Differences in pulmonary function measured by oscillometry between horses with mild–moderate equine asthma and healthy controls

Abstract

Background

The diagnosis of mild–moderate equine asthma (MEA) can be confirmed by airway endoscopy, bronchoalveolar lavage fluid (BALf) cytology, and lung function evaluation by indirect pleural pressure measurement. Oscillometry is a promising pulmonary function test method, but its ability to detect subclinical airway obstruction has been questioned.

Objectives

To evaluate the differences in lung function measured by oscillometry between healthy and MEA‐affected horses.

Study design

Prospective case–control clinical study.

Methods

Thirty‐seven horses were divided into healthy and MEA groups, based on history and clinical score; the diagnosis of MEA was confirmed by airway endoscopy and BALf cytology. Horses underwent oscillometry at frequencies ranging from 2 to 6 Hz. Obtained parameters included whole‐breath, inspiratory, expiratory, and the difference between inspiratory and expiratory resistance (Rrs) and reactance (Xrs). Differences between oscillometry parameters at different frequencies were evaluated within and between groups by repeated‐measures two‐way ANOVA and post hoc tests with Bonferroni correction. Frequency dependence was compared between groups by t test. For significant parameters, a receiver operating characteristics curve was designed, cut‐off values were identified and their sensitivity and specificity were calculated. Statistical significance was set at p < 0.05.

Results

No significant differences in Xrs and Rrs were observed between groups. The frequency dependence of whole‐breath and inspiratory Xrs significantly differed between healthy (respectively, −0.03 ± 0.02 and −0.05 ± 0.02 cmH₂O/L/s) and MEA (−0.1 ± 0.03 and −0.2 ± 0.02 cmH₂O/L/s) groups (p < 0.05 and p < 0.01). For inspiratory Xrs frequency dependence, a cut‐off value of −0.06 cmH₂O/L/s was identified, with 86.4% (95% CI: 66.7%–95.3%) sensitivity and 66.7% (95% CI: 41.7%–84.8%) specificity.

Main limitations

Sample size, no BALf cytology in some healthy horses.

Conclusions

Oscillometry can represent a useful non‐invasive tool for the diagnosis of MEA. Specifically, the evaluation of the frequency dependence of Xrs may be of special interest.

1. INTRODUCTION

Mild–moderate equine asthma (MEA) is a disease affecting the lower airways of horses, characterised by mild inflammation, subclinical airflow obstruction, excessive tracheobronchial mucus accumulation, and airway hyperreactivity. Clinical manifestations of MEA are typically subtle, including reduced athletic performance and chronic occasional or intermittent coughing. Therefore, ancillary procedures are essential for its diagnosis. These involve detecting tracheal mucus accumulation at airway endoscopy, observing mild increases in neutrophils, eosinophils, and/or mast cell percentages in the bronchoalveolar lavage fluid (BALf), and finding no evidence of airflow limitation based on the oesophageal balloon catheter technique. ¹ However, these methods require physical and/or pharmacological restraint. Moreover, they identify lower airway inflammation but not the subclinical airway obstruction responsible for the clinical signs. Thus, the development of non‐invasive techniques, not requiring sedation, and capable of identifying airway obstruction, represents a pivotal topic in respiratory research. ²

In human asthma, the detection of variable airflow limitation by spirometry or other pulmonary function tests (PFTs) is considered the diagnostic gold standard. ³ In equine medicine, various PFTs have been reported, including indirect transpleural pressure measurement by the oesophageal balloon catheter technique, spirometry, electrical impedance tomography, and oscillometry. Based on the method used to generate the oscillations, oscillometry techniques can be divided into forced oscillation technique (FOT) and impulse oscillometry system (IOS): the former uses monofrequency sinusoidal waves measured at multiple separated frequencies, while the latter consists of a train of pulses that decompose into a continuous frequency spectrum. ⁴ While most PFTs have demonstrated good sensitivity for the detection of severe airway obstructions, ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² contrasting results have been reported by different authors on their efficacy in identifying milder obstructive forms. Indeed, the oesophageal balloon catheter technique showed higher respiratory resistance in horses with MEA compared with healthy controls in one study, ⁵ but not in another one applying the same methodology. ¹³ In a study using spirometry, the only altered pulmonary function parameter in MEA‐affected horses was the forced expiratory flow rate at 95% of expired forced vital capacity, which was significantly decreased compared with their healthy counterparts. ⁵ However, both the oesophageal balloon catheter technique and spirometry are not easily applicable in clinical conditions: indeed, the former requires the use of a facemask, a pneumotachometer, and an oesophageal balloon connected to a pressure transducer, while the latter requires deep sedation, the use of a nasotracheal tube, and mechanical ventilation. ⁵ Electrical impedance tomography has the advantage of being completely non‐invasive, more easily applicable in daily practice, and portable, but lacks sensitivity to detect mild obstructions. ¹⁴ Finally, FOT and IOS do not need sedation and are generally well‐tolerated by horses, ¹⁵ but they are complex techniques not currently commercially available, which has mostly limited their application to research settings. ² However, they have demonstrated increased inspiratory and expiratory resistance in MEA‐affected clinical horses. ⁷ , ¹⁶

