Vet Res Commun DOI 10.1007/s11259-012-9541-y
ORIGINAL ARTICLE
Daily variability of forced oscillometry parameters in horses suffering recurrent airway obstruction, a pilot study Ali Cesur Onmaz & Christina Stoklas-Schmidt & René van den Hoven
Accepted: 4 October 2012 # Springer Science+Business Media Dordrecht 2012
Abstract The aim of the study was to analyse the day-today variability of the respiratory resistance (Rrs) and the reactance (Xrs) in 5 horses in a status of remission of recurrent airway obstruction by forced oscillometry system (FOS). Furthermore, the effects of stabling, outdoor and room temperature and humidity on these parameters were determined. Mean Rrs at oscillation frequencies 1, 1.5, 2 and 3 Hz were all significantly higher in the morning than in the afternoon, while Xrs was not significantly affected by time of the day. Rrs was significantly different on various sampling days at all frequencies, in the morning as well as in the afternoon. Xrs showed the same significant differences for frequencies greater than 1 Hz. Overall, Rrs showed a smaller variability than Xrs. Neither outdoor temperature, nor room temperature affected Rrs. This also applied to Xrs except for the measurements at 2 Hz that were significantly (p00.044) affected by outdoor temperature. Turning out these horses at paddock for day and night caused a significant lower mean Rrs that when kept at stable for all night. While Rrs showed a very weak positive frequency dependency, Xrs clearly showed a positive frequency dependance for all horses. Resonance frequency (fres) was between 2 and 4 Hz. These studies showed that FOS generated parameters are affected by environmental and management factors and thus only after well-designed standardized procedures FOS could be a
A. C. Onmaz (*) Department of Internal Medicine, Faculty of Veterinary Medicine, University of Erciyes, 38039, Kayseri, Turkey e-mail:
[email protected] C. Stoklas-Schmidt : R. van den Hoven Department of Small Animals and Horses, Equine Clinic, Section Internal Medicine and Infectious Diseases, Veterinary University of Vienna, Vienna 1210, Austria
useful diagnostic tool for the diagnosis and monitoring of equine respiratory disorders. Keywords Horse . Lung function . Oscillometry . Variance
Introduction The conventional pulmonary function test (PFT) as reviewed by Derksen and Robinson (1980) involves inserting an oesophageal balloon catheter via the nasal passage. However, if frequent measurements must be made, this procedure renders horses apprehensive and thus sedation becomes inevitable. Since sedatives affect the mechanics of breathing (Lavoie et al. 1992), blunting of measurements will occur. Moreover, PFT appeared not a very sensitive tool for detecting smaller changes in diseased airways (Robinson et al. 2000). Thus, alternative systems to measure changes in lung function parameters have been developed or are still under study (Desager et al. 1991; MacLeod and Birch 2001; Klein et al. 2006; Hoffman 2002, Winkler et al. 2009). Forced oscillation techniques are rapid and sensitive noninvasive techniques for measuring disease-induced changes in respiratory mechanics in many different species including human and domestic animals (Oostveen et al. 2003). Standardisation procedures for these techniques in human pulmonology have already been proposed (Oostveen et al. 2003). The principle of these techniques is that external forces are superimposed on the normal breathing via mouthpieces or piece in a face mask. This external energy is produced either by a loudspeaker in the impulse oscillometric systems (IOS) or by air pressure pulses in forced oscillometric mechanics. The generated oscillations of air flow and pressure are measured and the resistance against change, the so called respiratory mechanical impedance (Zrs) is calculated. Zrs is a complex vector variable that is
Vet Res Commun
composed of a real part, the resistance (Rrs), and an imaginary part, the reactance (Xrs). Xrs is considered as the sum of the elastance (the reciprocal of compliance) and the inertance. At low oscillation frequencies the elastance determines the magnitude of Xrs; whereas at higher frequencies the inertance becomes the determining factor. Because of its dependence on oscillation frequency, the resonance frequency (fres) of the respiratory system can be determined too. This may be helpfull to know, since values of Xrs measured at frequencies below fres represent the elastance. It is precisely the “heterogeneity” of airway function that is detected up by oscillometry (Hoffman and Mazan 1999). Horses suffering low-grade small