- Email: [email protected]
Bronchoalveolar lavage fluid from both lungs in horses: Diagnostic reliability of cytology from pooled samples
The Veterinary Journal 244 (2019) 28–33
Contents lists available at ScienceDirect
The Veterinary Journal journal homepage: www.elsevier.com/locate/tvjl
Original Article
Bronchoalveolar lavage fluid from both lungs in horses: Diagnostic reliability of cytology from pooled samples T. Hermangea,* , S. Le Correa , C. Bizona , E.A. Richardb,c, A. Couroucéa a Unité de nutrition, PhysioPathologie et Pharmacologie (NP3), Ecole nationale vétérinaire, agro-alimentaire et de l’alimentation Nantes-Atlantique (Oniris), Université Bretagne Loire (UBL), Nantes F-44307, France b LABÉO Frank Duncombe, 14053 Caen Cedex 4, France c Normandie Univ, UniCaen, Biotargen, 14280 Saint-Contest, France
A R T I C L E I N F O
A B S T R A C T
Article history: Received 28 February 2018 Received in revised form 17 July 2018 Accepted 4 December 2018
Cytology of bronchoalveolar lavage fluid (BALF) from one lung may not predict findings in the contralateral lung of the same horse. The aim of this study was to determine whether a pooled BALF from both lungs was representative of corresponding individual samples. Fifty-one horses referred for poor performance and/or respiratory signs and for which a BALF was collected from both lungs, were included in the study. Cytology of pooled and individual BALF samples were performed using a masked protocol. Based on clinical signs and individual BALF cytologies, horses were classified as control (CTL), mild equine asthma (mEA), severe equine asthma (sEA) and/or exercise-induced pulmonary haemorrhage (EIPH). No significant difference was observed between pooled and individual BALF samples for all cell types (P > 0.05). Correlations between pooled and individual BALF samples were good (r 0.9) for neutrophil proportions and haemosiderophages/macrophages ratio, and moderate (r 0.4) for metachromatic cell and eosinophil proportions. Similarly, intraclass correlation coefficient (ICC) were good (ICC 0.9) for neutrophil proportions and haemosiderophages/macrophages ratio and substantial (ICC 0.6) for metachromatic cell proportions. Based on threshold values for pooled samples as determined by receiver operating characteristic (ROC) analysis, categorical agreements were good (k 0.97) for diagnosis of mEA/sEA, and substantial (k = 0.74) for EIPH. Using a pooled BALF sample, only one horse was incorrectly classified as CTL instead of mEA and three horses were classified as EIPH instead of CTL. In conclusion, BALF cytology from pooled sample is representative of both individual lungs, and constitutes a valid method to diagnose EA. © 2018 Elsevier Ltd. All rights reserved.
Keywords: Bronchoalveolar lavage Cytology Equine asthma Exercise-induced pulmonary haemorrhage Pooled sample
Introduction Mild to moderate equine asthma (mEA) and severe equine asthma (sEA), were previously defined as inflammatory airway disease (IAD) and recurrent airway obstruction (RAO), respectively. Equine asthma and exercise-induced pulmonary haemorrhage (EIPH) constitute a common cause of poor performance in horses (Pirie, 2014; Hinchcliff et al., 2015; Couëtil et al., 2016). Diagnosis of equine asthma (EA) is based on (1) clinical signs and (2) tracheal mucus, bronchoalveolar lavage (BAL) fluid (BALF) cytology or pulmonary function testing (Couëtil et al., 2016). Mild and moderate EA has been defined as a neutrophilic and/or mastocytic and/or eosinophilic lower airway
* Corresponding author. E-mail address: [email protected] (T. Hermange). URL: http://mailto:[email protected] (T. Hermange) https://doi.org/10.1016/j.tvjl.2018.12.002 1090-0233/© 2018 Elsevier Ltd. All rights reserved.
inflammation, characterised by occasional poor performance and coughing without increased respiratory effort at rest (Robinson, 2003; Couëtil et al., 2007, 2016). Severe EA has been defined as a marked neutrophilic lower airway inflammation and obstruction with regular to frequent coughing, exercise intolerance and increased respiratory efforts at rest (Couëtil et al., 2016; Pirie et al., 2016). Diagnostic methodologies for EIPH include blood visualisation and quantification during post-exercise tracheobronchoscopy (Birks et al., 2002; Hinchcliff et al., 2005), and quantification of erythrocytes (Epp et al., 2006) or haemosiderophages in the BALF (Doucet and Viel, 2002; Depecker et al., 2015). Performing BAL seems to be the most sensitive diagnostic tool (Sullivan et al., 2015). To date, BALF cytology is considered the reference standard method for diagnosis of non-infectious lower airway diseases (Couëtil et al., 2016). Several authors have previously investigated regional variations in terms of BALF cytology. One study including
T. Hermange et al. / The Veterinary Journal 244 (2019) 28–33
control (CTL) and sEA horses showed no differences between several regions of both lungs (McGorum et al., 1993). Other studies comparing right and left lungs showed statistical differences for metachromatic cells only (Sweeney et al., 1992; Jean et al., 2011). More recently, a study suggested an alternative methodology for BALF cytological investigation, a 5-field leukocyte differential count (Fernandez et al., 2013). However, the reliability of mast cell percentages was acceptable only on slides with high cell density. A prospective study including a large population of clinically healthy French Trotters found that BALF cytology from the right lung was not equivalent to cytology from the left (Depecker et al., 2014). According to this study, between 12.8% and 37.1% of horses would have been incorrectly been classified as ‘control’ instead of mEA when only one lung was sampled, depending on the cut-off used (‘restrictive’: 5% neutrophils, 1% eosinophils, 2% metachromatic cells; or ‘less restrictive’ one: 10% neutrophils, 5% eosinophils, 5% metachromatic cell). Similarly, between 6.7% to 14.3% of horses would have been incorrectly classified as ‘control’ rather than EIPH, when only one lung was sampled (Depecker et al., 2014). Therefore, both right and left lungs should be sampled for a precise cytological diagnosis of mEA and EIPH. However, ultimately, sampling both lungs with distinct cytological analyses substantially increase the costs for the owner. The hypothesis of this study was that one single cytology on a pooled BALF sample from both lungs would be a consistent tool for diagnosis of EA and EIPH. Our aims were to determine the diagnostic reliability of cytology from pooled BALF samples in comparison with individual BALF cytologies from each lung of the same horse. Materials and methods Horses Prospective power calculation (StatMate, GraphPad) for paired comparisons was performed with expected standard deviation of 23% and correlation among pairs of 0.97. A sample size of 50 horses has 80% power to detect a difference of 2.26% with a significance level (alpha) of 0.05. For inclusion in this study, horses had to be referred to the hospital between April 2014 and March 2017 for poor performance (as subjectively defined by the owner/rider) and/or respiratory signs (cough, dyspnoea, nasal discharge, noise during exercise, increased respiratory effort at rest), and examined by at least one of the authors (TH, CB, AC). The decision of performing BAL was made at the discretion of the attending clinician. Prior to any procedure and at least 24 h after the last training or racing (if any), each horse was subjected to a thorough clinical examination. If observed, the presence of cough, nasal discharge (either serous or mucous), nostril flaring at rest, or dyspnoea was recorded. Further examinations were performed at the clinician discretion in order to rule out any other condition, including i.e. haematology, biochemistry and tracheal wash bacterial and mycological culture, and thoracic radiographies. The study was approved by the regional Animal Ethic Committee (CERVO-2016-14-V). A consent form was also signed by the owner. Bronchoalveolar lavage Horses were sedated using detomidine (10 mg/kg IV, Sédomidine, Audevard) and butorphanol (10 mg/kg IV, Dolorex, Intervet). A nose twitch was used as necessary. The BAL was performed using a flexible 3.2 m long, 12.8 mm tip diameter videoendoscope (Optomed) without previous lidocaine instillation. The endoscope was inserted into one main bronchus until wedged. A total of 250 mL (two boluses of 125 mL each) of warmed (37 C) sterile isotonic saline solution was instilled through the endoscope biopsy channel, using 60-mL pre-filled syringes. After each bolus, the liquid was aspirated with the same syringes and the first 20 mL corresponding to the endoscope channel volume that did not reach the lung were disposed of. At the end of the procedure, the biopsy channel was rinsed with 40 mL of saline solution with the endoscope wedged; the endoscope was then pulled back to the carina and inserted into the controlateral lung and the same procedure was repeated. The volume of fluid collected from each lung was recorded. The contents of the syringes corresponding to the left and right lungs were put into two different metal bowls; an aliquot of 4 mL each BALF sample was kept in EDTA tubes. An equivalent volume (50 mL) from each bowl was pooled in another bowl. An aliquot of the pooled sample was then kept in EDTA tubes.
