Bronchoalveolar lavage hemosiderosis in lightly active or sedentary horses

Abstract

Background

Hemosiderophages in bronchoalveolar lavage fluid (BALF) are commonly ascribed to exercise‐induced pulmonary hemorrhage (EIPH). Little information exists regarding the presence of these cells in horses that perform light or no work and that are referred for respiratory problems.

Objectives

Evaluate the presence of hemosiderophages in BALF of horses suspected of respiratory disease without history of or risk factors for EIPH and determine predictors of hemosiderophages in BALF in this population.

Methods

Observational retrospective cross‐sectional study using STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines. Bronchoalveolar lavage fluid cytology reports of 353 horses evaluated for respiratory disease between 2012 and 2022 at the Cummings School for Veterinary Medicine were reviewed retrospectively. Horses with a history or likelihood of having performed past strenuous exercise were removed, and the remaining 91 horses were divided into hemosiderin‐positive (HSD‐POS) and hemosiderin‐negative groups based on Perls' Prussian blue staining. Potential predictors for the presence of hemosiderophages in BALF (history, clinical evaluation, baseline lung function, airway reactivity, BALF cytology, and hemosiderin score) were compared between the 2 groups, using univariate and multivariate analyses.

Results

Horses with a diagnosis of severe equine asthma (sEA; odds ratio, 11.1; 95% confidence interval, 3.2‐38.5; P < .001) were significantly more likely to be HSD‐POS than horses with mild‐to‐moderate equine asthma.

Conclusions and Clinical Importance

Hemosiderophages were found in the BALF cytology in a subset of horses that perform light or no work and presented for respiratory signs; these cells were found more frequently in horses with sEA. The link between hemosiderophages and sEA highlights previously unstudied pathology associated with this common disease.

Abbreviations

BAL

bronchoalveolar lavage

BALF

bronchoalveolar lavage fluid

CI

confidence interval

COPD

chronic obstructive pulmonary disease

EA

equine asthma

EHV‐5

equine herpesvirus‐5

EIPH

exercise‐induced pulmonary hemorrhage

HSD‐NEG

hemosiderin‐negative

HSD‐POS

hemosiderin‐positive

HSD

hemosiderin

mEA

mild‐to‐moderate equine asthma

OR

odds ratio

PPB

Perls' Prussian blue

sEA

severe equine asthma

1. INTRODUCTION

Hemosiderophages (macrophages containing iron complexes known as hemosiderin) are an occasional finding in respiratory secretions of many species. Hemosiderin is generated as the result of breakdown of red blood cells by alveolar macrophages in the lungs and serves as an iron storage complex. ¹ Hemosiderin is identified in respiratory secretions, most commonly from bronchoalveolar lavage fluid (BALF) samples, based on the characteristic appearance on modified Wright‐Giemsa staining and by detection of iron, most often using the Perls' Prussian blue (PPB) reaction (Figure 2). ¹ When alveolar hemorrhage occurs, iron is partially stored as the nonheme iron complex hemosiderin (the proteolytic breakdown product of damaged ferritin in a phagolysosome). The PPB reaction detects ferric (Fe³⁺) iron in hemosiderin by generation of a blue pigment, insoluble ferric ferrocyanide, ¹ but does not detect the ferrous (Fe²⁺) iron in hemoglobin or myoglobin.

FIGURE 2.

Perls Prussian blue staining of BAL cytology. (A) No hemosiderophages present, nuclei of alveolar macrophages and neutrophils counterstained with Nuclear Fast Red. (B) Score = 1; occasional, or <20% of alveolar macrophages positive on PPB. (C) Score = 2; moderate, or 20% to 50% alveolar macrophages positive on PPB. (D) Score = 3; majority, or >50% alveolar macrophages positive on PPB. BAL, bronchoalveolar lavage; PPB, Perls' Prussian blue

Horses that exercise intensively in a variety of disciplines are reported to have hemosiderophages in their BALF, ² , ³ , ⁴ but in other species they are described in individuals without a history of strenuous exercise in association with respiratory or cardiopulmonary diseases of inflammatory, infectious, or neoplastic origin. ⁵ , ⁶ , ⁷ , ⁸ , ⁹ Indeed, hemosiderophages are not always diagnostic of alveolar hemorrhage, and can instead indicate the presence of ferruginous bodies, in which iron‐containing protein may coat inhaled particulates. ¹ Thus, it is reasonable to suppose that sedentary horses, especially those with pulmonary disease, might have alveolar hemosiderophages detectable in BALF.