Recently, a novel FOT device has been developed and preliminarily validated. ¹⁷ This device is wireless, portable, compact and, therefore, potentially applicable in clinical and field settings. The aim of the present study was to evaluate whether the new device can detect pulmonary function differences between healthy horses and those affected by MEA, and assess its sensitivity in diagnosing horses with MEA.

2. MATERIALS AND METHODS

2.1. Study design

This was a prospective case–control clinical study. According to the guidelines on the minimum sample size required for diagnostic studies, ¹⁸ and based on a prevalence of MEA up to 80%, ¹⁹ a sample size of 39 horses was required to achieve a minimum power of 80% (actual power = 80.7%) to detect a change in the percentage value of sensitivity from 0.70 to 0.90, based on a target significance level of 0.05 (actual p = 0.048). ¹⁸ Horses were selected among a population of clinical patients examined at the Equine Unit of the Veterinary Teaching Hospital and/or referred to the Equine Sports Medicine Laboratory of the University of Milan between January and August 2023, for which owners or caretakers had given informed consent. The horses varied in age, sex, and breed, and were referred for various purposes, including clinical check‐up examinations, athletic fitness assessment, or respiratory tract evaluation. Because of the clinical nature of the study and the limited period of enrolment, we were able to recruit a total of 37 horses.

2.2. Clinical history and examination

For all horses, clinical history was collected, and a questionnaire‐based 1–4 score for respiratory symptoms was assigned according to the HOARSI scoring system. ²⁰ A complete physical examination was performed, and a 0–23 score was assigned to the respiratory clinical signs using the weighed 23‐point scoring system. ²¹ Horses with a HOARSI <3/4 and a clinical score <15/23 were included in the study. A HOARSI score ≥3/4 and/or a clinical score ≥15/23 suggested a possible diagnosis of severe asthma, and were considered as exclusion criteria. Other exclusion criteria included: the administration of corticosteroids or bronchodilators within 14 days before examination (which could have altered clinical signs, airway endoscopy, BALf cytology, and oscillometry parameters), presence of signs of systemic illness (i.e., fever, anorexia, depression, etc.), current or historical presentation of dyspnoea at rest, history of respiratory noises compatible with upper airway obstructions, and present or historical detection of upper airway obstruction during resting or dynamic endoscopy.

Horses with a HOARSI score of 1/4 and a respiratory clinical score of 0/23 were included as healthy controls (n = 15). Among these, five horses had a recent history of reduced athletic performance and underwent a comprehensive diagnostic protocol to identify the underlying cause, which included, among other ancillary procedures, airway endoscopy and BALf cytology to rule out lower airway inflammation. Horses were confirmed as healthy from a respiratory point of view if they presented a tracheal mucus accumulation score <2/5 and BALf cytology results showing neutrophils ≤10%, mast cells ≤5%, and eosinophils ≤5%. ¹ The other 10 horses included in the healthy group were referred for disorders not related to the respiratory tract (i.e., gastrointestinal, metabolic, dermatological, and musculoskeletal disorders).

Horses with a HOARSI of 2/4 and/or a respiratory clinical score between 1/23 and 14/23 were considered possibly affected by MEA and underwent airway endoscopy and BALf cytology for diagnostic confirmation. The diagnosis was confirmed when horses had a tracheal mucus score ≥2/5 and a BALf cytology consisting of neutrophils >10%, and/or mast cells >5%, and/or eosinophils >5% ¹ (n = 22).

