This paper is dedicated to the memory of Dr Stuart Wood, Senior Lecturer in General Practice at the University of Glasgow, who died in March 2006. His kindness, encouragement and enthusiasm for the project are greatly missed.
Professor Neil C. Thomson Respiratory Medicine Section Division of Immunology Infection and Inflammation University of Glasgow Gartnavel General Hospital G12 OYN Glasgow UK
Background: The warm, humid environment in modern homes favours the dust mite population, but the effect of improved home ventilation on asthma control has not been established. We tested the hypothesis that a domestic mechanical heat recovery ventilation system (MHRV), in addition to allergen avoidance measures, can improve asthma control by attenuating re-colonization rates.
Methods: We conducted a randomized double-blind placebo-controlled parallel group trial of the installation of MHRV activated in half the homes of 120 adults with asthma, allergic to Dermatophagoides pteronyssinus. All homes had carpets steam cleaned and new bedding and mattress covers at baseline. The primary outcome was morning peak expiratory flow (PEF) at 12 months.
Results: At 12 months, the primary end-point; change in mean morning PEF as compared with baseline, did not differ between the MHRV group and the control group (mean difference 13.5 l/min, 95% CI: −2.6 to 29.8, P = 0.10). However, a secondary end-point; evening mean PEF, was significantly improved in the MHRV group (mean difference 24.5 l/min, 95% CI: 8.9–40.1, P = 0.002). Indoor relative humidity was reduced in MHRV homes, but there was no difference between the groups in Der p 1 levels, compared with baseline.
Conclusions: The addition of MHRV to house dust mite eradication strategies did not achieve a reduction in mite allergen levels, but did improve evening PEF.
The prevalence of asthma in the western world has increased over the last generation (1), in parallel with a warmer indoor microclimate (2). Increased insulation, double glazing and modern building construction have improved standards of heating and energy efficiency in homes, but with reduced ventilation (3). A warm, humid indoor environment favours the growth of the house dust mite population (4). Sensitivity to the house dust mite is the most common allergy associated with asthma in the UK (5).
Studies of occupational asthma (6) and altitude (7) infer that the environment may directly affect symptoms of asthma. Accordingly, allergen avoidance has been advocated as an important aspect of asthma management, yet the evidence for its efficacy is limited (8). Large studies of conventional measures to eradicate dust mites, such as mite-impermeable mattress covers, have not shown a benefit for asthma symptoms (9). However, Morgan et al. (10) found significant improvement in asthma when a number of allergens were targeted in combination with an educational intervention and smoking cessation advice. This study was not blinded and some positive results were subjective, nevertheless it addressed the concept of environmental control of allergens that has repeatedly been emphasized in statements from the Global Initiative for Asthma and the National Asthma Education and Prevention Program. Recent meta-analyses recommended that repeating common interventions is unlikely to be informative (9); however, more comprehensive interventional studies on house dust mite control are required (11), and more specifically, there is a need for studying the health benefits of dehumidification by a double-blind, randomized, controlled trial with adequate sample size measuring clinical outcomes in patients with asthma (12).
As house dust mites thrive in moist conditions, an additional eradication strategy would be to reduce indoor air humidity by improving ventilation. Mechanical heat recovery ventilation (MHRV) is a method of active ventilation using both an extract and a supply fan. Outdoor air is supplied at ambient humidity into the living room and bedroom, and extracted from the kitchen and bathroom. There is evidence that mechanical ventilation reduces indoor air humidity and the house dust mite allergen burden (13). This randomized controlled study in 40 homes of HDM-sensitive asthma subjects, with mechanical ventilation, achieved significantly lower humidity levels and Der p 1 allergen (means: 2.05 vs 5.82 μg/g dust, P < 0.001), with a trend for improved histamine (PC) (20) (P = 0.085) but no change in symptoms or lung function; therefore the clinical effects of mechanical ventilation on asthma remain unproven. We tested the hypothesis that domestic MHRV, in addition to allergen avoidance measures, can improve asthma control of subjects sensitive to house dust mite allergen, by attenuating re-colonization rates.