29
BALF analysis All three samples (left, right and pooled BALF) were kept refrigerated until handled by the laboratory within 24 h after collection. Two hundred microlitres of fluid were cytocentrifugated (80 g, 10 min) and stained with May-Grünwald-Giemsa (MGG). A differential cell count was performed on 300 cells and the number of each cell type was recorded as a percentage of total nucleated cells, excluding epithelial cells. The observer performing the cytologic examination was masked to the horse and the type of sample (left lung, right lung, or pooled sample). Case definition According to the revised American College of Veterinary Internal Medicine (ACVIM) consensus on IAD (Couëtil et al., 2016), two definitions were used to characterise normal cytology: ‘restrictive’ characterisation was 5% neutrophils, 1% eosinophils, 2% metachromatic cells; the ‘less restrictive’ characterisation was: 10% neutrophils, 5% eosinophils, 5% metachromatic cells. Horses without coughing, nasal discharge, or increased respiratory effort (dyspnoea or nasal flaring), and with normal BALF cytology from both lungs were classified as CTL. Horses with chronic (>3 weeks) coughing and/or history of poor performance, without increased respiratory effort at rest (nostril flaring or dyspnoea), and with abnormal BALF cytology (neutrophils, and/or eosinophils, and/or metachromatic cells) in at least one lung, were classified as mEA. Horses with chronic (>3 weeks) coughing, increased respiratory efforts at rest (nostril flaring or dyspnoea), and with abnormal BALF cytology (neutrophils only) in at least one lung were classified as sEA. Furthermore, recurrent and reversible airway obstruction for these horses was confirmed either by history or follow-up. Horses with cytology with a haemosiderophage/macrophage (H/M) ratio >17% (Depecker et al., 2015) were classified as EIPH. Statistical analysis Normality of continuous data distribution was evaluated using the Shapiro– Wilk W tests. When necessary, data were log 10 transformed to normalise distribution. Multiple comparisons for cytological parameters were performed using one-way ANOVA, with Dunnett’s post-hoc test. Pearson’s correlation test was performed to evaluate the association between pooled and respectively right and left lung BALF samples. Agreement between pooled sample and both left and right samples, for numerical (cell proportions) and categorical variables (CTL, mEA, sEA, EIPH), respectively, was measured with the intra-class correlation coefficient (ICC) and the Cohen’s kappa coefficient (k). Data were presented as mean standard deviation, unless stated otherwise. Values of P < 0.05 were considered statistically significant. Cut-off values for pooled samples were established based on a receiver operating characteristic (ROC) curve and associated area under the curve (AUC), with combined individual BALF samples used as cytological reference standard. The best cut-off was considered to be the data with the highest true-positive rate (sensitivity) and the lowest false-positive rate (1– specificity), which represents the point closest to the top left corner of the ROC curve.
Results Horses A total of 51 horses (females, n = 20; geldings n = 23; males n = 8), aged 3–22 years old (8.6 4.9 years) were included in the study. Breeds represented were as follows: Warmbloods (n = 17), Trotters (total n = 17; racing n = 9, non-racing, n = 8), Thoroughbreds (total n = 4; racing n = 2, non-racing, n = 2), ponies (n = 8), and miscellaneous horses (n = 5). BALF volume and cell proportions Mean volume proportions of fluid retrieved were 43.2 10.9% for the right lung and 45.7 11.5% for the left lung and were not significantly different between lungs (P = 0.36). Overall, there was no significant difference between pooled BALF and samples from right/left lungs for both neutrophil (P = 0.11) and metachromatic cell proportions (P = 0.07), and H/M ratio (P = 0.50). Eosinophil proportions in pooled BALF were significantly lower when compared with left lung samples (P = 0.02), but not with right lung samples (P = 0.47; Fig. 1).
30
T. Hermange et al. / The Veterinary Journal 244 (2019) 28–33
Fig. 1. Mean standard deviation for (a) neutrophil proportions, (b) metachromatic cells proportions, (c) eosinophil proportions, and (d) Haemosiderophages/Macrophages (H/M) ratio; corresponding mean difference with 95% confidence interval within each group for (a’) neutrophil proportions, (b’) metachromatic cells proportions, (c’) eosinophil proportions, and (d’) H/M ratio. Symbol * indicates P < 0.05.
Correlations and numerical agreements Good correlations between pooled BALF and each lung sample were found for neutrophil proportions and H/M ratio. Significant correlations were fair to moderate for metachromatic cell and eosinophil proportions (Table 1). Numerical agreements between
pooled BALF and right/left lung samples were good for neutrophil proportions and H/M ratio, and substantial for metachromatic cell proportions. For eosinophil proportions, numerical agreement was good between pooled BALF and left lung, and poor between pooled BALF and the right lung (Table 1; Supplementary Fig. S1 in the online version at DOI: 10.1016/j.tvjl.2018.12.002).
T. Hermange et al. / The Veterinary Journal 244 (2019) 28–33
31
Table 1 Correlations (Pearson coefficient, r), numerical agreements (intra-class coefficient, ICC), and corresponding 95% confidence intervals (95% CI), between bronchoalveolar lavage fluid cytology from pooled samples and respective right and left lung. Pooled sample vs. right lung Neutrophils
H/M ratio
Metachromatic cells
Eosinophils
r = 0.95 (95% CI: 0.90–0.98) ICC = 0.93 (95% CI: 0.89–0.96) r = 0.99 (95% CI: 0.99–0.99) ICC = 0.99 (95% CI: 0.98–0.99) r = 0.47 (95% CI: 0.14–0.71) ICC = 0.64 (95% CI: 0.47–0.77) r = 0.35 (95% CI: 0.00–0.63) ICC = 0.00 (95% CI: 0.26–0.25)
Pooled sample vs. left lung P < 0.01 P < 0.01 P < 0.01 P < 0.01 P = 0.01 P < 0.01 P = 0.03 P = 0.95
Left vs. right lung P < 0.01
r = 0.93 (95% CI: 0.86–0.97) ICC = 0.97 (95% CI: 0.94–0.98) r = 0.89 (95% CI: 0.79–0.95) ICC = 0.90 (95% CI: 0.84–0.94) r = 0.37 (95% CI: 0.02–0.65) ICC = 0,68 (95% CI: 0.51–0.80) r = 0.64 (95% CI: 0.36–0.81) ICC = 0.98 (95% CI: 0.97–0.99)
r = 0.89 (95% CI: 0.82–0.94) r = 0.88 (95% CI: 0.83–0.92) r = 0.87 (95% CI: 0.78–0.92) r = 0.85 (95% CI: 0.79–0.90) r = 0.51 (95% CI: 0.28–0.69) r = 0.47 (95% CI: 0.30–0.61) r = 0.014 (95% CI: 0.26–0.29) r = 0.01 (95% CI: 0.20–0.19)
P < 0.01 P < 0.01 P < 0.01 P = 0.01 P < 0.01 P < 0.01 P < 0.01
P < 0.01 P < 0.01 P < 0.01 P < 0.01 P < 0.01 P < 0.01 P = 0.90 P = 0.96
H/M, Haemosiderophages/Macrophage.
Cut-off values for pooled samples Based on the ‘restrictive’ definition, thresholds for pooled samples, as determined by ROC, were set at 3% neutrophils and 1% metachromatic cells. Based on the ‘less restrictive’ definition, the thresholds, as determined by ROC, were set at 9% neutrophils, 3% metachromatic cells. Due to the low number of samples with increased eosinophil proportions, thresholds for pooled sample could not be established. For EIPH, threshold for pooled samples was set at 9% H/M. The corresponding sensitivities and specificities are presented in Table 2 and graphical representations are in Supplementary Fig. S2 in the online version at DOI: 10.1016/j. tvjl.2018.12.002.
definition (Table 3). However, when considering BALF cytology from one lung only (and irrespective of the lung side), five and six horses with mEA were classified as controls when using respectively the ‘restrictive’ and ‘less restrictive’ definition, respectively. There was substantial agreement between pooled BALF sample and combined individual BALF samples (k = 0.74; 95% CI: 0.46– 1.00) for the final diagnosis of EIPH, based on H/M ratio. The same 5/51 (9.8%) horses were classified as EIPH when performing cytology on either both lungs or pooled sample. Three other horses were categorised as EIPH based on the pooled sample, but not on samples from the individual lungs. Table 3 Comparison of final diagnosis based on clinical signs and cytological data from either both lungs or pooled sample.