2.3. Oscillometry

Pulmonary function was tested in all horses using a recently validated oscillometry device, based on the FOT method. ¹⁷ During the tests, horses were either secured in stocks or kept in their boxes, with their heads loosely contained with halter and lead rope to maintain a neutral position and prevent any interference with spontaneous breathing. Measurements were taken at the horse's airway opening by applying small‐amplitude sinusoidal pressure waveforms with frequencies ranging from 2 to 6 Hz, through a custom‐designed horse mask able to superimpose 1.5 cmH₂O (peak‐to‐peak) oscillations to horses' breathing. Pressure and flows were sampled at 250 Hz using two 12‐bit Successive Approximation Register Analog to Digital Converters (MCP3201; Microchip) after being filtered with a low pass filter with a cut‐off frequency of 50 Hz. Each stimulation frequency was applied for 30 s. The pressure, flow, and stimulation frequency signals sampled by the device were stored on a laptop calibrated in cmH₂O, L/s, and Hz as time series for subsequent analysis. Oscillometry measurements lasted for a total of 2.5 min and were performed once for each horse. If the horse moved its head or any other potential interference occurred during the measurement, the oscillometry measurement was repeated to obtain clear tracings and limit artefacts. The measurements were analysed using MATLAB (MATLAB; MathWorks). Pressure and flow digital values were filtered with a 4th‐order Butterworth bandpass filter (±1 Hz around the stimulation frequency) and used to estimate the impedance using the least square algorithm. The flow signal was digitally integrated to obtain tidal volume, which was then used to identify the inspiratory and expiratory phases (vertical lines in Figures 1 and 2 represent the beginning and end of each identified breath, and the red marks represent the end of the identified inspiratory phase). Breaths with respiratory rates higher than 45 bpm were considered outliers and excluded from further analysis. The identified and accepted breaths were used to compute intra‐breath Rrs and Xrs. Inspiratory Rrs and Xrs were calculated by dividing each inspiratory phase into three equal‐duration parts and by computing the median and IQR of the central third only. Expiratory Rrs and Xrs were calculated using the same approach on the identified expiratory phases. Total resistance and reactance were computed for the duration of the whole breath. ΔRrs and ΔXrs were calculated by the difference between the inspiratory and expiratory values. Breaths with a median flow shape index or pressure shape index higher than 0.2 were excluded. Further outliers were manually excluded by visual inspection of the pressure, flow, Rrs, and Xrs traces (Figure 1). The average values for inspiratory, expiratory, total, and Δ Rrs and Xrs of the whole measurement were calculated as mean ± SD of the accepted breaths. Data from all horses were managed by the same operator (D.B.), blinded to the group allocation of horses and unaware of their diagnosis. Frequency dependence of Rrs and Xrs, indicative of distal obstruction, was calculated by the difference between values obtained at 2 and 6 Hz (respectively, Rrs₂ – Rrs₆ and Xrs₂ – Xrs₆), ²² , ²³ during inspiration, expiration, and whole‐breath.

FIGURE 1.

Example of traces of a control group horse's flow, pressure, Rrs, Xrs, and volume of a measurement at 4 Hz. The vertical lines indicate the beginning and end of each identified breath; the red cross is the identified end‐inspiratory sample. The median inspiratory and expiratory resistance and reactance are computed by considering the central third of each phase.

FIGURE 2.

Example of traces of a MEA group horse's flow, pressure, Rrs, Xrs, and volume of a measurement at 4 Hz. The vertical lines indicate the beginning and end of each identified breath; the red cross is the identified end‐inspiratory sample. The median inspiratory and expiratory resistance and reactance are computed by considering the central third of each phase.

2.4. Airway endoscopy, BALf collection and cytology

Airway endoscopy and BALf collection were performed in horses considered possibly affected by MEA based on HOARSI and clinical scores for diagnostic confirmation (22/22 MEA‐affected horses), and in those presenting with poor athletic performance to rule out subclinical lower airway inflammation (5/15 healthy horses). Endoscopy and BALf collection were performed after oscillometry, by two operators blinded to the oscillometry results (L.S. and C.M.L.F.). Before endoscopy, horses were sedated with detomidine hydrochloride (0.01 mg/kg IV) and restrained with a twitch. The endoscope was passed through the nasal passages and the upper and lower tracts of the respiratory system were examined. ²⁴ A 0–5 score was assigned to tracheal mucus accumulation by one operator blinded to the oscillometry results (L.S.). ²⁵ For BALf collection, 60 mL of 0.5% lidocaine hydrochloride was sprayed at the level of the carina to suppress the coughing reflex, and the endoscope was advanced into the bronchial tree, until firmly wedged. Then, 300 mL of sterile saline 0.9% was instilled and re‐aspirated. An aliquot of 10 mL was preserved in an EDTA tube. Within 90 minutes, 300 μL of pooled BALf was cytocentrifuged at 25 g for 5 minutes. The resulting slides were air‐dried, stained with May‐Grünwald Giemsa, and a 400‐cell leukocyte differential count was performed by one operator blinded to the oscillometry results (C.M.L.F.). ²⁴