Participants 16–60 years of age were eligible if they had asthma for more than 1 year, were on regular inhaled corticosteroids and had daily symptoms. Participants were recruited from general practice (GP) and hospital clinics in Lanarkshire, Scotland, UK. Variable airflow obstruction of ≥12% on spirometry (14) or ≥15% on peak expiratory flow (PEF) readings (15) or a symptom score of ≥0.86 on the Asthma Control Questionnaire (ACQ) (16, 17) was required for inclusion. Participants had a minimum forced expiratory volume in 1 s (FEV1) of >50% predicted at baseline and had not had an exacerbation in the previous month. Spirometric measurements were recorded using an electronic spirometer (Vitalograph, Buckingham, UK), before and after inhaled salbutamol (400 μg) (18). Peak expiratory flow measurements were taken at home using a mini-Wright peak flow meter (Clement Clarke, Harlow, UK). Allergy to Dermatophagoides pteronyssinus was determined by positive skin prick test, defined as a wheal diameter of ≥3 mm greater than that of negative control at 15 min; solutions supplied by ALK Abello (Hungerford, UK) (19). Participants were excluded if they were likely to move house or had a pet that provoked their symptoms. The Lanarkshire Research Ethics Committee approved the study. All participants gave written informed consent.
This was a randomized, double-blind, placebo-controlled, parallel group study to evaluate the effect of home installation of MHRV, in addition to conventional eradication strategies, in adults with asthma who were sensitive to D. pteronyssinus.
Homes of eligible participants were surveyed to assess suitability for installation. Homes were excluded if installation was technically difficult or if there was asbestos in ceiling materials. MHRV units (HR250 or HR800) were fitted in the roof space or hallway cupboard in 120 suitable homes by ‘Vent-Axia™’ (Crawley, UK). These energy efficient units extract air continuously from the kitchen and bathroom and deliver prewarmed air via insulated ducts into the bedroom and living room (Fig. 1). The system provided an additional 0.5 air exchanges per hour to the living room and bedroom.
Participants attended a baseline visit where spirometry was recorded, and the ACQ (16) and St George’s Respiratory Questionnaire (20) were completed. The EQ-5D questionnaire (21) was used as a standardized generic instrument for valuing health-related quality of life. Each score is categorized and ‘quality of life years’ (QALYs) can be calculated from the difference in health states due to an intervention. Participants compiled a 2-week PEF diary prior to the visit. Nasal symptoms were recorded on a visual analogue scale (22). Serum total IgE and D. pteronyssinus and pollen-specific IgE antibody were measured by commercial enzyme-immunoassay (Sweden Diagnostics Ltd, Milton Keynes, UK) (23).
House dust allergen, humidity and other environmental measurements
The architect team attended the home for a baseline visit. The bedroom, living room floor and bed surface were vacuumed at a rate of 1 m2/min to obtain complete dust samples using a Dyson model – DC14 (Dyson, London, UK). The dust samples were filtered and weighed, and a standardized soluble extract prepared. The extract concentration of allergens from house dust mite (Der p 1, Der p 2), cat dander (Fel d1) and dog dander (Can f1) was measured using fluorescent multiplex array technology (Indoor Biotechnologies, Charlottesville, VA, USA) (24). The microbial content of dust was estimated by measuring soluble extract concentrations of bacterial endotoxin and fungal β (1–3) glucan (Associates of Cape Cod Inc., East Falmouth, MA, USA).
Temperature and humidity were recorded at 90-min intervals for 12 months in the living room and bedroom using thermohygrographs, ‘Gemini Tiny Tag Ultra two channel’ data loggers (Gemini, Chichester, UK). The critical equilibrium humidity, at which no water is gained or lost by the house dust mite, is 73% relative humidity at 25°C (4). As maintaining relative humidity below 50% is a reasonable recommendation for reducing dust mites in the home in terms of practicability and efficacy (25), the proportion of time that relative humidity fell below 50% was calculated.
Cigarette smoking history. A volunteered smoking history was recorded and objectively confirmed by quantifying plasma cotinine, a stable nicotine metabolite and an indirect measure of individual tobacco smoking. Cotinine was measured using a microplate competitive enzyme-immunoassay (Cozart Bioscience Ltd, Abingdon, Oxford, UK).
Allergen avoidance measures. Once baseline measurements were complete, allergen eradication was carried out in all homes. Carpets were cleaned with a ‘Medivac’ steam cleaner at the rate of 1 m2/min (Medivac Healthcare Ltd, Cambridge, UK). New pillows, duvets and mattress covers were supplied to all participants (‘Naturelle’ range; Medivac Healthcare Ltd, Cambridge, UK).
Randomization and blinding
Randomization was created using a random number generator (sas version 9 software, SAS Institute Inc. Cary, NC, USA), and performed in sequential blocks of four using an automated telephone answering system at the Robertson Centre for Biostatistics, University of Glasgow, UK, by the architect team. Accordingly, a fused electrical spur was switched in the roof space by the architect to activate half of the units. The unit activation device was concealed from the patient and the clinical research team.