Final diagnosis and categorical agreement Based on BALF cytology from both lungs separately, classification as CTL, mEA, sEA and EIPH when using either ‘restrictive’ or ‘less restrictive’ definition is detailed in Table 3. Among horses classified as mEA, variable combinations of abnormal BALF cytological profiles were present (Table 4). Among horses classified as sEA, all but one exhibited neutrophil proportions >30% in both lungs (Fig. 2). One horse had 12% neutrophils in BALF from the left lung and 43% from the right lung. Good agreement between pooled BALF samples and combined individual BALF samples was found for the final diagnosis of CTL, mEA or sEA, when using both the ‘restrictive’ and the ‘less restrictive’ definition (k = 1.00 and k = 0.97; 95% CI: 0.91–1.00), respectively. All 51 horses were correctly classified when using the ‘restrictive’ definition and one horse classified as mEA was incorrectly classified as CTL when using the ‘less restrictive’
Groups
CTL mEA sEA EIPH total EIPH only Virtual totalb
‘Restrictive’ definitiona
‘Less restrictive’ definitionb
Left and right BALF samples
Pooled BALF sample
Left and right BALF samples
Pooled BALF sample
n=3 n = 36 n = 11 n=5 n=1 n = 56
n=3 n = 36 n = 11 n=8 n=1 n = 59
n = 11 n = 27 n = 11 n=5 n=2 n = 56
n = 12 n = 26 n = 11 n=8 n=2 n = 59
BALF, bronchoalveolar lavage fluid; CTL, control; mEA, mild equine asthma; sEA, severe equine asthma; EIPH, exercise-induced pulmonary haemorrhage. a This definition is used for the cytological discrimination between CTL and mEA horses only. b A total of 51 horses were included in this study. EIPH was detected in five horses by evaluating both lungs separately (virtual total of 56), and in eight horses when using pooled BALF samples (virtual total of 59).
Table 2 Reliability (sensitivity, specificity, corresponding area under the curve [AUC] and respective 95% confidence intervals [CI]) of cut-off values for pooled bronchoalveolar lavage fluid (BALF) samples, as defined according to the thresholds used for individual BALF samples. Neutrophils Individual threshold Pooled cut-off Sensitivity (95% CI) Specificity (95% CI) AUC (95% CI)
5% 3% 1.00 (0.91–1.00) 0.86 (0.42–1.00) 1.00 (0.98–1.00)
H/M, Haemosiderophages/Macrophages.
Metachromatic cells 10% 9% 0.93 (0.77–0.99) 1.00 (0.78–1.00) 0.98 (0.94–1.00)
2% 1% 0.96 (0.78–1.00) 0.64 (0.44–0.81) 0.88 (0.78–0.97)
H/M 5% 3% 1.00 (0.29–1.00) 0.73 (0.58–0.85) 0.94 (0.84–1.00)
17% 9% 1.00 (0.59–1.00) 0.98 (0.88–1.00) 1.00 (0.98–1.00)
32
T. Hermange et al. / The Veterinary Journal 244 (2019) 28–33
Table 4 Number of horses for each combination of cells classified as mild and moderate equine asthma (mEA), based on bronchoalveolar lavage fluid cytological evaluation from both lungs individually. Definition
‘Restrictive’ ‘Less restrictive’
Neutrophilic mEA
Non-neutrophilic mEA
Total
NEU
NEU + MET
NEU + EOS
NEU + MET + EOS
MET
EOS
MET + EOS
20 23
12 1
0 1
1 0
2 2
0 0
1 0
36 27
NEU, neutrophils; MET, metachromatic cells; EOS, eosinophils.
Fig. 2. Graphic representations of bronchoalveolar lavage fluid neutrophil proportions according to sample type and final diagnosis. CTL, control; mEA, mild to moderate equine asthma; sEA, severe equine asthma.
Discussion To our knowledge, this is the first report of pooled equine BALF samples from both lungs compared to each lung separately. Excellent correlation and numerical agreement were demonstrated for neutrophil proportions between pooled BALF samples and samples from both lungs. For metachromatic cells, correlation and numerical agreement were only moderate. This may partly be explained by their low proportions in BALF, and also by their low prevalence in the population under study (14/51 and 3/51 horses with increased metachromatic cell proportions when using the ‘restrictive’ and the ‘less restrictive’ definition, respectively). For the same reasons, correlation and numerical agreement for eosinophils was also only moderate (prevalence: 2/51 and 1/51 horses with ‘restrictive’ and ‘less restrictive’ definitions, respectively). Furthermore, among the 2/51 horses with BALF eosinophilia (>1%), one horse exhibited 15% eosinophils in the left lung and no eosinophils in the right lung. As 13% eosinophils were found in the pooled BALF sample, the discrepant findings in this horse likely contributed to the moderate correlation and numerical agreement between pooled and right BALF; this limits the clinical relevance of the eosinophil data. The current data confirm those of previous studies, in which good correlations and agreements were found between lungs for BALF neutrophil proportions but not for metachromatic cell and eosinophil proportions (Jean et al., 2011; Depecker et al., 2014). However, neutrophil proportions between the left and right lungs were not significantly different, in contrast with previously reported results (Depecker et al., 2014). This might be explained by comparisons between study populations; horses of any age were referred for respiratory investigation in our study, while prospective evaluation of clinically healthy young racehorses was performed in the previous study. For BALF inflammatory cells, the pooled BALF sample was representative of combined individual BALF samples, particularly for neutrophil proportions. Most cases of mEA are associated with increased neutrophil proportions in BALF (Fogarty and Buckley,
1991; Couëtil et al., 2001; Depecker et al., 2014; Richard et al., 2014), even if cytological profiles of mEA may also include increased proportions of other cell types (metachromatic cells and eosinophils) (Couëtil et al., 2007; Couëtil et al., 2016). However, sEA has been associated with increased proportions of neutrophils only (Pirie, 2014). When considering the final diagnosis of equine asthma, pooled BALF samples were reliable, since excellent categorical agreements were found for the discrimination between CTL, mEA and sEA. Importantly, no discordancy (0/51 horses) was found between methods for the final diagnosis of EA when using the ‘restrictive’ definition, and only 1/51 discordant diagnosis was obtained when using the ‘less restrictive’ definition (one horse with mEA incorrectly classified as CTL). In contrast, sampling one lung only resulted in 6/51 discordant diagnoses. The minimum sample size (50 horses) was determined based on BALF neutrophil proportions. The mEA phenotype (neutrophilic, metachromatic, eosinophilic, mixt) has not been further subdivided for investigation of categorical agreements. Among the 36 horses diagnosed with mEA and according to the ‘restrictive’ definition, 33/36 horses exhibited BALF neutrophilia, with or without increased proportions of either metachromatic cells or eosinophils. Similarly, increased neutrophil proportions were found in 25/27 horses with mEA when using the ‘less restrictive’ definition. Although metachromatic cells and eosinophils correlations and agreements were only moderate, their impact on the final diagnosis of equine asthma was limited. According to the revised ACVIM consensus on IAD (Couëtil et al., 2016), mild and moderate EA were not separated within the ‘mEA group’ in the current study. However, it has recently been suggested that both entities could be separated, based on the presence/absence of coughing in addition to poor performance (Bullone and Lavoie, 2017). In contrast, coughing was a specific but insensitive clinical sign associated with non-severe lower airway inflammation in previous studies (Wasko et al., 2011; Cardwell et al., 2014). Coughing may not be reliability reported by owners/ trainers, leading to some horses being incorrectly classified as ‘mild’ EA instead of ‘moderate’ EA. Three horses presented for poor performance and high neutrophil proportions in BALF (neutrophils >25%) were classified as mEA because of the absence of any respiratory signs. Apart from BALF cytology, this emphasises the importance of clinical investigations (i.e. respiratory effort at rest) to discriminate between mild/moderate and severe EA. However, no sEA horse in the current study exhibited increased respiratory effort at rest associated with low proportions of neutrophils (<25%) in the BALF. This paucigranulocytic phenotype has previously been mentioned by some authors (Bullone and Lavoie, 2017). The absence of discordant diagnoses for sEA between both methods might be explained by the relatively high neutrophil proportions (all samples from horses with sEA except for one had neutrophil proportions >30%). However, 1/11 (9%) horse had 12% neutrophils in the BALF from the left lung and 46% from the right lung. This discrepancy (paucigranulocytic pattern on one lung and neutrophilic on the other one), highlights the relevance of sampling both lungs for a thorough assessment.
