Similar effect of therapeutic ultrasound and antibiotics for acute bacterial rhinosinusitis: a randomised trial.
Sinusitis is frequently encountered in general practice. The one-year incidence in primary care in Norway has been reported to be approximately 3.5 per 100 adults (Lindbaek 2004). In the United States, sinusitis is reported to affect 1 in 7 adults each year (Rosenfeld et al 2007a), and sinusitis accounts for 15-21% of antibiotic prescriptions for adult outpatients (Ahovuo-Saloranta et al 2008). The term rhinosinusitis is often used and acute rhinosinusitis may be classified further into acute bacterial rhinosinusitis and viral rhinosinusitis based on symptoms (Rosenfeld et al 2007a). Antibiotics should only be prescribed for acute bacterial rhinosinusitis.
Distinguishing viral from bacterial infections is particularly challenging in the acute stages (Lindbaek 2007). Diagnosis in primary care is normally based on symptoms such as facial and tooth pain, pain when bending forward, and purulent nasal discharge (Lindbaek and Hjortdahl 2002, Meltzer et al 2004). Antibiotics have been the most common intervention for both acute and chronic sinusitis, and when antibiotics are prescribed for acute bacterial rhinosinusitis, amoxicillin has been recommended as the first choice (Rosenfeld et al 2007a). Frequent prescription of antibiotics can lead to an increase in antibiotic resistance (Ahovuo-Saloranta et al 2008, Ferech et al 2006) and current guidelines provide more conservative recommendations for antibiotic prescription for acute bacterial rhinosinusitis (Ahovuo-Saloranta et al 2008, Lindbaek 2004, Rosenfeld et al 2007a). Current guidelines recommend delaying antibiotic prescription for up to 7 days in patients without severe illness (Rosenfeld et al 2007a). Although reviews report superior effect of antibiotics compared with placebo after seven days (Lindbaek 2004, Rosenfeld et al 2007a), others claim that antibiotics are not justified even after 7-10 days (Williamson et al 2007, Young et al 2008). However, physicians often feel pressured by patients to prescribe antibiotics (Varonen et al 2004). Perhaps it is not surprising therefore that the practice of prescribing antibiotics for common infectious diseases, including sinusitis, has not changed significantly in spite of new recommendations and efforts to implement them (Ferech et al 2006, Neumark et al 2009, Varonen et al 2007).
The continuing debate and controversy about prescribing antibiotics for acute bacterial rhinosinusitis, and the resistance to change in practice, motivate a search for alternative interventions. Rapid reduction of the symptoms of acute bacterial rhinosinusitis with therapeutic ultrasound has been observed in the clinic. However, no controlled studies have been conducted. The purpose of this study was to compare the effect of antibiotics with therapeutic ultrasound in patients with clinically diagnosed acute bacterial rhinosinusitis in primary care. The specific research questions were:
1. Is there any difference in the effect of therapeutic ultrasound and antibiotics (amoxicillin) on pain and congestion for acute bacterial rhinosinusitis in the short-term?
2. Is there any difference in patient satisfaction, preferred future intervention, side-effects and relapses in the long-term?
If therapeutic ultrasound gives symptomatic relief equivalent to amoxicillin, it may serve as an alternative to antibiotics.
A randomised trial was conducted in a primary care setting in Norway. Participants were recruited from consecutive patients coming to a single general practice with sinusitis-like symptoms, where they were diagnosed by a physician (AL). After collection of baseline measures, the participants were randomly allocated to an experimental or a control group. The allocation sequence was computer generated in random permutated blocks of 6 or 8 and was concealed from the recruiter and participants in sealed envelopes which were opened by a nurse. The experimental group received four consecutive days of ultrasound and the control group received a 10-day course of antibiotics. The outcomes were all self-reports by the participants who along with the therapists were not blind to group allocation. However, the person analysing the data was blind to group allocation. Pain and congestion were measured at baseline, Day 4, and Day 21. Day 4 coincided with the last day of ultrasound, while Day 21 was 11 days after the end of the course of antibiotics. Satisfaction with the intervention, preferred future intervention, side-effects and relapses were measured one year later.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
Patients with sinusitis-like symptoms were included if they were over 15 years old and had one of the following: pain when bending forward, headache, or pain in the teeth. They must also have had purulent nasal secretion; 'double worsening', ie, worsening of symptoms within 10 days after initial improvement (Lindbaek and Hjortdahl 2002, Meltzer et al 2004, Rosenfeld et al 2007a); and a bacterial infection as indicated by an increased number of granulocytes (neutrophils) relative to lymphocytes on white blood cell count. They were excluded if they had had antibiotics or allergy medication within the last three weeks, were allergic to antibiotics, or were pregnant.