airway disease or having agonist-induced bronchoconstriction showed that Rrs is dependent on oscillation frequency (Hoffman et al. 1998; Hoffman and Mazan 1999; Young et al. 1997). IOS and FOS are non-invasive techniques and appear sensitive, especially for the evaluation of peripheral airway obstruction. Hence, it is an attractive alternative to conventional PFT. Despite this potential advantage, relatively few studies have been reported so far (Art and Lekeux 1993; Young and Tesarowski 1994; Tesarowski et al. 1996; Young et al. 1997; Hoffman et al. 1998; Hoffman and Mazan 1999; Mazan et al. 1999; Roberts et al. 1999; MacLeod and Birch 2001; Hoffman 2002; Klein 2002; van Erck et al. 1998, 2003, 2004, 2006, Smith et al. 2005; Winkler et al. 2009; Richard et al. 2009) after the initial work of Young and Hall (1989). The techniques are still not used routinely in equine referral clinics, likely because the most suitable method has not been decided on. Most studies reported results obtained by IOS. Results obtained by air pressure generated oscillations were only given in a few studies (Young and Tesarowski 1994; Hoffman et al. 2001; van Erck et al. 2003, 2004, 2006; Winkler et al. 2009). In this study we used the forced oscillometry system (FOS) as was described by Young and Tesarowski (1994). For FOS Young et al. (1997) showed that measuring at frequencies of 1, 1.5, 2 and 3 Hz was the best option to detect air small air way pathology. Also Hoffman (2002) recommends determining Rrs and Xrs at this frequency range in order to obtain appropriate data of horses suffering heaves. Results shown by Richard et al. (2009) using IOS and the fact that the frequency range is dependent on the size of the lung (Smith et al. 2005) further supports the choice of this frequency range. There is substantial evidence that lung function of horses with recurrent airway obstruction (RAO) measured with conventional PFT shows a daily and a seasonal variance (Stadler and Deegen 1986; Jean et al. 1999). Despite this variance, Jean et al. 1999 reported that these classical PFT measurements were repeatable during short and long periods. On the other hand,
Young (1991) showed that even healthy ponies have marked differences in the respiratory impedance between summer and spring. The degree of observed difference may be explained by the different sensitivity of the applied analytical tools. In order to be useful in pulmonary diagnostics, the system must be sensitive but still robust. The aim of the present pilot study was to quantify the daily variance in resistance (Rrs) and reactance (Xrs) during a winter month (February), using MFOS in horses with variable expression of RAO symptoms. Furthermore, the effects of out-door temperature, out-door humidity and in-door temperature and -humidity were studied. Finally, the effects on Rrs and Xrs of keeping the horses at the paddock or stabled all day were analysed.
Materials and methods Animals Initially 6 experimental horses all owned by the equine hospital of Vienna were used for this study, but after a few measurements, one horse became apprehensive, resulting in poor coherence of measurements. This animal was taken out of the experiment since sedation is no option because this changes airway resistance (Lavoie et al. 1992; Klein et al. 2006). Details of the 5 remaining horses are given in Table 1. Before the measurements started, horses were accustomed to the FOS procedure. During the whole of February, airway mechanics was measured with the FOS on consecutive working days. Twice daily, at 8.00 am and at 3:00 pm lung function was measured. The pulmonary status of 4 horses had been characterized before by conventional PFT and broncho-alveolar lavage (BAL) according to the Criteria for defining the heaves phenotype set by the International Workshop on Equine Chronic Airway Disease 2000 (Robinson 2001). Cells were stained with Wright-Giemsa and Toluidine Blue dye. One horse initially (Nr. 3) considered as healthy, turned out to be a mild sufferer of RAO based on later endoscopic findings and its response to dusty environment. All 5 horses were kept out on a paddock when weather conditions were good. They were fed 6–8 kg medium quality moistened pasture hay and 2–4 kg muesli mix. Half of the ration was given in the morning and half in the late afternoon. When kept in-doors, they were on wood shavings and received moistened hay. The experimental protocol was approved by the Ethical Review Committee of the Veterinary University of Vienna Animal and by the Austrian Ministry of Education, Sciences and Culture, licence number: GZ 68.205/143-BrGT/2004.