T. Hermange et al. / The Veterinary Journal 244 (2019) 28–33
Although categorical agreement was also substantial for the final diagnosis of EIPH, a few discordant results were found between both methods. A total of 3/51 horses were classified as EIPH based on the pooled samples, but were classified as CTL based on individual lungs. However, the presence of false positive results was unlikely, based on previous or concomitant clinical observations. One horse was diagnosed with atrial fibrillation, which is often associated with EIPH (Reef et al., 1995). In another horse, EIPH had been diagnosed (endoscopically) a few months previously; the third horse exhibited mild epistaxis at the end of the exercise test. However, the pooled sample threshold used for the diagnosis of EIPH should be interpreted with caution, since the confidence interval for sensitivity was relatively large (95% CI: 0.59–1.00). This might be associated with the low prevalence of EIPH (5/51 horses) in the current population, which likely reflects the relatively low number of racing horses included in the study (11/51). Conclusions When using a standardised collection procedure, cytology from pooled BALF samples is representative of both individual lungs, according to both correlations and numerical agreements. Based on categorical agreement (final diagnosis), cytology of pooled BALF samples has been validated for equine asthma. Further studies using larger populations of racing horses are required to definitively validate the threshold value for EIPH cytological detection in pooled BALF samples. Conflict of interest statement None of the authors of this paper has a financial or personal relationship with other people or organizations that could inappropriately influence or bias the content of the paper. Acknowledgements Preliminary results were presented to the 9th ECEIM congress 2016, Helsinki, Finland (oral communication) and to the WEAS symposium 2017 (poster), Copenhagen, Denmark. Financial support was provided by LABÉO, CISCO-Oniris and AVEF (veterinary association of French practitioner). References Birks, E.K., Shuler, K.M., Soma, L.R., Martin, B.B., Marconato, L., Del Piero, F., Teleis, D. C., Schar, D., Hessinger, A.E., Uboh, C.E., 2002. EIPH: postrace endoscopic evaluation of Standardbreds and Thoroughbreds. Equine Vet. J. Suppl. 375–378. Bullone, M., Lavoie, J.-P., 2017. Science-in-brief: equine asthma diagnosis: beyond bronchoalveolar lavage cytology. Equine Vet. J. 49, 263–265. Cardwell, J.M., Smith, K.C., Wood, J.L.N., Newton, J.R., 2014. Infectious risk factors and clinical indicators for tracheal mucus in British National Hunt racehorses. Equine Vet. J. 46, 150–155.
33
Couëtil, L.L., Cardwell, J.M., Gerber, V., Lavoie, J.-P., Léguillette, R., Richard, E.A., 2016. Inflammatory airway disease of horses-revised consensus statement. J. Vet. Intern. Med. 30, 503–515. Couëtil, L.L., Hoffman, A.M., Hodgson, J., Buechner-Maxwell, V., Viel, L., Wood, J.L.N., Lavoie, J.-P., 2007. Inflammatory airway disease of horses. J. Vet. Intern. Med. 21, 356–361. Couëtil, L.L., Rosenthal, F.S., DeNicola, D.B., Chilcoat, C.D., 2001. Clinical signs, evaluation of bronchoalveolar lavage fluid, and assessment of pulmonary function in horses with inflammatory respiratory disease. Am. J. Vet. Res. 62, 538–546. Depecker, M., Couroucé-Malblanc, A., Leleu, C., Genneviève, V., Pitel, P.-H., Richard, E.A., 2015. Comparison of two cytological methods for detecting pulmonary haemorrhage in horses. Vet. Rec. 177, 305.2–305. Depecker, M., Richard, E.A., Pitel, P.-H., Fortier, G., Leleu, C., Couroucé-Malblanc, A., 2014. Bronchoalveolar lavage fluid in Standardbred racehorses: influence of unilateral/bilateral profiles and cut-off values on lower airway disease diagnosis. Vet. J. 199, 150–156. Doucet, M.Y., Viel, L., 2002. Alveolar macrophage graded hemosiderin score from bronchoalveolar lavage in horses with exercise-induced pulmonary hemorrhage and controls. J. Vet. Intern. Med. 16, 281–286. Epp, T.S., McDonough, P., Padilla, D.J., Gentile, J.M., Edwards, K.L., Erickson, H.H., Poole, D.C., 2006. Exercise-induced pulmonary haemorrhage during submaximal exercise. Equine Vet. J. Suppl. 502–507. Fernandez, N.J., Hecker, K.G., Gilroy, C.V., Warren, A.L., Léguillette, R., 2013. Reliability of 400-cell and 5-field leukocyte differential counts for equine bronchoalveolar lavage fluid. Vet. Clin. Pathol. 42, 92–98. Fogarty, U., Buckley, T., 1991. Bronchoalveolar lavage findings in horses with exercise intolerance. Equine Vet. J. 23, 434–437. Hinchcliff, K.W., Couetil, L.L., Knight, P.K., Morley, P.S., Robinson, N.E., Sweeney, C.R., van Erck, E., 2015. Exercise induced pulmonary hemorrhage in horses: American College of Veterinary Internal Medicine Consensus Statement. J. Vet. Intern. Med. 29, 743–758. Hinchcliff, K.W., Jackson, M.A., Brown, J.A., Dredge, A.F., O’Callaghan, P.A., McCaffrey, J.P., Morley, P.S., Slocombe, R.E., Clarke, A.F., 2005. Tracheobronchoscopic assessment of exercise-induced pulmonary hemorrhage in horses. Am. J. Vet. Res. 66, 596–598. Jean, D., Vrins, A., Beauchamp, G., Lavoie, J.-P., 2011. Evaluation of variations in bronchoalveolar lavage fluid in horses with recurrent airway obstruction. Am. J. Vet. Res. 72, 838–842. McGorum, B.C., Dixon, P.M., Halliwell, R.E.W., Irving, P., 1993. Comparison of cellular and molecular components of bronchoalveolar lavage fluid harvested from different segments of the equine lung. Res. Vet. Sci. 55, 57–59. Pirie, R.S., 2014. Recurrent airway obstruction: a review. Equine Vet. J. 46, 276–288. Pirie, R.S., Couëtil, L.L., Robinson, N.E., Lavoie, J.-P., 2016. Equine asthma: an appropriate, translational and comprehendible terminology? Equine Vet. J. 48, 403–405. Reef, V.B., Reimer, J.M., Spencer, P.A., 1995. Treatment of atrial fibrillation in horses: new perspectives. J. Vet. Intern. Med. 9, 57–67. Richard, E.A., Depecker, M., Defontis, M., Leleu, C., Fortier, G., Pitel, P.-H., CouroucéMalblanc, A., 2014. Cytokine concentrations in bronchoalveolar lavage fluid from horses with neutrophilic inflammatory airway disease. J. Vet. Intern. Med. 28, 1838–1844. Robinson, N.E., 2003. Inflammatory airway disease: defining the syndrome. Conclusions of the Havemeyer Workshop. Equine Vet. Educ. 15, 61–63. Sullivan, S.L., Whittem, T., Morley, P.S., Hinchcliff, K.W., 2015. A systematic review and meta-analysis of the efficacy of furosemide for exercise-induced pulmonary haemorrhage in Thoroughbred and Standardbred racehorses: furosemide for exercise-induced pulmonary haemorrhage. Equine Vet. J. 47, 341–349. Sweeney, C.R., Rossier, Y., Ziemer, E.L., Lindborg, S., 1992. Effects of lung site and fluid volume on results of bronchoalveolar lavage fluid analysis in horses. Am. J. Vet. Res. 53, 1376–1379. Wasko, A.J., Barkema, H.W., Nicol, J., Fernandez, N., Logie, N., Léguillette, R., 2011. Evaluation of a risk-screening questionnaire to detect equine lung inflammation: results of a large field study. Equine Vet. J. 43, 145–152.