The experimental group received therapeutic ultrasound (a) at 1.0 W/[cm.sup.2] in continuous mode for 10 minutes each day for four days. The transducer was moved constantly in small circular movements on both sides of the nose and over the forehead, ie, over the sinuses (Figure 1). The same machine was used to deliver all ultrasound.
The control group was prescribed antibiotics--500 mg of amoxicillin three times a day for 10 days.
Pain and congestion around the nose and in the forehead and teeth were measured on a numeric rating scale, where 0 represented no pain/congestion and 10 represented the worst pain/congestion possible. Pain around the nose was considered the primary outcome.
Satisfaction with intervention (Y/N), preferred intervention to manage a future episode (same as allocated/opposite of allocated), number of side-effects, and number of relapses were measured using a postal questionnaire.
A change in pain of 2 points on an 11-point numeric rating scale has been shown to represent a clinically important difference (Farrar et al 2003). To have 80% power to detect a between-group difference in pain around the forehead of 2 points on an 11-point numeric rating scale, with alpha at 0.05 and assuming a SD of 2 points, 17 participants were needed in each group. Considering the uncertainty of the SD, to increase the likelihood of normally distributed data, and to account for drop-outs, it was decided to recruit 48 participants.
All participants with follow-up data were analysed according to their group allocation, ie, using an intention-to-treat principle. Due to a low drop-out rate of 6% in the short-term and 12% in the long-term, no attempt was made to impute missing data. Between-group effects for pain and congestion by Days 4 and 21 were reported as mean (95% CI) difference while satisfaction with intervention, preferred future intervention, side-effects and relapses at one year follow-up were reported as relative risk (95% CI).
Flow of participants, therapists, and centres through the trial
Forty-eight patients with acute bacterial rhinosinusitis participated in the trial; 24 were allocated to the experimental group to receive ultrasound and 24 to the control group to receive antibiotics. In the short-term, there were 3 dropouts so that 94% of data was collected and in the long-term there were 6 dropouts so that 88% of data was collected. Figure 2 shows the flow of participants through the trial and reasons for dropping out. The baseline characteristics of the participants are presented in Table 1. The groups were similar in age, gender, smoking habits, duration of current symptoms, previous episodes of sinusitis, and previous intervention except that the experimental group had more experience with nasal irrigation than the control group. Three out of four participants (77%) reported having symptoms for more than 7 days and 41 participants (85%) had had sinusitis previously. White blood cell counts at baseline showed an increase in granulocytes indicative of bacterial infection.
One general practitioner in general practice recruited all the participants and prescribed the antibiotics for the control group. One physiotherapist in a private physiotherapy practice delivered all ultrasound interventions (Table 1).
Compliance with trial method
All participants in the experimental group completed the four sessions of ultrasound. Compliance with taking the antibiotics was not formally assessed, but there were no reports of interruption. The side-effects reported by the experimental group were nausea/stomach pain (n= 1) and headache (n = 2), and by the control group were nausea/ stomach pain (n = 1), fungal infection (n = 1), headache (n = 1) and allergy (n = 1).
Effect of intervention
Group data for pain and congestion in the short-term is presented in Table 2 and satisfaction, preferred future intervention, side-effects, and relapses in the long-term are presented in Table 3.
By Day 4, pain and congestion had decreased markedly in both groups. Pain around the nose had decreased by 1.5 points out of 10 (95% CI 0.6 to 2.5) more in the experimental group than in the control group. There was also a trend for pain in the teeth to decrease more in the experimental group than the control group (mean difference -1.5 points out of 10, 95% CI -3.3 to 0.3). There were no other differences in decrease in pain and congestion between the groups.
By Day 21, pain and congestion had decreased to low levels in both groups. However, there were no differences in decrease in pain and congestion between the groups in any area.