To ensure blinding of intervention status, all participants were told not to expect any significant sound or air movement. All the electric switchgear were boxed into prevent tampering. To control for any audible signs, all the low-level electric motors were switched on, but were not connected to the fans in the control group. The fans were ceiling mounted and were relatively inaccessible. Furthermore, the air coming in had been warmed, which along with the low flow rate would have been very difficult to detect. Thereafter, the topic of which group the subjects were assigned to was never brought up and the subjects were remarkably uncurious about this.
Participants were followed up at 3, 6, 9 and 12 months after randomization. Participants measured morning and evening PEF for 2 weeks before each visit. At each visit, spirometry was performed, ACQ score was recorded and requirements for oral corticosteroids, hospitalizations, GP or Emergency Department (ED) visits and any adverse reactions were noted. The St George’s Respiratory questionnaire, the EuroQol questionnaire and nasal visual analogue scale and questionnaires were repeated at 6 and 12 months after randomization. At 12 months, blood samples for IgE serology, dust samples and humidity measurements were taken and placebo units activated.
The sample size was based upon a parallel group design, using a standard deviation of 40 l/min for mean morning PEF. The study was intended to have 64 evaluable participants per group (n = 128), to have 80% power (at the 5% significance level) to calculate a difference of 20 l/min.
The primary analysis was a comparison between groups of the change over baseline in morning PEF. Secondary end-points were evening PEF, ACQ scores, exacerbation and hospitalization rates, spirometry, Der p 1 levels and humidity readings in the homes, IgE levels and economic evaluations. If 12-month data were not available, 9-month data would be used instead. The main analyses were carried out with ancova models adjusted for baseline severity. The analyses were firstly carried out on an intention-to-treat basis. A list of major protocol violators, consisting of those MHRV units that were prematurely activated by the electrician and any randomization errors, was created and the remaining population was denoted per the protocol set. The primary and secondary end-points were repeated for the ‘per protocol’ set. Binary end-points such as hospitalizations were compared by odds ratios, the 95% confidence interval and tested by Fisher’s exact test.
Baseline characteristics of participants
A total of 4986 participants were invited to participate, 482 attended clinical screening. Of these, 188 did not have a positive skin test to house dust mite allergens and 78 were clinically unsuitable (35 were asymptomatic, 22 were lost to follow-up, 12 had low FEV1, nine were symptomatic to their own pet). A total of 216 subjects fulfilled the clinical entry criteria (Fig. 2). Of these 216, 96 were excluded after a house survey which found that 53 that did not fulfil all housing criteria, and the reasons for exclusion were firstly, possible asbestos in ceiling material and secondly, technical difficulties in installing the MHRV unit. A further 43 subjects did not wish to have the MHRV unit installed after discussing the practical details of the unit with the house surveyor. Over time, recruitment and drop-out rates settled at 120 appropriate candidates and MHRV units were installed, one subject defaulted and therefore 119 underwent randomization. Baseline demographic characteristics of those randomized were similar (Table 1).
Table 1. Baseline demographic and clinical characteristics
The values within parenthesis are given as mean and standard deviation, or median and inter-quartile range.
Eighteen major protocol violators were excluded from the per protocol analysis as the MHRV machine had been inadvertently prematurely activated by the electrician. Of these, five were further complicated, three homes had delays of 6 months in randomization and were withdrawn from follow-up. One withdrew due to ill health. One unit appeared to have been reactivated twice prior to randomization and was withdrawn. A further three simple randomization errors were made because the MHRV system had already been activated by the electrician and the randomization was not followed appropriately. These were included in the per protocol analysis in the correct treatment arm. This was partly due to the unusual form of a randomized clinical trial within an architecture study and lack of familiarity with this procedure. In summary of loss to follow-up, 19 patients did not attend the final follow-up: three major protocol violators were excluded by the team, one withdrew due to ill health, one withdrew due to repeated reactivation, five moved house and nine failed to attend appointments.
Clinical outcomes. A total of 100 participants, 53 MHRV and 47 placebo, attended follow-up at 12 months. The clinical outcome measures are listed in Table 2. The primary end-point; change in mean morning PEF, did not differ between the MHRV group and the control group (mean difference: 13.5 l/min, 95% CI: −2.6 to 29.8, P = 0.10). However, there was a significant improvement in the MHRV group compared with the control group in mean evening PEF (mean difference: 24.5 l/min, 95% CI: 8.9–40.1, P = 0.002) (Fig. 3).