2.5. Data analysis

Data were gathered on an electronic spreadsheet (Excel; Microsoft Corp) and analysed using two statistical software packages (Prism 10.0.3; GraphPad Software and JASP 0.17.3; JASP Team). The normality of data distribution was assessed by the Shapiro–Wilk test, and descriptive statistics were performed. Normally distributed data are expressed as mean and 95% confidence interval (CI), while non‐normally distributed data are presented as median and interquartile ranges (IQR). Age was compared between healthy and MEA groups by the Mann–Whitney test, whereas sex and breed distributions were compared between groups by the Chi‐square test. The number of accepted breaths from the two groups was compared by unpaired Student's t test. Differences in lung function at different frequencies within and between the healthy and MEA groups were evaluated using repeated‐measures two‐way ANOVA with the Geisser–Greenhouse correction, weighed on the age variable. In case of significant time effect, group effect, or group × time effect, a post hoc test was performed with Bonferroni correction for multiple comparisons. Finally, the values of frequency dependence of whole‐breath, inspiratory, and expiratory Xrs and Rrs were compared between groups using the t test, adjusted on the age variable. For the parameters significantly differing between the healthy and MEA groups, a receiver operating characteristics (ROC) curve was used to quantify the diagnostic accuracy of these parameters and their capacity to discriminate between healthy and MEA‐affected horses. The parameters showing a significant area under the curve (AUC) ²⁶ were selected for calculating optimal cut‐off values for the diagnosis of MEA, according to the closest‐to‐(0,1) criterion. ²⁷ Sensitivity and specificity of the identified cut‐off values were calculated from the ROC analyses. Statistical significance was set at p < 0.05.

3. RESULTS

3.1. Horses

Among the 37 horses included in the study, 15 were included in the healthy group, while 22 were included in the MEA group. Healthy horses were aged between 2 and 14 years (median 6, IQR 4–11 years), and included 9 mares, 4 geldings, and 2 stallions, of various breeds (5 Thoroughbreds, 4 Standardbreds, 4 Warmbloods, and 2 Quarter Horses). Horses in the MEA group were aged between 2 and 12 years (median 4, IQR 3–7 years), and comprised 4 mares, 8 geldings, and 10 stallions, including 11 Thoroughbred, 7 Standardbreds, 2 Arabians, 1 Warmblood, and 1 Appaloosa. No statistically significant differences in age or breed distribution were observed between the healthy and MEA groups. Conversely, a significant difference in sex distribution was observed (p = 0.02).

3.2. History and clinical examination

In the healthy group, all horses had a HOARSI of 1/4 (no history of cough, nasal discharge, nor dyspnoea), and a respiratory clinical score of 0/23 (absence of respiratory signs). Conversely, horses in the MEA group had a HOARSI of 2/5 (history of mucous nasal discharge, occasional coughing, or both, but with normal respiratory pattern), and respiratory clinical scores varying from 1/23 to 7/23 (median 4/23, IQR 2/23–5/23).

3.3. Airway endoscopy and BALf analyses

The 5 examined healthy horses presented with poor performance had a median tracheal mucus score of 1 (0–1), while the 22 horses with MEA showed a median score of 3 (2–4). Within the healthy group, the median percentages of BALf leukocytes were: 50 (49–50)% macrophages, 39 (39–41)% lymphocytes, 9 (8–9)% neutrophils, 0 (0–0)% eosinophils, and 2 (2–3)% mast cells. Conversely, in the MEA group, the median percentages were: 47 (41–51)% macrophages, 26 (24–33)% lymphocytes, 19 (16–23)% neutrophils, 0 (0–2)% eosinophils, and 3 (2–5)% mast cells. All horses were negative at BALf bacteriological examination.