Several quantitative scoring systems have been proposed to evaluate hemosiderophage numbers and extent of hemosiderin within individual alveolar macrophages, but no accepted gold standard has been proposed, and reported methods of nonautomated counting suffered from inter‐rater variability. ¹⁰ , ¹¹ Recently, automated image analysis using deep learning ¹² , ¹³ has been developed to avoid this problem, but this software is not readily available. Our study used a binary approach to assess presence of hemosiderophages in the BALF of horses without history or suspicion of exercise‐induced pulmonary hemorrhage (EIPH). The goal of our retrospective cross‐sectional study was to investigate the presence of hemosiderophages in the BALF cytology of lightly active or sedentary horses presented to a referral hospital for suspected respiratory disease, and to identify predictors for the presence of these cells in a population not suspected to have EIPH.

2. METHODS

Case records from 353 horses presented to Cummings School of Veterinary Medicine over an 11‐year period (2012 through 2022) for which history, clinical examination, and BALF cytology were available, and were used for a retrospective cross‐sectional study. All records were examined to determine the equine current and historical activity level, and horses that were known to have undergone strenuous exercise in the past or the signalment of which increased the likelihood of their having undergone strenuous activity were excluded (Figure 1). Thus, the study excluded any horse with a history of racing on flat or over jumps, Western performance horses, polo ponies, event horses, endurance horses, jumpers, and upper level dressage horses, and all subsets of horses in which EIPH has been documented. In addition, even if their exercise history was unknown, Thoroughbreds and Standardbreds also were excluded, ¹⁴ with 106 horses considered sedentary. The BALF cytology reports of these sedentary horses were examined for description of PPB staining for confirmation of hemosiderophages; if available, archived slides were stained with PPB. Fifteen horses were removed from the study because of incomplete information or lack of archived slides, leaving 91 horses in the study (Figure 1).

FIGURE 1.

Study design.

Bronchoalveolar lavage was performed using either a commercial cuffed BAL tube (Bivona Inc, Gary, Indiana) or with a 3‐m bronchoscope, using 2 aliquots of 250 mL of warmed physiologic saline. The 2 samples subsequently were pooled, and slides were prepared by cytocentrifugation. Slides were stained with modified Wright stain and toluidine blue, the latter for enumeration of mast cells. The PPB  method was performed either at the time of original presentation or on archived slides (Figure 2). Cells (n = 500) were classified as the percentage of total cells that were macrophages, lymphocytes, neutrophils, eosinophils, and mast cells, and examined for the presence of hemosiderin or other abnormalities (×1000 magnification). Hemosiderin was identified as PPB‐positive inclusions within alveolar macrophages and was either reported categorically (none, occasional, moderate, or majority) or as a percentage of alveolar macrophages containing hemosiderin. All reports subsequently were converted to categorical scoring on a scale of 0 to 3, with 0 (no hemosiderin was seen within alveolar macrophages), 1 (mild, <20% of alveolar macrophages contained hemosiderin), 2 (moderate, 20%‐50% of alveolar macrophages contained hemosiderin), or 3 (marked, >50% of alveolar macrophages contained hemosiderin; Table 1).

TABLE 1.

Hemosiderin scoring system.