Table 2. Comparison of clinical outcomes at baseline and 12 months
MHRV mean (SD)
Placebo mean (SD)
Adjusted difference ancova (95% Cl)
Data represented as mean (SD). Values represent mean difference (CI) compared with baseline. Allergen concentration in μg/g dust.
CI, confidence interval; PEF, peak expiratory flow; FEV1, forced expiratory volume in 1 s; ACQ, Asthma Control Questionnaire score (range: 0–6, with higher scores indicating worse asthma control); St George’s Respiratory Questionnaire (range: 0–100, with higher scores indicating worse quality of life); ED, Emergency Department, GP, General Practitioner; VAS, visual analogue scale (range: 1–10, with higher scores indicating worse symptoms); IgE, immunoglobulin E; Der p 1 and Der p 2, Dermatophagoides pteronyssinus allergen 1 and 2; EU, endotoxin units.
PEF a.m. (l/min)
13.59 (−2.66 to 29.85)
PEF p.m. (l/min)
24.56 (8.97 to 40.15)
−0.25 (−0.57 to 0.08)
Rescue medicine (no. puffs)
−0.04 (−1.00 to 0.92)
St George’s Q (0–100)
−2.83 (−7.82 to 2.16)
FEV1 (% pred)
1.32 (−2.56 to 5.19)
GP out of hours
−0.76 (−1.70 to 0.18)
−0.46 (−0.47 to 0.55)
−0.51 (−1.68 to 0.66)
Der p 1
−0.32 (−0.84 to 0.21)
−3.2 (−6.7 to 0.4)
−1.3 (−2.3 to −0.2)
0.1 (−0.8 to 0.9)
−1.9 (−4.0 to 0.2)
−2.8 (−4.7 to −0.9)
1.46 (−2.65 to 5.57)
−0.4 (−4.7 to 3.8)
−0.5 (−1.6 to 0.6)
Der p 2
−0.04 (−0.16 to 0.08)
−0.6 (−1.0 to −0.3)
−0.6 (−1.0 to −0.1)
0.56 (−0.65 to 1.77)
−0.2 (−1.5 to 1.1)
−1.3 (−2.4 to −0.2)
1.07 (−1.63 to 3.76)
0.1 (−2.4 to 2.7)
−0.3 (−0.7 to 0.2)
−0.29 (−2.63 to 2.06)
0.0 (−1.3 to 1.4)
0.4 (−1.7 to 2.5)
−1.81 (−4.35 to 0.73)
1.1 (−0.5 to 2.7)
2.9 (0.8 to 5.0)
0.61 (−1.59 to 2.81)
0.3 (−1.1 to 1.8)
−1.1 (−3.9 to 1.7)
7.22 (−8.64 to 23.07)
−1.7 (−12.1 to 8.7)
−8.4 (−23.5 to 6.7)
−3.17 (−29.3 to 23.0)
−67.1 (−207.1 to 73.0)
5.8 (−11.9 to 23.4)
−5.20 (−22.4 to 12.0)
−1.4 (−13.9 to 11.2)
7.2 (−6.6 to 21.1)
Mould (μg/g dust) β(1–3) glucan
−7.8 (−26.0 to 10.2)
22.7 (−0.4 to 45.9)
−15.4 (−40.4 to 9.5)
−1187 (−2935 to 560)
497 (−679 to 1674)
−64.6 (−1465 to 1336)
IgE to HDM
2.09 (−5.67 to 9.85)
Values for spirometry, use of rescue medication, ACS, St George’s Respiratory Questionnaire score, rhinitis scale, requirement for oral corticosteroids, GP or ED visits or hospitalizations with asthma did not differ between the two groups. In the economic analysis, there was an insignificant gain of 0.02 QALYs per MHRV patient. The ‘per protocol’ analysis provided similar results to the intention-to-treat analysis. No adverse event was reported relating to the installation of the MHRV unit.
Indoor relative humidity and temperature. Mechanical heat recovery ventilation system significantly reduced mean relative humidity in the bedrooms for a sustained period from October to February (P < 0.05 each month) and in the living room from December to February (P < 0.05 each month) (Fig. 4). The median (inter-quartile range) per cent of time homes achieved a reduction in the indoor relative humidity below 50% was greater in the MHRV group than in the control group in the bedroom [45.1% (30.0–55.1) vs 21.0 (8.5–49.0), P = 0.001], but not in the living room [51.5% (35.4–58.7) vs 40.6% (12.8–63.5), P = 0.26].