Contents lists available at ScienceDirect
The Veterinary Journal journal homepage: www.elsevier.com/locate/tvjl
Original Article
Bronchoalveolar lavage fluid from both lungs in horses: Diagnostic reliability of cytology from pooled samples T. Hermangea,* , S. Le Correa , C. Bizona , E.A. Richardb,c, A. Couroucéa a Unité de nutrition, PhysioPathologie et Pharmacologie (NP3), Ecole nationale vétérinaire, agro-alimentaire et de l’alimentation Nantes-Atlantique (Oniris), Université Bretagne Loire (UBL), Nantes F-44307, France b LABÉO Frank Duncombe, 14053 Caen Cedex 4, France c Normandie Univ, UniCaen, Biotargen, 14280 Saint-Contest, France
A R T I C L E I N F O
A B S T R A C T
Article history: Received 28 February 2018 Received in revised form 17 July 2018 Accepted 4 December 2018
Cytology of bronchoalveolar lavage fluid (BALF) from one lung may not predict findings in the contralateral lung of the same horse. The aim of this study was to determine whether a pooled BALF from both lungs was representative of corresponding individual samples. Fifty-one horses referred for poor performance and/or respiratory signs and for which a BALF was collected from both lungs, were included in the study. Cytology of pooled and individual BALF samples were performed using a masked protocol. Based on clinical signs and individual BALF cytologies, horses were classified as control (CTL), mild equine asthma (mEA), severe equine asthma (sEA) and/or exercise-induced pulmonary haemorrhage (EIPH). No significant difference was observed between pooled and individual BALF samples for all cell types (P > 0.05). Correlations between pooled and individual BALF samples were good (r 0.9) for neutrophil proportions and haemosiderophages/macrophages ratio, and moderate (r 0.4) for metachromatic cell and eosinophil proportions. Similarly, intraclass correlation coefficient (ICC) were good (ICC 0.9) for neutrophil proportions and haemosiderophages/macrophages ratio and substantial (ICC 0.6) for metachromatic cell proportions. Based on threshold values for pooled samples as determined by receiver operating characteristic (ROC) analysis, categorical agreements were good (k 0.97) for diagnosis of mEA/sEA, and substantial (k = 0.74) for EIPH. Using a pooled BALF sample, only one horse was incorrectly classified as CTL instead of mEA and three horses were classified as EIPH instead of CTL. In conclusion, BALF cytology from pooled sample is representative of both individual lungs, and constitutes a valid method to diagnose EA. © 2018 Elsevier Ltd. All rights reserved.
Keywords: Bronchoalveolar lavage Cytology Equine asthma Exercise-induced pulmonary haemorrhage Pooled sample
Introduction Mild to moderate equine asthma (mEA) and severe equine asthma (sEA), were previously defined as inflammatory airway disease (IAD) and recurrent airway obstruction (RAO), respectively. Equine asthma and exercise-induced pulmonary haemorrhage (EIPH) constitute a common cause of poor performance in horses (Pirie, 2014; Hinchcliff et al., 2015; Couëtil et al., 2016). Diagnosis of equine asthma (EA) is based on (1) clinical signs and (2) tracheal mucus, bronchoalveolar lavage (BAL) fluid (BALF) cytology or pulmonary function testing (Couëtil et al., 2016). Mild and moderate EA has been defined as a neutrophilic and/or mastocytic and/or eosinophilic lower airway
* Corresponding author. E-mail address: [email protected] (T. Hermange). URL: http://mailto:[email protected] (T. Hermange) https://doi.org/10.1016/j.tvjl.2018.12.002 1090-0233/© 2018 Elsevier Ltd. All rights reserved.
inflammation, characterised by occasional poor performance and coughing without increased respiratory effort at rest (Robinson, 2003; Couëtil et al., 2007, 2016). Severe EA has been defined as a marked neutrophilic lower airway inflammation and obstruction with regular to frequent coughing, exercise intolerance and increased respiratory efforts at rest (Couëtil et al., 2016; Pirie et al., 2016). Diagnostic methodologies for EIPH include blood visualisation and quantification during post-exercise tracheobronchoscopy (Birks et al., 2002; Hinchcliff et al., 2005), and quantification of erythrocytes (Epp et al., 2006) or haemosiderophages in the BALF (Doucet and Viel, 2002; Depecker et al., 2015). Performing BAL seems to be the most sensitive diagnostic tool (Sullivan et al., 2015). To date, BALF cytology is considered the reference standard method for diagnosis of non-infectious lower airway diseases (Couëtil et al., 2016). Several authors have previously investigated regional variations in terms of BALF cytology. One study including
T. Hermange et al. / The Veterinary Journal 244 (2019) 28–33
control (CTL) and sEA horses showed no differences between several regions of both lungs (McGorum et al., 1993). Other studies comparing right and left lungs showed statistical differences for metachromatic cells only (Sweeney et al., 1992; Jean et al., 2011). More recently, a study suggested an alternative methodology for BALF cytological investigation, a 5-field leukocyte differential count (Fernandez et al., 2013). However, the reliability of mast cell percentages was acceptable only on slides with high cell density. A prospective study including a large population of clinically healthy French Trotters found that BALF cytology from the right lung was not equivalent to cytology from the left (Depecker et al., 2014). According to this study, between 12.8% and 37.1% of horses would have been incorrectly been classified as ‘control’ instead of mEA when only one lung was sampled, depending on the cut-off used (‘restrictive’: 5% neutrophils, 1% eosinophils, 2% metachromatic cells; or ‘less restrictive’ one: 10% neutrophils, 5% eosinophils, 5% metachromatic cell). Similarly, between 6.7% to 14.3% of horses would have been incorrectly classified as ‘control’ rather than EIPH, when only one lung was sampled (Depecker et al., 2014). Therefore, both right and left lungs should be sampled for a precise cytological diagnosis of mEA and EIPH. However, ultimately, sampling both lungs with distinct cytological analyses substantially increase the costs for the owner. The hypothesis of this study was that one single cytology on a pooled BALF sample from both lungs would be a consistent tool for diagnosis of EA and EIPH. Our aims were to determine the diagnostic reliability of cytology from pooled BALF samples in comparison with individual BALF cytologies from each lung of the same horse. Materials and methods Horses Prospective power calculation (StatMate, GraphPad) for paired comparisons was performed with expected standard deviation of 23% and correlation among pairs of 0.97. A sample size of 50 horses has 80% power to detect a difference of 2.26% with a significance level (alpha) of 0.05. For inclusion in this study, horses had to be referred to the hospital between April 2014 and March 2017 for poor performance (as subjectively defined by the owner/rider) and/or respiratory signs (cough, dyspnoea, nasal discharge, noise during exercise, increased respiratory effort at rest), and examined by at least one of the authors (TH, CB, AC). The decision of performing BAL was made at the discretion of the attending clinician. Prior to any procedure and at least 24 h after the last training or racing (if any), each horse was subjected to a thorough clinical examination. If observed, the presence of cough, nasal discharge (either serous or mucous), nostril flaring at rest, or dyspnoea was recorded. Further examinations were performed at the clinician discretion in order to rule out any other condition, including i.e. haematology, biochemistry and tracheal wash bacterial and mycological culture, and thoracic radiographies. The study was approved by the regional Animal Ethic Committee (CERVO-2016-14-V). A consent form was also signed by the owner. Bronchoalveolar lavage Horses were sedated using detomidine (10 mg/kg IV, Sédomidine, Audevard) and butorphanol (10 mg/kg IV, Dolorex, Intervet). A nose twitch was used as necessary. The BAL was performed using a flexible 3.2 m long, 12.8 mm tip diameter videoendoscope (Optomed) without previous lidocaine instillation. The endoscope was inserted into one main bronchus until wedged. A total of 250 mL (two boluses of 125 mL each) of warmed (37 C) sterile isotonic saline solution was instilled through the endoscope biopsy channel, using 60-mL pre-filled syringes. After each bolus, the liquid was aspirated with the same syringes and the first 20 mL corresponding to the endoscope channel volume that did not reach the lung were disposed of. At the end of the procedure, the biopsy channel was rinsed with 40 mL of saline solution with the endoscope wedged; the endoscope was then pulled back to the carina and inserted into the controlateral lung and the same procedure was repeated. The volume of fluid collected from each lung was recorded. The contents of the syringes corresponding to the left and right lungs were put into two different metal bowls; an aliquot of 4 mL each BALF sample was kept in EDTA tubes. An equivalent volume (50 mL) from each bowl was pooled in another bowl. An aliquot of the pooled sample was then kept in EDTA tubes.