At one year follow-up, there were no differences between the groups in terms of satisfaction with intervention (RR 0.77, 95% CI 0.50 to 1.04), number of side-effects (RR 0.71, 95% CI 0.20 to 2.56), or number of relapses (RR 1.83, 95% CI 0.87 to 4.12). However, the experimental group were more likely to prefer ultrasound than the control group were to prefer antibiotics for a future episode (RR 2.75, 95% CI 1.19 to 7.91).
Despite rapid and clinically significant reductions in facial pain and congestion in both groups, there was little or no difference between the effect of therapeutic ultrasound and antibiotics on pain and congestion in acute bacterial rhinosinusitis in the short-term, although a trend for a faster effect for ultrasound was indicated by a larger decrease in pain around the nose by Day 4. The experimental group were more likely to prefer ultrasound than the control group were to prefer antibiotics as an intervention for a future episode of sinusitis, possibly reflecting a concern for antibiotic resistance. Few side-effects were reported. Four days were required to administer the ultrasound as opposed to 10 days for the course of antibiotics. Delivery of the ultrasound necessitated four visits to a professional whereas prescription of the antibiotics only needed one attendance. The direct costs are probably only marginally different.
There are a number of potential causes of sinusitis (such as bacteria, viruses, fungi, parasites, allergies) and there is lack of consensus on diagnostic criteria and classification (Benninger et al 2003). Distinguishing between viral and bacterial infection in the clinic is difficult (Hickner et al 2001, Young et al 2008) and we cannot rule out that participants with viral infections or other causes of sinusitis were included in our sample. However, symptom duration for most participants of above seven days suggests a bacterial infection (Rosenfeld et al 2007a) and an increase of granulocytes (neutrophils) rather than lymphocytes favours a bacterial rather than a viral infection (Table 1). This is, however, only an indication and not conclusive evidence of a bacterial origin for acute bacterial rhinosinusitis. Imaging, laboratory tests or bacterial culture are not recommended for routine use in primary care (Hickner et al 2001, Rosenfeld et al 2007a). The primary care clinician is thus left to base the diagnosis of acute bacterial rhinosinusitis on signs and symptoms seen in the clinic in line with the procedures used in this study.
We cannot say whether the rapid reduction of symptoms observed in both groups reflects an effect of intervention, placebo, or natural history. Natural history of sinusitis has not been documented (Gwaltney et al 2004). Information on the clinical course of untreated sinusitis comes from patients receiving a placebo in randomised trials for acute bacterial rhinosinusitis, but there are conflicting results. Lindbaek et al (1996) reported a significantly faster and superior effect of amoxicillin compared to placebo within 30 days of symptom onset. However, Rosenfeld et al (2007b) reported improvement after seven days with and without antimicrobial intervention and Bucher et al (2003) reported no advantage of antibiotics over placebo. Since no placebo group was included in our study, we cannot distinguish the effect of intervention from placebo. However, there is some evidence for a superior effect of antibiotic compared to placebo within a window 7-12 days after initiating intervention (Haye et al 1998, Rosenfeld et al 2007b).
The mechanism for a beneficial effect of ultrasound is unknown. Clinically, coloured and purulent discharge is regularly observed during or immediately after intervention. Ultrasound works by transporting mechanical energy through local vibration of tissue particles (Leighton 2007). Perhaps mechanical vibration detaches purulent matter from the walls of the sinuses, independent of a viral or bacterial cause, relieving the pressure and thus easing the pain. Bartley and Young (2009) point to enhanced bacterial death from low frequency, high intensity ultrasound in laboratory settings. When bacteria density reaches a critical level they organize within 'slimy' biofilms for protection, a potential reason for the ineffectiveness of antibiotics. Bartley and Young hypothesise that ultrasound may break down biofilms and that this could either kill or reduce the viability of bacteria directly or make bacteria more accessible to antibiotic intervention by increasing cell membrane permeability.
There is growing concern about resistance and overutilisation of antibiotics for sinusitis-like symptoms in primary care. By confirming that there is no difference between the effect of therapeutic ultrasound compared with antibiotics, except for a faster benefit in terms of pain around the nose, this study provides evidence that ultrasound can be used as an alternative intervention to antibiotics for acute sinusitis. Furthermore, therapeutic ultrasound had no serious side-effects. However, it should be kept in mind that both interventions may have a marginal impact on the natural course of the disease. The combined effect of ultrasound and antibiotics for sinusitis should be investigated.