Vet Res Commun Table 1 Horses, gender, breed, weight, age and respiratory disease status Horse
Gender
Breed
Weight (kg)
Age (Years)
Disease status (Robinson 2001)
1 2 3
gelding mare gelding
trotter trotter thoroughbred
511 530 488
15 5 16
4 5
gelding mare
warmblood warmblood
566 464
15 16
mild clinical RAO (ΔPplmax 14 cm H2O; BAL:9,300 cell/μl) variable RAO (ΔPplmax 15 cm H2O; BAL: 12,000 cells/μl) subclinical RAO (endoscopic mucus score 2, coughing; increased lung percussion field) subclinical RAO (ΔPplmax 13 cmH2O; BAL: 7,400 cells/μl) mild clinical RAO (episodes of severe respiratory distress; wheezing; TBS: 550/ml; 32–80 % neutrophils)
Methods FOS A face mask was put over the animal’s nose. With an elastic bandage the cheeks were stabilized. Dead space in the system was limited as much as possible and animals were put in a physiological head-neck position (Lavoie et al. 1992). The applied system (On The Nose; Scientific Solutions, Loughborough, UK) has previously been described by Young and Tesarowski (1994). Amplified pressure and flow signals were digitized at 25.6 Hz for 22 s by using a personal computer and a data-acquisition/analysis package (MacADIOS 8ain and Superscope, GW Instruments). The signals were band-pass filtered (12th order digital Butterworth filter with a 0.2-Hzwide passband centred at the measurement frequency) and divided into consecutive 5-s epochs with 50% overlap from which impedance was calculated. Impedance (Zrs) was measured at 1, 1.5, 2 and 3 Hz. Coherence values of ≥0.90 were accepted for 2 to 3 Hz measurements and ≥0.80 for measurements at 1 and 1.5 Hz (Young et al.1997). The pressure generator automatically applied sinusoidal pressure changes of maximal 2 atm at the 4 chosen frequencies for respectively 30, 20, 10 and 10 s. This procedure was repeated 5 times. Data were sampled on a total of 22 working days. No data were sampled during week-ends. All measurements took place in the same room. Horses were always measured in the same order. Statistical analysis Descriptive statistical data are given as mean±standard deviation (sd) if not specified otherwise. Mean and standard error of mean (SE) of Rrs and Xrs are used in graphs. Effects were analysed by a general or by a mixed linear models (PASW 17.0, SPSS Inc.). In a first model, horse was taken as random factor and sampling day and housing conditions as fixed factors. In a second model, indoor and outdoor climate conditions were taken as fixed factors. Both models showed a good fit. Significance was set at p < 0.05.
Furthermore, linear regression analysis was performed of temperature and humidity on Rrs and Xrs. The frequency dependence of Rrs and Xrs was estimated by linear regression. fres was estimated by linear regression of Xrs on frequency, the value of the intersection of the calculated line with the X-axis represented the fres.
Results The daily mean Rrs for 1.5 Hz is shown in Fig. 1 for the days that horses had been kept insides and days when they were 24 h outdoors. Similar patterns were seen for the other frequencies. Mean Rrs was significantly different between morning and afternoon for all oscillation frequencies (Table 2). No significant difference between morning and afternoon was calculated for Xrs. The coefficients of variation (CV) of Rrs and Xrs for each horse at each test frequency are given in Table 3. Both Rrs and Xrs show large variance, Xrs even more than Rrs, especially at 2 and 3 Hz. At all test frequencies the effect of sampling day on Rrs was significant for measurements in the morning (1 Hz: p