Concentration of allergens and microbial products in house dust, and serum IgE levels. At 12 months, the changes in mean Der p 1 and Der p 2 concentrations in the bed, bedroom and living room carpets, as compared with baseline concentrations, did not differ between the MHRV group and the control group, nor were there differences in total or house dust mite specific IgE levels. There were no significant differences in secondary analyses of cat dander allergen (Fel d1), dog dander allergen (Can f1), β(1–3) glucan or endotoxin (Table 2).
This randomized, double-blind, placebo-controlled study examined the effect of the installation of domestic MHRV on asthma control in adults sensitive to house dust mite allergen. It was based on the hypothesis that a warm, humid environment favours the growth of the house dust mite population and that decreasing indoor air humidity with mechanical ventilation would reduce the dust mite allergen burden and improve asthma control.
We found that the MHRV intervention reduced indoor relative humidity in the autumn and winter months, but with no associated effects on Der p 1 burden. We found at 12 months that there was no change in morning PEF; which was our primary end-point, but there was an improvement in evening PEF. The evening PEF is a secondary outcome measure and although the observed improvement was statistically significant, this might have been by statistical chance. Interpretation of this result, therefore, will need to be speculative; however, because the mean improvement of 24.5 l/min was also of clinical significance, this should not be discounted. The morning PEF changes were internally consistent with these evening PEF changes, but they did not quite achieve statistical significance. One reason for this is probably because the study was insufficiently powered to demonstrate a clinical response; only 100 of a projected optimum number of 128 participants completed follow-up. This level of subject attrition reflected the complexity and difficulties in project managing this level of domestic intervention and should be factored in any future study.
It was clear that the reduced relative humidity was insufficient to impact on Der p1 burden, and there was also no difference between the groups in change in serum house dust mite specific IgE antibody. Domestic dehumidification has reduced mite allergen burden in some (13) but not in this, or other (26, 27) trials. Even within overall humidity control, minor fluctuations of humidity levels may permit mite survival. For example, a New Zealand study (26) showed that, although active ventilation did reduce relative humidity to less than 50% for 7 months of the year, there was no effect on mite levels because values were below the critical equilibrium humidity for mite survival for only 39% of the total of 24-h periods for which measurements were made. In another UK study, Fletcher (27) also found no impact of MHRV on Der p 1. It is also possible that 12 months was too short a period to measure a difference in seasonally affected mite re-colonization rates. One underlying reason for this variable lack of efficacy in mite control may be related to climate (28). For ventilation to reduce indoor humidity, the outdoor air humidity must be sufficiently lower; and this may be relevant for our study located in the particularly humid West of Scotland and in others, for example, New Zealand (26). Based on these observations, a future development in the intervention would be a humidistat controller, set at the optimum proportion of relative humidity for temperature, linked to a variable flow fan unit to ensure more refined humidity suppression.
Although we found no effect of MHRV on mite burden, we did not measure airborne allergen concentrations, and it is possible that the increased ventilation rather than the, albeit inadequate, dehumidification could be helpful. We would suggest that ventilation and air sampling for allergen removal be considered in any future intervention trial. There is some evidence for this concept. Burr et al. (29) recently conducted an unblinded mould eradication trial that included improved home ventilation and found symptomatic improvement in wheeze, medication use and rhinitis. However, in our study no difference between fungal glucan exposures was observed between the groups. Other possible factors affected by improved ventilation would include a reduction in respiratory viruses (30), but in our study there was no associated change in GP visits, and other components of indoor air quality, such as particulate matter or volatile organic compounds, seem unlikely. In this case, we can speculate on some of the secondary outcome results of our study. We found that with the MHRV intervention, there was a more prolonged and more significant reduction in relative humidity in the bedrooms, and that the Der p1 levels were lower in bedding, which is arguably the most important exposure. There may be some merit in considering simple improved ventilation in bedrooms of asthmatic children.
The authors are grateful for the initial planning and advice from Professor George Morris from the Scottish Centre for Infection and Environmental Health. They acknowledge the secretarial support of Mrs Janice Reid and the technical assistance of Mrs Christine Downie and her staff at Monklands Hospital Respiratory Investigation Centre. They thank Mr Alan Lawson, Mr Robert Gemmel, Mr Paul McBrearty and Mr Denis Dixon as part of the architect study team. They also thank the Lanarkshire GPs and practice staff for their help with recruitment, North and South Lanarkshire Councils for arranging asbestos checks and ‘Vent-Axia’ for their co-operation in planning surveys and installation of MHRV units. They thank the participants and their families for their helpful participation in the study. The development of the Multiplex Array for Indoor Allergens (MARIA) has been supported in part by the National Institute of Environmental Health Sciences, Small Business Innovation Research contract ES55545.