Footnote: (a) Sonopuls 492, Enraf-Nonius BV, PO Box 12080, 3004, The Netherlands.
Ethics: The study was approved by the Regional Committee for Medical and Health Research Ethics in Trondheim, Norway (2004). Written consent was obtained from all participants before the study began.
Competing interests: None declared.
Support: S0r-Tr0ndelag chapter of the Norwegian Physiotherapist Association for financial support.
Acknowledgements: R0ros Medical Centre for assistance in patient recruitment.
Correspondence: Dr Ottar Vasseljen, Department of Public Health and General Practice, Faculty of Medicine, Norwegian University of Science and Technology, Norway. Email: firstname.lastname@example.org
Ahovuo-Saloranta A, Borisenko OV, Kovanen N, Varonen H, Rautakorpi UM, Williams JW Jr., et al (2008) Antibiotics for acute maxillary sinusitis. Cochrane Database of Systematic Reviews April 16: CD000243.
Bartley J, Young D (2009) Ultrasound as a treatment for chronic rhinosinusitis. Medical Hypotheses 73: 15-17.
Benninger MS, Ferguson BJ, Hadley JA, Hamilos DL, Jacobs M, Kennedy DW, et al (2003) Adult chronic rhinosinusitis: definitions, diagnosis, epidemiology, and pathophysiology. Otolaryngology-Head and Neck Surgery 129: S1-S32.
Bucher HC, Tschudi P, Young J, Periat P, Welge-Luussen A, Zust H, et al (2003) Effect of amoxicillin-clavulanate in clinically diagnosed acute rhinosinusitis: a placebocontrolled, double-blind, randomized trial in general practice. Archives of Internal Medicine 163: 1793-1798.
Farrar JT, Berlin JA, Strom BL (2003) Clinically important changes in acute pain outcome measures: a validation study. Journal of Pain and Symptom Management 25: 406-411.
Ferech M, Coenen S, Malhotra-Kumar S, Dvorakova K, Hendrickx E, Suetens C, et al (2006) European Surveillance of Antimicrobial Consumption (ESAC): outpatient antibiotic use in Europe. Journal of Antimicrobial Chemotherapy 58: 401-407.
Gwaltney JM, Wiesinger BA, Patrie JT (2004) Acute community-acquired bacterial sinusitis: the value of antimicrobial treatment and the natural history. Clinical Infectious Disease 38: 227-233.
Haye R, Lingaas E, Hoivik HO, Odegard T (1998) Azithromycin versus placebo in acute infectious rhinitis with clinical symptoms but without radiological signs of maxillary sinusitis. European Journal of Clinical Microbiology and Infectious Diseases 17: 309-312.
Hickner JM, Bartlett JG, Besser RE, Gonzales R, Hoffman JR, Sande MA (2001) Principles of appropriate antibiotic use for acute rhinosinusitis in adults: background. Annals of Internal Medicine 134: 498-505.
Leighton TG (2007) What is ultrasound? Progress in Biophysics and Molecular Biology 93: 3-83.
Lindbaek M (2004) Acute sinusitis: guide to selection of antibacterial therapy. Drugs 64: 805-819.
Lindbaek M (2007) Acute sinusitis--to treat or not to treat? JAMA 298: 2543-2544.
Lindbaek M, Hjortdahl P (2002) The clinical diagnosis of acute purulent sinusitis in general practice--a review. British Journal of General Practice 52: 491-495.
Lindbaek M, Hjortdahl P, Johnsen UL (1996) Randomised, double blind, placebo controlled trial of penicillin V and amoxycillin in treatment of acute sinus infections in adults. British Medical Journal 313: 325-329.
Meltzer EO, Hamilos DL, Hadley JA, Lanza DC, Marple BF, Nicklas RA, et al (2004) Rhinosinusitis: establishing definitions for clinical research and patient care. Otolaryngology-Head and Neck Surgery 131: S1-S62.
Neumark T, Brudin L, Engstrom S, Molstad S (2009) Trends in number of consultations and antibiotic prescriptions for respiratory tract infections between 1999 and 2005 in primary healthcare in Kalmar County, Southern Sweden. Scandinavian Journal of Primary Health Care 27: 18-24.