3.4. Between and within groups oscillometry variations

The FOT device and the oscillometry measurements were well‐tolerated by all horses, and no horses had respiratory rate >45 bpm. The average accepted breaths were 26.7 (95% CI: 24.7–28.6) for healthy group and 26.5 (95% CI: 24.2–28.8) for MEA group (p = 0.9). Whole‐breath, inspiratory, expiratory, and Δ Xrs and Rrs obtained at each frequency in the healthy and MEA groups are displayed in Figure 3. Regarding reactance, for whole‐breath Xrs a non‐significant effect was observed for frequency × group (p = 0.07). Post hoc test revealed significant differences within the MEA group, with lower values at 2 Hz (mean −0.02, 95% CI: −0.05 to 0.01) compared with 4 Hz (mean 0.06, 95% CI: 0.01–0.10) (p < 0.001), 5 Hz (mean 0.08, 95% CI: 0.02–0.14) (p < 0.001), and 6 Hz (mean 0.09, 95% CI: 0.02–0.16) (p < 0.001), and at 3 Hz (mean 0.02, 95% CI: −0.02 to 0.07) compared with 6 Hz (p < 0.01). Similarly, significant frequency (p = 0.004) and frequency × group effects (p < 0.01) were observed for the inspiratory Xrs. In MEA‐affected horses, inspiratory Xrs at 2 Hz (mean −0.04, 95% CI: −0.07 to −0.004) was significantly lower than values at 3 Hz (mean 0.04, 95% CI: −0.002 to 0.07) (p < 0.001), 4 Hz (mean 0.07, 95% CI: 0.03–0.10) (p < 0.001), 5 Hz (mean 0.1, 95% CI: 0.05–0.14) (p < 0.001), and 6 Hz (mean 0.11, 95% CI: 0.06–0.16) (p < 0.001), and at 3 Hz compared with 5 Hz (p = 0.001) and 6 Hz (p < 0.001). Finally, a significant frequency effect was detected for expiratory Xrs (p < 0.05), with values at 2 Hz (mean −0.03, 95% CI: −0.06 to −0.002) being lower than those at 4 Hz (mean 0.05, 95% CI: −0.01 to 0.10) (p = 0.002), 5 Hz (mean 0.08, 95% CI: 0.02–0.14) (p < 0.001) and 6 Hz (mean 0.07, 95% CI: −0.01 to 0.15) (p < 0.001) in the MEA group. No differences between groups or within the healthy group were observed for any of the reactance variables. Concerning the resistance variables, significant frequency effects were observed for whole‐breath (p = 0.001), inspiratory (p = 0.002), and expiratory Rrs (p < 0.001). In the healthy group, whole‐breath Rrs was lower at 2 Hz (mean 0.49, 95% CI: 0.44–0.54) than at 4 Hz (mean 0.62, 95% CI: 0.54–0.69) (p = 0.04), 5 Hz (mean 0.66, 95% CI: 0.57–0.75) (p < 0.001), and 6 Hz (mean 0.69, 95% CI: 0.59–0.78) (p < 0.001), and at 3 Hz (mean 0.56, 95% CI: 0.50–0.62) compared with 6 Hz (p = 0.04). Inspiratory Rrs was lower at 2 Hz (mean 0.44, 95% CI: 0.39–0.49) compared with 5 Hz (mean 0.55, 95% CI: 0.49–0.60) (p < 0.01) and 6 Hz (mean 0.55, 95% CI: 0.48–0.63) (p < 0.01). Expiratory Rrs was lower at 2 Hz (mean 0.55, 95% CI: 0.48–0.62) than at 5 Hz (mean 0.78, 95% CI: 0.66–0.90) (p < 0.001) and 6 Hz (mean 0.85, 95% CI: 0.69–1) (p < 0.001), and at 3 Hz (mean 0.62, 95% CI: 0.56–0.68) compared with 6 Hz (p = 0.002). In the MEA group, whole‐breath Rrs was lower at 2 Hz (mean 0.45, 95% CI: 0.38–0.53) than at 4 Hz (mean 0.62, 95% CI: 0.54–0.69) (p < 0.001), 5 Hz (mean 0.67, 95% CI: 0.58–0.75) (p < 0.001), and 6 Hz (mean 0.7, 95% CI: 0.63–0.77) (p < 0.001), and at 3 Hz (mean 0.53, 95% CI: 0.45–0.62) compared with 5 Hz (p < 0.001) and 6 Hz (p < 0.001). Inspiratory Rrs was lower at 2 Hz compared (mean 0.39, 95% CI: 0.33–0.45) to 4 Hz (mean 0.5, 95% CI: 0.45–0.55) (p < 0.001), 5 Hz (mean 0.55, 95% CI: 0.50–0.61) (p < 0.001), and 6 Hz (mean 0.59, 95% CI: 0.52–0.65) (p < 0.001), at 3 Hz (mean 0.46, 95% CI: 0.40–0.52) compared with 5 Hz (p < 0.01) and 6 Hz (p < 0.001), and at 4 Hz compared with 6 Hz (p = 0.02). Finally, expiratory Rrs was lower at 2 Hz (mean 0.51, 95% CI: 0.43–0.59) than at 4 Hz (mean 0.71, 95% CI: 0.61–0.81) (p < 0.001), 5 Hz (mean 0.8, 95% CI: 0.66–0.94) (p < 0.001), and 6 Hz (mean 0.83, 95% CI: 0.72–0.94) (p < 0.001), and at 3 Hz (mean 0.64, 95% CI: 0.54–0.74) compared with 5 Hz (p < 0.01) and 6 Hz (p < 0.001). No differences between groups were observed for any of the resistance variables. No significant group or time effects were observed for ΔRrs or ΔXrs.

FIGURE 3.

Oscillometry parameters, displayed as mean and 95% confidence intervals, obtained in 15 healthy and 22 MEA horses at frequencies ranging from 2 to 6 Hz.