29
BALF analysis All three samples (left, right and pooled BALF) were kept refrigerated until handled by the laboratory within 24 h after collection. Two hundred microlitres of fluid were cytocentrifugated (80 g, 10 min) and stained with May-Grünwald-Giemsa (MGG). A differential cell count was performed on 300 cells and the number of each cell type was recorded as a percentage of total nucleated cells, excluding epithelial cells. The observer performing the cytologic examination was masked to the horse and the type of sample (left lung, right lung, or pooled sample). Case definition According to the revised American College of Veterinary Internal Medicine (ACVIM) consensus on IAD (Couëtil et al., 2016), two definitions were used to characterise normal cytology: ‘restrictive’ characterisation was 5% neutrophils, 1% eosinophils, 2% metachromatic cells; the ‘less restrictive’ characterisation was: 10% neutrophils, 5% eosinophils, 5% metachromatic cells. Horses without coughing, nasal discharge, or increased respiratory effort (dyspnoea or nasal flaring), and with normal BALF cytology from both lungs were classified as CTL. Horses with chronic (>3 weeks) coughing and/or history of poor performance, without increased respiratory effort at rest (nostril flaring or dyspnoea), and with abnormal BALF cytology (neutrophils, and/or eosinophils, and/or metachromatic cells) in at least one lung, were classified as mEA. Horses with chronic (>3 weeks) coughing, increased respiratory efforts at rest (nostril flaring or dyspnoea), and with abnormal BALF cytology (neutrophils only) in at least one lung were classified as sEA. Furthermore, recurrent and reversible airway obstruction for these horses was confirmed either by history or follow-up. Horses with cytology with a haemosiderophage/macrophage (H/M) ratio >17% (Depecker et al., 2015) were classified as EIPH. Statistical analysis Normality of continuous data distribution was evaluated using the Shapiro– Wilk W tests. When necessary, data were log 10 transformed to normalise distribution. Multiple comparisons for cytological parameters were performed using one-way ANOVA, with Dunnett’s post-hoc test. Pearson’s correlation test was performed to evaluate the association between pooled and respectively right and left lung BALF samples. Agreement between pooled sample and both left and right samples, for numerical (cell proportions) and categorical variables (CTL, mEA, sEA, EIPH), respectively, was measured with the intra-class correlation coefficient (ICC) and the Cohen’s kappa coefficient (k). Data were presented as mean standard deviation, unless stated otherwise. Values of P < 0.05 were considered statistically significant. Cut-off values for pooled samples were established based on a receiver operating characteristic (ROC) curve and associated area under the curve (AUC), with combined individual BALF samples used as cytological reference standard. The best cut-off was considered to be the data with the highest true-positive rate (sensitivity) and the lowest false-positive rate (1– specificity), which represents the point closest to the top left corner of the ROC curve.
Results Horses A total of 51 horses (females, n = 20; geldings n = 23; males n = 8), aged 3–22 years old (8.6 4.9 years) were included in the study. Breeds represented were as follows: Warmbloods (n = 17), Trotters (total n = 17; racing n = 9, non-racing, n = 8), Thoroughbreds (total n = 4; racing n = 2, non-racing, n = 2), ponies (n = 8), and miscellaneous horses (n = 5). BALF volume and cell proportions Mean volume proportions of fluid retrieved were 43.2 10.9% for the right lung and 45.7 11.5% for the left lung and were not significantly different between lungs (P = 0.36). Overall, there was no significant difference between pooled BALF and samples from right/left lungs for both neutrophil (P = 0.11) and metachromatic cell proportions (P = 0.07), and H/M ratio (P = 0.50). Eosinophil proportions in pooled BALF were significantly lower when compared with left lung samples (P = 0.02), but not with right lung samples (P = 0.47; Fig. 1).
30
T. Hermange et al. / The Veterinary Journal 244 (2019) 28–33
Fig. 1. Mean standard deviation for (a) neutrophil proportions, (b) metachromatic cells proportions, (c) eosinophil proportions, and (d) Haemosiderophages/Macrophages (H/M) ratio; corresponding mean difference with 95% confidence interval within each group for (a’) neutrophil proportions, (b’) metachromatic cells proportions, (c’) eosinophil proportions, and (d’) H/M ratio. Symbol * indicates P < 0.05.
Correlations and numerical agreements Good correlations between pooled BALF and each lung sample were found for neutrophil proportions and H/M ratio. Significant correlations were fair to moderate for metachromatic cell and eosinophil proportions (Table 1). Numerical agreements between
pooled BALF and right/left lung samples were good for neutrophil proportions and H/M ratio, and substantial for metachromatic cell proportions. For eosinophil proportions, numerical agreement was good between pooled BALF and left lung, and poor between pooled BALF and the right lung (Table 1; Supplementary Fig. S1 in the online version at DOI: 10.1016/j.tvjl.2018.12.002).
T. Hermange et al. / The Veterinary Journal 244 (2019) 28–33
31
Table 1 Correlations (Pearson coefficient, r), numerical agreements (intra-class coefficient, ICC), and corresponding 95% confidence intervals (95% CI), between bronchoalveolar lavage fluid cytology from pooled samples and respective right and left lung. Pooled sample vs. right lung Neutrophils
H/M ratio
Metachromatic cells
Eosinophils
r = 0.95 (95% CI: 0.90–0.98) ICC = 0.93 (95% CI: 0.89–0.96) r = 0.99 (95% CI: 0.99–0.99) ICC = 0.99 (95% CI: 0.98–0.99) r = 0.47 (95% CI: 0.14–0.71) ICC = 0.64 (95% CI: 0.47–0.77) r = 0.35 (95% CI: 0.00–0.63) ICC = 0.00 (95% CI: 0.26–0.25)
Pooled sample vs. left lung P < 0.01 P < 0.01 P < 0.01 P < 0.01 P = 0.01 P < 0.01 P = 0.03 P = 0.95
Left vs. right lung P < 0.01
r = 0.93 (95% CI: 0.86–0.97) ICC = 0.97 (95% CI: 0.94–0.98) r = 0.89 (95% CI: 0.79–0.95) ICC = 0.90 (95% CI: 0.84–0.94) r = 0.37 (95% CI: 0.02–0.65) ICC = 0,68 (95% CI: 0.51–0.80) r = 0.64 (95% CI: 0.36–0.81) ICC = 0.98 (95% CI: 0.97–0.99)
r = 0.89 (95% CI: 0.82–0.94) r = 0.88 (95% CI: 0.83–0.92) r = 0.87 (95% CI: 0.78–0.92) r = 0.85 (95% CI: 0.79–0.90) r = 0.51 (95% CI: 0.28–0.69) r = 0.47 (95% CI: 0.30–0.61) r = 0.014 (95% CI: 0.26–0.29) r = 0.01 (95% CI: 0.20–0.19)
P < 0.01 P < 0.01 P < 0.01 P = 0.01 P < 0.01 P < 0.01 P < 0.01
P < 0.01 P < 0.01 P < 0.01 P < 0.01 P < 0.01 P < 0.01 P = 0.90 P = 0.96
H/M, Haemosiderophages/Macrophage.
Cut-off values for pooled samples Based on the ‘restrictive’ definition, thresholds for pooled samples, as determined by ROC, were set at 3% neutrophils and 1% metachromatic cells. Based on the ‘less restrictive’ definition, the thresholds, as determined by ROC, were set at 9% neutrophils, 3% metachromatic cells. Due to the low number of samples with increased eosinophil proportions, thresholds for pooled sample could not be established. For EIPH, threshold for pooled samples was set at 9% H/M. The corresponding sensitivities and specificities are presented in Table 2 and graphical representations are in Supplementary Fig. S2 in the online version at DOI: 10.1016/j. tvjl.2018.12.002.
definition (Table 3). However, when considering BALF cytology from one lung only (and irrespective of the lung side), five and six horses with mEA were classified as controls when using respectively the ‘restrictive’ and ‘less restrictive’ definition, respectively. There was substantial agreement between pooled BALF sample and combined individual BALF samples (k = 0.74; 95% CI: 0.46– 1.00) for the final diagnosis of EIPH, based on H/M ratio. The same 5/51 (9.8%) horses were classified as EIPH when performing cytology on either both lungs or pooled sample. Three other horses were categorised as EIPH based on the pooled sample, but not on samples from the individual lungs. Table 3 Comparison of final diagnosis based on clinical signs and cytological data from either both lungs or pooled sample.