Rosenfeld RM, Andes D, Bhattacharyya N, Cheung D, Eisenberg S, Ganiats TG, et al (2007a) Clinical practice guideline: adult sinusitis. Otolaryngology-Head and Neck Surgery 137: S1-S31.
Rosenfeld RM, Singer M, Jones S (2007b) Systematic review of antimicrobial therapy in patients with acute rhinosinusitis. Otolaryngology-Head and Neck Surgery 137: S32-S45.
Varonen H, Rautakorpi UM, Huikko S, Honkanen PO, Klaukka T, Laippala P, et al P (2004) Management of acute maxillary sinusitis in Finnish primary care. Results from the nationwide MIKSTRA study. Scandinavian Journal of Primary Health Care 22: 122-127.
Varonen H, Rautakorpi UM, Nyberg S, Honkanen PO, Klaukka T, Palva E, et al (2007) Implementing guidelines on acute maxillary sinusitis in general practice-a randomized controlled trial. Family Practice 24: 201-206.
Williamson IG, Rumsby K, Benge S, Moore M, Smith PW, Cross M, et al (2007) Antibiotics and topical nasal steroid for treatment of acute maxillary sinusitis: a randomized controlled trial. The Journal of the American Medical Association 298: 2487-2496.
Young J, De SA, Merenstein D, van Essen GA, Kaiser L, Varonen H, Williamson I, et al (2008) Antibiotics for adults with clinically diagnosed acute rhinosinusitis: a meta-analysis of individual patient data. Lancet 371: 908-914.
Eli Hosoien (1), Anne B Lund (2) and Ottar Vasseljen (3)
(1) Physiotherapy Private Practice, Roros, (2) Roros Medical Centre, (3) Norwegian University of Science and Technology Norway
Table 1. Baseline characteristics of participants, therapists and centres. Characteristic Randomised (n = 48) Exp Con (n = 24) (n = 24) Participants Age (yr), mean (SD) 42 (15) 47 (14) Gender, n females (%) 17 (71) 17 (71) Smokers, n (%) 5 (21) 7/23 (30) Duration of current symptoms, n (%) > 7 days 20 (83) 17 (71) > 30 days 4 (17) 2/23 (9) First episode of sinusitis, n (%) 3 (13) 4 (17) More than one episode per year, n (%) 7 (29) 9/23 (39) Previous intervention, n (%) Antibiotics 21 (88) 20 (87) Acupuncture 4 (18) 2 (13) Homeopathy 3 (14) 5 (29) Nasal spray 20 (87) 15 (79) Saline nasal irrigation 14 (67) 6 (33) WBC count (%), mean (range) Lymphocytes * 25 (14-34) 23 (9-37) Granulocytes (neutrophils) ** 70 (59-83) 72 (58-88) Therapists, n participants (%) Physiotherapist 24 (100) 0 (0) General practitioner 0 (0) 24 (100) Centres, n participants (%) Physiotherapy private practice 24 (100) 0 (0) General medical practice 0 (0) 24 (100) Characteristic Lost to follow-up (n = 6) Exp Con (n = 3) (n = 3) Participants Age (yr), mean (SD) 36 (6) 60 (15) Gender, n females (%) 3 (100) 2 (67) Smokers, n (%) 1 (33) 2 (67) Duration of current symptoms, n (%) > 7 days 2 (67) 3 (100) > 30 days 1 (33) 1 (33) First episode of sinusitis, n (%) 0 (0) 0 (0) More than one episode per year, n (%) 1 (33) 1 (33) Previous intervention, n (%) Antibiotics 3 (100) 3 (100) Acupuncture 2 (67) 0 (0) Homeopathy 1 (33) 0 (0) Nasal spray 3 (100) 2 (67) Saline nasal irrigation 1 (33) 1 (33) WBC count (%), mean (range) Lymphocytes * 30 (25-33) 19 (9-25) Granulocytes (neutrophils) ** 65 (60-72) 76 (68-88) Therapists, n participants (%) Physiotherapist 3 (100) 0 (0) General practitioner 0 (0) 3 (100) Centres, n participants (%) Physiotherapy