3.5. Frequency dependence

The frequency dependence (difference between oscillometry values at 2 Hz and those at 6 Hz) of whole‐breath, inspiratory, and expiratory Xrs and Rrs in the healthy and MEA groups are displayed in Table 1. The values of frequency dependence of whole‐breath and inspiratory Xrs were significantly more negative in the MEA group compared with the healthy group (respectively, p < 0.05 and p < 0.01), showing a stronger frequency dependence of Xrs in MEA‐affected horses. Conversely, no significant differences were observed for the frequency dependences of neither expiratory Xrs nor the Rrs parameters. For the frequency dependence of whole‐breath and inspiratory Xrs, ROC curves were performed. Whole‐breath Xrs frequency dependence showed an AUC of 0.68 (95% CI: 0.50–0.85), which is considered as “poor” based on previously reported interpretation guidelines. ²⁶ A possible cut‐off value of −0.038 cmH₂O/L/s was identified, showing 68.2% (95% CI: 47.3%–83.6%) sensitivity and 66.7% (95% CI: 41.7%–84.8%) specificity. Inspiratory Xrs frequency dependence showed an AUC of 0.75 (95% CI: 0.59–0.92), classified as “fair.” ²⁶ A possible cut‐off value of −0.06 cmH₂O/L/s was identified, showing 86.4% (95% CI: 66.7%–95.3%) sensitivity and 66.7% (95% CI: 41.7%–84.8%) specificity.

TABLE 1.

Frequency dependence of whole‐breath, inspiratory, and expiratory Xrs and Rrs, expressed as mean and 95% confidence intervals, in the healthy and MEA groups.

Frequency dependence parameter	Healthy group	MEA group
Whole‐breath Xrs (cmH2O/L/s)	−0.03 (−0.08 to 0.02)*	−0.11 (−0.17 to −0.05)*
Inspiratory Xrs (cmH2O/L/s)	−0.05 (−0.10 to −0.01)**	−0.15 (−0.19 to −0.1)**
Expiratory Xrs (cmH2O/L/s)	−0.03 (−0.08 to −0.03)	−0.10 (−0.16 to −0.04)
Whole‐breath Rrs (cmH2O/L/s)	−0.20 (−0.28 to −0.12)	−0.25 (−0.33 to −0.17)
Inspiratory Rrs (cmH2O/L/s)	−0.11 (−0.17 to −0.05)	−0.20 (−0.26 to −0.13)
Expiratory Rrs (cmH2O/L/s)	−0.29 (−0.41 to −0.17)	−0.32 (−0.43 to −0.21)

Note: Statistically significant differences between groups are shown as *(p < 0.05) and **(p < 0.01).

4. DISCUSSION

The diagnosis of MEA currently relies on airway endoscopy, BALf collection for cytological examination, and pulmonary function evaluation by indirect transpleural pressure measurement. ¹ The development of non‐invasive and sensitive systems capable of identifying mild airway obstruction is pivotal in equine respiratory medicine. ² With this purpose, oscillometry has shown promise in previous studies; however, to date, its application has been mostly limited to research settings. The present study showed that a novel portable and wireless FOT device, designed for easy use in clinical and field settings, ¹⁷ could detect differences in the pulmonary reactance frequency dependence between healthy and MEA‐affected horses.