Final diagnosis and categorical agreement Based on BALF cytology from both lungs separately, classification as CTL, mEA, sEA and EIPH when using either ‘restrictive’ or ‘less restrictive’ definition is detailed in Table 3. Among horses classified as mEA, variable combinations of abnormal BALF cytological profiles were present (Table 4). Among horses classified as sEA, all but one exhibited neutrophil proportions >30% in both lungs (Fig. 2). One horse had 12% neutrophils in BALF from the left lung and 43% from the right lung. Good agreement between pooled BALF samples and combined individual BALF samples was found for the final diagnosis of CTL, mEA or sEA, when using both the ‘restrictive’ and the ‘less restrictive’ definition (k = 1.00 and k = 0.97; 95% CI: 0.91–1.00), respectively. All 51 horses were correctly classified when using the ‘restrictive’ definition and one horse classified as mEA was incorrectly classified as CTL when using the ‘less restrictive’
Groups
CTL mEA sEA EIPH total EIPH only Virtual totalb
‘Restrictive’ definitiona
‘Less restrictive’ definitionb
Left and right BALF samples
Pooled BALF sample
Left and right BALF samples
Pooled BALF sample
n=3 n = 36 n = 11 n=5 n=1 n = 56
n=3 n = 36 n = 11 n=8 n=1 n = 59
n = 11 n = 27 n = 11 n=5 n=2 n = 56
n = 12 n = 26 n = 11 n=8 n=2 n = 59
BALF, bronchoalveolar lavage fluid; CTL, control; mEA, mild equine asthma; sEA, severe equine asthma; EIPH, exercise-induced pulmonary haemorrhage. a This definition is used for the cytological discrimination between CTL and mEA horses only. b A total of 51 horses were included in this study. EIPH was detected in five horses by evaluating both lungs separately (virtual total of 56), and in eight horses when using pooled BALF samples (virtual total of 59).
Table 2 Reliability (sensitivity, specificity, corresponding area under the curve [AUC] and respective 95% confidence intervals [CI]) of cut-off values for pooled bronchoalveolar lavage fluid (BALF) samples, as defined according to the thresholds used for individual BALF samples. Neutrophils Individual threshold Pooled cut-off Sensitivity (95% CI) Specificity (95% CI) AUC (95% CI)
5% 3% 1.00 (0.91–1.00) 0.86 (0.42–1.00) 1.00 (0.98–1.00)
H/M, Haemosiderophages/Macrophages.
Metachromatic cells 10% 9% 0.93 (0.77–0.99) 1.00 (0.78–1.00) 0.98 (0.94–1.00)
2% 1% 0.96 (0.78–1.00) 0.64 (0.44–0.81) 0.88 (0.78–0.97)
H/M 5% 3% 1.00 (0.29–1.00) 0.73 (0.58–0.85) 0.94 (0.84–1.00)
17% 9% 1.00 (0.59–1.00) 0.98 (0.88–1.00) 1.00 (0.98–1.00)
32
T. Hermange et al. / The Veterinary Journal 244 (2019) 28–33
Table 4 Number of horses for each combination of cells classified as mild and moderate equine asthma (mEA), based on bronchoalveolar lavage fluid cytological evaluation from both lungs individually. Definition
‘Restrictive’ ‘Less restrictive’
Neutrophilic mEA
Non-neutrophilic mEA
Total
NEU
NEU + MET
NEU + EOS
NEU + MET + EOS
MET
EOS
MET + EOS
20 23
12 1
0 1
1 0
2 2
0 0
1 0
36 27
NEU, neutrophils; MET, metachromatic cells; EOS, eosinophils.
Fig. 2. Graphic representations of bronchoalveolar lavage fluid neutrophil proportions according to sample type and final diagnosis. CTL, control; mEA, mild to moderate equine asthma; sEA, severe equine asthma.
Discussion To our knowledge, this is the first report of pooled equine BALF samples from both lungs compared to each lung separately. Excellent correlation and numerical agreement were demonstrated for neutrophil proportions between pooled BALF samples and samples from both lungs. For metachromatic cells, correlation and numerical agreement were only moderate. This may partly be explained by their low proportions in BALF, and also by their low prevalence in the population under study (14/51 and 3/51 horses with increased metachromatic cell proportions when using the ‘restrictive’ and the ‘less restrictive’ definition, respectively). For the same reasons, correlation and numerical agreement for eosinophils was also only moderate (prevalence: 2/51 and 1/51 horses with ‘restrictive’ and ‘less restrictive’ definitions, respectively). Furthermore, among the 2/51 horses with BALF eosinophilia (>1%), one horse exhibited 15% eosinophils in the left lung and no eosinophils in the right lung. As 13% eosinophils were found in the pooled BALF sample, the discrepant findings in this horse likely contributed to the moderate correlation and numerical agreement between pooled and right BALF; this limits the clinical relevance of the eosinophil data. The current data confirm those of previous studies, in which good correlations and agreements were found between lungs for BALF neutrophil proportions but not for metachromatic cell and eosinophil proportions (Jean et al., 2011; Depecker et al., 2014). However, neutrophil proportions between the left and right lungs were not significantly different, in contrast with previously reported results (Depecker et al., 2014). This might be explained by comparisons between study populations; horses of any age were referred for respiratory investigation in our study, while prospective evaluation of clinically healthy young racehorses was performed in the previous study. For BALF inflammatory cells, the pooled BALF sample was representative of combined individual BALF samples, particularly for neutrophil proportions. Most cases of mEA are associated with increased neutrophil proportions in BALF (Fogarty and Buckley,
1991; Couëtil et al., 2001; Depecker et al., 2014; Richard et al., 2014), even if cytological profiles of mEA may also include increased proportions of other cell types (metachromatic cells and eosinophils) (Couëtil et al., 2007; Couëtil et al., 2016). However, sEA has been associated with increased proportions of neutrophils only (Pirie, 2014). When considering the final diagnosis of equine asthma, pooled BALF samples were reliable, since excellent categorical agreements were found for the discrimination between CTL, mEA and sEA. Importantly, no discordancy (0/51 horses) was found between methods for the final diagnosis of EA when using the ‘restrictive’ definition, and only 1/51 discordant diagnosis was obtained when using the ‘less restrictive’ definition (one horse with mEA incorrectly classified as CTL). In contrast, sampling one lung only resulted in 6/51 discordant diagnoses. The minimum sample size (50 horses) was determined based on BALF neutrophil proportions. The mEA phenotype (neutrophilic, metachromatic, eosinophilic, mixt) has not been further subdivided for investigation of categorical agreements. Among the 36 horses diagnosed with mEA and according to the ‘restrictive’ definition, 33/36 horses exhibited BALF neutrophilia, with or without increased proportions of either metachromatic cells or eosinophils. Similarly, increased neutrophil proportions were found in 25/27 horses with mEA when using the ‘less restrictive’ definition. Although metachromatic cells and eosinophils correlations and agreements were only moderate, their impact on the final diagnosis of equine asthma was limited. According to the revised ACVIM consensus on IAD (Couëtil et al., 2016), mild and moderate EA were not separated within the ‘mEA group’ in the current study. However, it has recently been suggested that both entities could be separated, based on the presence/absence of coughing in addition to poor performance (Bullone and Lavoie, 2017). In contrast, coughing was a specific but insensitive clinical sign associated with non-severe lower airway inflammation in previous studies (Wasko et al., 2011; Cardwell et al., 2014). Coughing may not be reliability reported by owners/ trainers, leading to some horses being incorrectly classified as ‘mild’ EA instead of ‘moderate’ EA. Three horses presented for poor performance and high neutrophil proportions in BALF (neutrophils >25%) were classified as mEA because of the absence of any respiratory signs. Apart from BALF cytology, this emphasises the importance of clinical investigations (i.e. respiratory effort at rest) to discriminate between mild/moderate and severe EA. However, no sEA horse in the current study exhibited increased respiratory effort at rest associated with low proportions of neutrophils (<25%) in the BALF. This paucigranulocytic phenotype has previously been mentioned by some authors (Bullone and Lavoie, 2017). The absence of discordant diagnoses for sEA between both methods might be explained by the relatively high neutrophil proportions (all samples from horses with sEA except for one had neutrophil proportions >30%). However, 1/11 (9%) horse had 12% neutrophils in the BALF from the left lung and 46% from the right lung. This discrepancy (paucigranulocytic pattern on one lung and neutrophilic on the other one), highlights the relevance of sampling both lungs for a thorough assessment.
