunately,
analysts, is to use
the patients as their
own controls in an open trial, comparing symptoms, peak flow rates, and other data before and after the treatment. This was the method used in the four-year trials I ran on Acarosan, a German spray or powder for bedding, carpets and soft furniture, which contains benzyl benzoate to kill the mites, plus a mixture of detergents to make the mite faeces stick together. (For a detailed description of the trial see
Allergic patients, amongst whom individual variation can be extreme, cannot be squeezed into a ‘one size fits all’ category for the convenience of statisticians.
Many patients and some doctors may disagree with the result of this meta-analysis because they have had great benefit from anti-mite measures, while those who have been disappointed may welcome it. The makers of drugs which suppress symptoms but can never cure will certainly welcome this conclusion, and their representatives will make certain that GPs and nurses are aware of it.
(I never forget an employee of a very large drug company who was present when I first presented excellent results from the Acarosan trial and who remarked to my wife ‘I cannot understand why your husband wants to kill our best friends.’ What a significant remark!)
It is over fifty years since mites and their droppings were identified as the major cause of asthma, rhinitis, and many cases of eczema, and we have been trying to kill them off ever since.
Many mite killers – from liquid nitrogen to strong tea – have been advocated or marketed. My personal experience has been with Acarosan – see above. Acarosan proved to be more successful and last longer as a result of the mites’ habit of eating their own droppings and thus recycling the benzyl benzoate. Unfortunately this effective mite killer is not available in the UK.
Children with severe asthma sent to special schools in the Alps improve considerably after a period in that almost mite-free environment. Their asthma may relapse when they are sent home, on holiday or for good, although some maintain their improvement indefinitely. This has been known for many years.
In my trials with Acarosan we also found that, once the mites had been killed off, some of the younger patients never relapsed. For a while these particular young people would become wheezy whenever they visited relatives with dusty houses but they ceased to react as long as they remained in a mite-free home.
Mite avoidance regimes are
routinely recommended by allergists and chest specialists, and are often included in official treatment guidelines although the results can be disappointing, as they depend totally on the
co-operation and enthusiasm of patients.
On the other hand a film shown on Channel 4 TV recently showed what dramatic improvements can be achieved in asthmatic children by getting rid of the mites in their home environment, rather than increasing the dosage of drugs to suppress symptoms.
Current UK research into mites is supported by grants
totalling nearly £400,000
from the Engineering
and Physical Science
Research Council. This
research is mainly directed
at controlling mite infestation by manipulating the temperature and humidity of the home and seems to apply mainly to the indoor climate of the buildings of the future.
One of the discoveries of the research team is that mites have a unique mechanism for taking up water from the air, which consists of ‘dribbling a salt solution from under their eight armpits to their mouth, which enables them to absorb moisture from the air’. Nice to know that!
If the humidity is too low mites dry out and die, but obviously this can rarely be accomplished in the damp temperate climate of the British Isles, so can only apply to the design of new buildings where the indoor humidity can be kept very low. A huge report on asthma and mites from the point of view of building research was submitted, in 2006, to the office of the deputy prime minister, at the time Mr Prescott. I wonder what happened to it?
Keep after them...
It is my considered opinion that the results of this meta-analysis, a very destructive piece of statistical research, should be disregarded. Based on many years of clinical experience, I would like to reassure Foods Matter readers that measures to control mites are really worth while, that you have been wasting neither your money nor your effort, and that you should keep up the good work.
Click here for more articles on dustmites
First Published in 2008
Back to top