First described in 1956 by Dubois et al., ²⁸ oscillometry assesses the respiratory system's response to external superimposed forces, filtered from the signals obtained from spontaneous breathing, ⁴ by measuring resistance and reactance at different frequencies. Resistance reflects the resistive properties of the respiratory system, which depend on the calibre and architecture of the airways, while reactance describes its elastic and inertial properties. ⁷ , ¹² , ²⁸ Both IOS and FOT have demonstrated good sensitivity in horses affected by severe forms of equine asthma, showing increased Rrs and decreased Xrs compared with healthy horses, ⁷ , ¹² , ²⁹ but their use has been confined to research settings because of their complexity and lack of commercial availability. ² The newly developed FOT device used in this study, specifically assembled and preliminarily validated for the equine species, may overcome this limitation. Indeed, this device is based on cost‐effective technology and predominantly employs automated data processing algorithms. ¹⁷ However, in the present study, it failed to reveal significant variations in Rrs or Xrs between healthy horses and those affected by MEA, which contrasts with the findings of two prior oscillometry studies on MEA. Specifically, a study using the FOT reported elevated Rrs within the 1–3 Hz frequency range in MEA‐affected horses compared with controls, with no differences in Xrs. ⁷ In another study, applying the IOS, increased Rrs was observed at frequencies within 1 and 10 Hz, along with decreased Xrs at higher frequencies from 5 to 20 Hz; within‐breath analysis confirmed these outcomes during both inspiratory and expiratory phases. ¹⁶ To properly interpret these results, it is important to note that the resonance frequency in horses typically falls between 2 and 4 Hz, ³⁰ , ³¹ and consequently the frequencies ranging from 1 to 3 Hz are considered the most sensitive for detecting lower airway obstruction. ⁴ , ⁶ Therefore, while both studies found Rrs to be significantly higher at low frequencies, distinctions in Xrs were only detectable in the IOS study at high frequencies: this makes it unlikely that differences in Xrs were indicative of distal airway obstruction. As MEA is a disorder affecting the lower airways, in the present study we decided to consider a lower range of frequencies (from 2 to 6 Hz), analogously to some of the first oscillometry studies. ⁶ , ⁷ Even if we did not observe statistically significant differences between groups, it is noteworthy that, at the 2 Hz frequency, the mean Xrs was negative in horses with MEA and positive in healthy horses, during each phase of the respiratory cycle. A larger population might enable the detection of significant differences that could aid in diagnosing MEA. The absence of observed differences in Rrs in our study population, although different from the aforementioned studies, ⁷ , ¹⁶ aligns with a previous IOS study, which found that horses with severe asthma during the remission phase did not exhibit variations in Rrs compared with healthy horses. ²⁹ Since Rrs depends on airway calibre, ³² our hypothesis is that airway obstruction in the horses with MEA included in our study was too mild to be detected. ² One plausible explanation may reside in the young age of the horses in the MEA group (median age of 4 years). Previous studies have demonstrated progressive airway remodelling in horses with MEA, determining alterations such as epithelial hyperplasia, smooth muscle fibrosis, and thickening of the lamina propria. ³³ , ³⁴ This process may lead to increased lung stiffness and airway narrowing, ³² resulting in worsened pulmonary resistance and reactance. ³⁵ , ³⁶ , ³⁷ However, airway remodelling is a chronic process, strongly correlated with increasing age. ³⁸ Therefore, in our study, while lower airway inflammation was confirmed through BALf cytology, it is possible that pulmonary dysfunction was still too mild to be detected, because of the early phase of the disease. However, this interpretation requires confirmation through future studies.

To assess the differences between inspiratory and expiratory components, we also calculated the Δ values. Indeed, higher ΔXrs values have been reported in horses with severe asthma in exacerbation, ²⁹ as well as in human patients with chronic obstructive pulmonary disease, ³⁹ which may be attributed to tidal expiratory flow limitation, a consequence of the narrowing of airway segments, known as choke points. Choke points result from dynamic compression of the peripheral airways and determine the decrease of expiratory reactance and subsequent increase in ΔXrs. ⁴⁰ In our study, mean ΔXrs values appeared higher in MEA‐affected horses compared with their healthy counterparts at frequencies ranging from 3 to 6 Hz, although no statistical significance was detected. This finding is consistent with results of a previous IOS study on horses with severe asthma during the remission phase, ²⁹ where affected horses did not differ from the healthy ones.

In addition to assessing FOT parameters, we also compared the frequency dependence of healthy and MEA‐affected horses. Both groups showed a significant frequency dependence of Rrs with a positive slope, indicating that Rrs increased with increasing frequency. No significant differences were observed between groups. Consistent with our findings, previous studies reported a positive slope of the frequency dependence of Rrs at low frequencies in both healthy horses ¹² and horses with severe asthma during remission, ³¹ measured respectively by IOS and FOT. Conversely, other IOS and FOT studies have suggested that Rrs does not physiologically change as a function of frequency, whereas horses with severe airway obstruction show a negative frequency dependence of Rrs. ⁶ , ⁷ , ⁴¹ Another FOT study, including only MEA‐affected horses, found that those with coughing exhibited a more negative slope of frequency dependence of Rrs compared with non‐coughing horses. ⁴² Regarding Xrs, in our study, it showed frequency dependence with a positive slope, exclusively in MEA‐affected horses. Moreover, significant differences between groups were observed for the values of frequency dependence of whole‐breath and inspiratory Xrs, but not of expiratory Xrs. A positive‐slope frequency dependence of Xrs, measured by both IOS and FOT, has been reported in both healthy horses and those with lower airway obstruction. ¹⁶ , ³¹ , ⁴¹ However, to the best of the authors' knowledge, no previous studies have investigated differences in the frequency dependence of Xrs between healthy and MEA‐affected horses. For the first time, our study suggests that the evaluation of the frequency dependence of whole‐breath and inspiratory Xrs may be of interest in the diagnosis of MEA and warrants further investigation. Despite the relatively small population size, we aimed to propose potential cut‐off values for distinguishing between healthy and MEA‐affected horses. The results indicated that the frequency dependence of inspiratory Xrs may be a suitable parameter for consideration, identifying the value of −0.06 cmH₂O/L/s as a cut‐off with a good sensitivity (86.4%), even if with an unsatisfactory specificity (66.7%). Therefore, we suggest considering the calculation of this parameter as a preliminary screening tool, emphasising that diagnostic confirmation through endoscopy and BALf cytology remains essential.