T. Hermange et al. / The Veterinary Journal 244 (2019) 28–33
Although categorical agreement was also substantial for the final diagnosis of EIPH, a few discordant results were found between both methods. A total of 3/51 horses were classified as EIPH based on the pooled samples, but were classified as CTL based on individual lungs. However, the presence of false positive results was unlikely, based on previous or concomitant clinical observations. One horse was diagnosed with atrial fibrillation, which is often associated with EIPH (Reef et al., 1995). In another horse, EIPH had been diagnosed (endoscopically) a few months previously; the third horse exhibited mild epistaxis at the end of the exercise test. However, the pooled sample threshold used for the diagnosis of EIPH should be interpreted with caution, since the confidence interval for sensitivity was relatively large (95% CI: 0.59–1.00). This might be associated with the low prevalence of EIPH (5/51 horses) in the current population, which likely reflects the relatively low number of racing horses included in the study (11/51). Conclusions When using a standardised collection procedure, cytology from pooled BALF samples is representative of both individual lungs, according to both correlations and numerical agreements. Based on categorical agreement (final diagnosis), cytology of pooled BALF samples has been validated for equine asthma. Further studies using larger populations of racing horses are required to definitively validate the threshold value for EIPH cytological detection in pooled BALF samples. Conflict of interest statement None of the authors of this paper has a financial or personal relationship with other people or organizations that could inappropriately influence or bias the content of the paper. Acknowledgements Preliminary results were presented to the 9th ECEIM congress 2016, Helsinki, Finland (oral communication) and to the WEAS symposium 2017 (poster), Copenhagen, Denmark. Financial support was provided by LABÉO, CISCO-Oniris and AVEF (veterinary association of French practitioner). References Birks, E.K., Shuler, K.M., Soma, L.R., Martin, B.B., Marconato, L., Del Piero, F., Teleis, D. C., Schar, D., Hessinger, A.E., Uboh, C.E., 2002. EIPH: postrace endoscopic evaluation of Standardbreds and Thoroughbreds. Equine Vet. J. Suppl. 375–378. Bullone, M., Lavoie, J.-P., 2017. Science-in-brief: equine asthma diagnosis: beyond bronchoalveolar lavage cytology. Equine Vet. J. 49, 263–265. Cardwell, J.M., Smith, K.C., Wood, J.L.N., Newton, J.R., 2014. Infectious risk factors and clinical indicators for tracheal mucus in British National Hunt racehorses. Equine Vet. J. 46, 150–155.
33
Couëtil, L.L., Cardwell, J.M., Gerber, V., Lavoie, J.-P., Léguillette, R., Richard, E.A., 2016. Inflammatory airway disease of horses-revised consensus statement. J. Vet. Intern. Med. 30, 503–515. Couëtil, L.L., Hoffman, A.M., Hodgson, J., Buechner-Maxwell, V., Viel, L., Wood, J.L.N., Lavoie, J.-P., 2007. Inflammatory airway disease of horses. J. Vet. Intern. Med. 21, 356–361. Couëtil, L.L., Rosenthal, F.S., DeNicola, D.B., Chilcoat, C.D., 2001. Clinical signs, evaluation of bronchoalveolar lavage fluid, and assessment of pulmonary function in horses with inflammatory respiratory disease. Am. J. Vet. Res. 62, 538–546. Depecker, M., Couroucé-Malblanc, A., Leleu, C., Genneviève, V., Pitel, P.-H., Richard, E.A., 2015. Comparison of two cytological methods for detecting pulmonary haemorrhage in horses. Vet. Rec. 177, 305.2–305. Depecker, M., Richard, E.A., Pitel, P.-H., Fortier, G., Leleu, C., Couroucé-Malblanc, A., 2014. Bronchoalveolar lavage fluid in Standardbred racehorses: influence of unilateral/bilateral profiles and cut-off values on lower airway disease diagnosis. Vet. J. 199, 150–156. Doucet, M.Y., Viel, L., 2002. Alveolar macrophage graded hemosiderin score from bronchoalveolar lavage in horses with exercise-induced pulmonary hemorrhage and controls. J. Vet. Intern. Med. 16, 281–286. Epp, T.S., McDonough, P., Padilla, D.J., Gentile, J.M., Edwards, K.L., Erickson, H.H., Poole, D.C., 2006. Exercise-induced pulmonary haemorrhage during submaximal exercise. Equine Vet. J. Suppl. 502–507. Fernandez, N.J., Hecker, K.G., Gilroy, C.V., Warren, A.L., Léguillette, R., 2013. Reliability of 400-cell and 5-field leukocyte differential counts for equine bronchoalveolar lavage fluid. Vet. Clin. Pathol. 42, 92–98. Fogarty, U., Buckley, T., 1991. Bronchoalveolar lavage findings in horses with exercise intolerance. Equine Vet. J. 23, 434–437. Hinchcliff, K.W., Couetil, L.L., Knight, P.K., Morley, P.S., Robinson, N.E., Sweeney, C.R., van Erck, E., 2015. Exercise induced pulmonary hemorrhage in horses: American College of Veterinary Internal Medicine Consensus Statement. J. Vet. Intern. Med. 29, 743–758. Hinchcliff, K.W., Jackson, M.A., Brown, J.A., Dredge, A.F., O’Callaghan, P.A., McCaffrey, J.P., Morley, P.S., Slocombe, R.E., Clarke, A.F., 2005. Tracheobronchoscopic assessment of exercise-induced pulmonary hemorrhage in horses. Am. J. Vet. Res. 66, 596–598. Jean, D., Vrins, A., Beauchamp, G., Lavoie, J.-P., 2011. Evaluation of variations in bronchoalveolar lavage fluid in horses with recurrent airway obstruction. Am. J. Vet. Res. 72, 838–842. McGorum, B.C., Dixon, P.M., Halliwell, R.E.W., Irving, P., 1993. Comparison of cellular and molecular components of bronchoalveolar lavage fluid harvested from different segments of the equine lung. Res. Vet. Sci. 55, 57–59. Pirie, R.S., 2014. Recurrent airway obstruction: a review. Equine Vet. J. 46, 276–288. Pirie, R.S., Couëtil, L.L., Robinson, N.E., Lavoie, J.-P., 2016. Equine asthma: an appropriate, translational and comprehendible terminology? Equine Vet. J. 48, 403–405. Reef, V.B., Reimer, J.M., Spencer, P.A., 1995. Treatment of atrial fibrillation in horses: new perspectives. J. Vet. Intern. Med. 9, 57–67. Richard, E.A., Depecker, M., Defontis, M., Leleu, C., Fortier, G., Pitel, P.-H., CouroucéMalblanc, A., 2014. Cytokine concentrations in bronchoalveolar lavage fluid from horses with neutrophilic inflammatory airway disease. J. Vet. Intern. Med. 28, 1838–1844. Robinson, N.E., 2003. Inflammatory airway disease: defining the syndrome. Conclusions of the Havemeyer Workshop. Equine Vet. Educ. 15, 61–63. Sullivan, S.L., Whittem, T., Morley, P.S., Hinchcliff, K.W., 2015. A systematic review and meta-analysis of the efficacy of furosemide for exercise-induced pulmonary haemorrhage in Thoroughbred and Standardbred racehorses: furosemide for exercise-induced pulmonary haemorrhage. Equine Vet. J. 47, 341–349. Sweeney, C.R., Rossier, Y., Ziemer, E.L., Lindborg, S., 1992. Effects of lung site and fluid volume on results of bronchoalveolar lavage fluid analysis in horses. Am. J. Vet. Res. 53, 1376–1379. Wasko, A.J., Barkema, H.W., Nicol, J., Fernandez, N., Logie, N., Léguillette, R., 2011. Evaluation of a risk-screening questionnaire to detect equine lung inflammation: results of a large field study. Equine Vet. J. 43, 145–152.