private practice 3 (100) 0 (0) General medical practice 0 (0) 3 (100) Exp = experimental group (ultrasound), Con = control group (antibiotics), WBC = White blood cell, * = High values suggest viral infection (normal reference range for adults is 21-43), ** = High values suggest bacterial infection (normal reference range for adults is 47-75) Table 2. Mean (SD) of groups, mean (SD) difference within groups, and mean (95% CI) difference between groups. Outcome Groups Day 1 Day 4 Exp Con Exp Con (n = 24) (n = 24) (n = 23) (n = 24) Pain NRS (0 to 10) Nose# 5.0# 4.2# 1.7# 2.4# (2.2)# (2.3)# (1.3)# (1.9)# Forehead 4.8 4.9 1.9 2.3 (2.3) (2.0) (1.5) (1.8) Teeth 4.2 3.0 1.0 1.5 (3.4) (3.0) (1.2) (2.1) Congestion NRS (0 to 10) Nose 6.3 5.8 2.6 3.0 (2.1) (1.7) (2.0) (1.7) Forehead 5.2 5.1 1.7 2.3 (2.5) (1.5) (1.7) (1.9) Outcome Groups Day 21 Exp Con (n = 21) (n = 24) Pain NRS (0 to 10) Nose# 1.2# 0.8# (1.8)# (0.9)# Forehead 1.1 0.5 (1.6) (0.9) Teeth 0.6 0.2 (1.7) (0.5) Congestion NRS (0 to 10) Nose 1.6 1.5 (2.0) (1.0) Forehead 0.8 0.5 (1.3) (0.8) Outcome Difference within groups Day 4 minus Day 1 Day 21 minus Day 1 Exp Con Exp Con Pain NRS (0 to 10) Nose# -3.3# -1.8# -3.9# -3.4# (1.6)# (1.6)# (2.1)# (2.0)# Forehead -2.8 -2.5 -3.6 -4.4 (2.2) (1.5) (2.1) (1.8) Teeth -3.0 -1.5 -3.4 -2.8 (2.7) (3.3) (3.2) (2.7) Congestion NRS (0 to 10) Nose -3.5 -2.8 -4.6 -4.3 (2.3) (1.7) (2.3) (1.9) Forehead -3.3 -3.0 -3.9 -4.7 (2.5) (1.7) (2.4) (1.4) Outcome Difference between groups Day 4 minus Day 1 Day 21 minus Day 1 Exp minus Con Exp minus Con Pain NRS (0 to 10) Nose# -1.5# -0.5# (-2.5 to-0.6)# (-1.7 to 0.7)# Forehead -0.2 0.8 (-1.3 to 0.9) (-0.4 to 2.0) Teeth -1.5 -0.6 (-3.3 to 0.3) (-2.4 to 1.1) Congestion NRS (0 to 10) Nose -0.7 -0.3 (-1.9 to 0.4) (-1.5 to 1.0) Forehead -0.3 0.8 (-1.6 to 1.0) (-0.4 to 2.0) Exp = experimental group (ultrasound), Con = control group (antibiotic), shaded row = primary outcome, NRS = numeric rating scale Note: Primary outcome indicated with#. Table 3. Number of participants (%) in each group * for satisfaction, side-effects, relapse, and preferred future intervention, and relative risk (95 % CI) between groups at one year. Outcome Groups Relative risk between groups Exp Con Exp relative (n = 21) (n = 21) to Con Satisfaction with 11/18 (61) 18/19 (95) 0.77 intervention, n yes (%) (0.50 to 1.04) Preferred future 12/16 (63) 3/11 (17) 2.75 intervention, n same as (1.19 to 7.91) allocated (%) Side-effects of 3/21 (14) 4/20 (20) 0.71 intervention, n (%) (0.20 to 2.56) Relapse, n (%) 11/21 (52) 6/21 (29) 1.83 (0.87 to 4.12) * = excluding participants who answered 'uncertain', Exp = experimental group (ultrasound), Con = control group (antibiotics)
|Printer friendly Cite/link Email Feedback|
|Author:||Hosoien, Eli; Lund, Anne B.; Vasseljen, Ottar|
|Publication:||Australian Journal of Physiotherapy|
|Article Type:||Clinical report|
|Date:||Mar 1, 2010|
|Previous Article:||Education improves bra knowledge and fit, and level of breast support in adolescent female athletes: a cluster-randomised trial.|
|Next Article:||Conical-PEP is safe, reduces lung hyperinflation and contributes to improved exercise endurance in patients with COPD: a randomised cross-over trial.|