4.1. Limitations

Overall, the results of the present study suggest that oscillometry testing could contribute to the diagnosis of MEA in clinical practice. However, limitations persist and further studies are needed to establish its role as a standalone diagnostic tool. Notably, the sample size in our population was found to be too small to determine cut‐off values with robust diagnostic utility. Moreover, data were obtained from one single measurement, and repeatability was not evaluated. Therefore, future studies should involve a larger cohort to attempt to define normal ranges, with a specific focus on inspiratory Xrs frequency dependence values, as indicated by our results.

Various variables beyond health status may influence lung function, including age, sex, and breed. Therefore, in the present study, we adjusted statistical analyses for age, even though it did not significantly differ between healthy and MEA groups. Conversely, groups were not sex‐matched, because our study population consisted of referred clinical patients rather than a standardised research herd. While gender did not affect lung function in one study on horses, ¹² recent preliminary research suggested a potential influence of sex hormones on lung function in mares with severe asthma. ⁴³ Similarly, horses of different breeds and uses were included in this study and, although breed distribution did not differ between groups, the percentage of racehorses was higher in the MEA group. To the authors' knowledge, the influence of breed on lung function has never been investigated and cannot be ruled out.

Another limitation of this study was our inability to perform a BALf on all healthy horses, but only in those showing clinical signs potentially related to respiratory diseases, because of ethical reasons. Therefore, in the remaining patients, the diagnosis of MEA was excluded based on clinical history and physical examination, albeit without cytological confirmation.

Finally, from a more practical point of view, a limitation inherent in the PFTs that depend on the use of a facemask is associated with the varying sizes of horses' muzzles, which differ according to age, breed, and individual head morphology. In our study, this limitation did not have a meaningful impact, as all horses were adult of medium size, and the facemask used had a silicone gasket that ensured a proper fit, preventing significant air leaks. However, a range of facemasks in different sizes may be needed to fit foals, ponies, or larger horses such as draught horses. Additional studies should be conducted to confirm or refute our findings, including a larger, better‐defined and homogeneous population of horses.

5. CONCLUSIONS

In conclusion, the present study showed no significant differences in oscillometry parameters between healthy and MEA‐affected horses. However, noteworthy variations in Xrs as a function of frequency were observed only in horses with MEA, which was confirmed by the differences in frequency dependence values of whole‐breath and inspiratory Xrs between the two groups. Therefore, calculating the difference between Xrs at 2 and 6 Hz may offer an indication for the diagnosis of MEA, yet confirmation is still necessary. These findings highlight the potential utility of the novel FOT device in the assessment of MEA, which merits further investigation. Future studies on larger and more standardised populations of horses are warranted to validate and expand our preliminary observations.

FUNDING INFORMATION

Not applicable.

CONFLICT OF INTEREST STATEMENT

The authors have declared no conflicting interests.

AUTHOR CONTRIBUTIONS

Chiara Maria Lo Feudo: Conceptualization; investigation; writing – original draft; methodology; writing – review and editing; formal analysis; data curation. Francesco Ferrucci: Visualization; writing – review and editing; project administration; supervision. Davide Bizzotto: Validation; software; data curation. Raffaele Dellacà: Validation; software; data curation; supervision; visualization. Jean‐Pierre Lavoie: Writing – review and editing; visualization; supervision. Luca Stucchi: Conceptualization; investigation; methodology; writing – review and editing; visualization.

DATA INTEGRITY STATEMENT

Chiara Maria Lo Feudo had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of data analysis.

ETHICAL ANIMAL RESEARCH

Ethical approval for the study was obtained from the Animal Welfare Organisation of the University of Milan (Protocol Number OPBA_136_2022).

INFORMED CONSENT

Written informed consent for the inclusion of patients was obtained from horse owners or trainers.

Lo Feudo CM, Ferrucci F, Bizzotto D, Dellacà R, Lavoie J‐P, Stucchi L. Differences in pulmonary function measured by oscillometry between horses with mild–moderate equine asthma and healthy controls. Equine Vet J. 2025;57(3):619–628. 10.1111/evj.14206

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are openly available in Figshare at https://figshare.com/s/7ab430f0b524347d9b4a.