Percentage of macrophages with hemosiderin	Categorical assessment	Final score used
0%	None seen	0
<20%	Occasional	1
20%‐50%	Moderate	2
>50%	Majority	3

Mono‐sinusoidal, multifrequency (1‐3 Hz) forced oscillatory mechanics were performed on the majority of cases to measure total respiratory system resistance (R _RS) as described previously. ¹⁵ Histamine bronchoprovocation was completed as previously described in horses with baseline total R _RS <0.80 cm H₂O/L/s (within 1 SD of the expected reference range; n = 59). ¹⁶ After testing, a dose‐response curve was generated to determine the histamine dose required to reach a 75% increase in R _RS (PC75 R _RS). Airway hyperreactivity was classified as a PC75 R _RS ≤6.0 mg/mL histamine.

Cases were divided into 3 main categories: mild‐to‐moderate equine asthma (mEA), sEA (heaves), and other cardiopulmonary disease (non‐EA). Horses were placed in the sEA category if they showed the following combined clinical abnormalities: current or historical labored breathing at rest, baseline respiratory system resistance >1.0 cm H₂O/L/s, or chronic cough or nasal discharge, in conjunction with evidence of disease chronicity based on history, with at least 2 episodes in the past and improvement when treated with corticosteroids, bronchodilators, environmental remediation, or a combination of these treatments. Horses therefore were classified as having sEA based on history and clinical assessment or lung function testing, rather than solely by BALF neutrophil proportions. ¹⁷ Horses were placed in the mEA category based on the criteria outlined in the most recent American College of Veterinary Internal Medicine consensus statement ¹⁸ with a history of poor performance or exercise intolerance, occasional coughing >3 weeks in duration, BALF cytology characterized by mild increases in neutrophils (>5%), eosinophils (>1%), metachromatic cells (>2%) or some combination of these, lack of visible respiratory effort at rest, and resting respiratory system resistance <0.6 cm H₂O/L/s. Horses were placed in the non‐EA category if the criteria for mEA or sEA did not apply, and further diagnostic testing, including echocardiography, ECG, thoracic radiographs, upper airway endoscopy, and blood chemistry or CBC were pursued as clinically indicated to identify a disease process other than EA.

2.1. Statistical analysis

All data were found to be non‐normally distributed based on Kolmogorov‐Smirnov testing and are therefore presented as median and interquartile range (continuous variables) or frequencies and percentages (categorical data). Univariate statistical analyses employed Mann‐Whitney U and Pearson chi‐square analyses to compare groups (accepted significance level: P < .05). In addition, stepwise forward logistic regression was performed using all continuous and categorical variables that were obtained in at least 70% of animals at the time of patient admission as predictors, with presence of hemosiderin as the outcome. These were used to derive odds ratio (OR) for analysis of outcome, using standard criteria of 0.05 for entry and 0.2 for removal. All data were analyzed using commercially available statistical software (IBM SPSS Statistics 28.0, Chicago, Illinois).

3. RESULTS

In total, 26 mares and 65 geldings were included in the final analysis. The most common breeds were Quarter Horses (32%), Warmbloods (24%), and Morgans (12%). In terms of performance, 19% were backyard companions, 59% were considered pleasure or light trail horses, and 22% performed at a low level (dressage, pleasure or flat show, hill‐topping, or light driving). A final diagnosis of mEA was reached in 39 horses (42.8%) and sEA (active disease or in remission) in 42 horses (46.2%). A diagnosis of “other cardiopulmonary disease” was made in 10 horses (10.9%). Of the latter group, 1 had grade 3 left laryngeal hemiplegia, 1 had immune‐mediated thrombocytopenia, 1 had atrial fibrillation, 1 had birefringent glass‐like particulates found in alveolar macrophages, 1 had pneumonia, and 2 had interstitial pulmonary fibrosis not associated with equine herpesvirus‐5 (EHV‐5), 2 had equine multinodular pulmonary fibrosis positive for EHV‐5 using PCR on BALF, and 1 had epiglottic entrapment. Only 35 horses had radiographs performed; 28 of these consisted of a diffuse, mild broncho‐interstitial pattern, and 1 had an alveolar pattern consistent with pneumonia. Six horses had a description of diffuse marked or severe interstitial pattern, with a radiographic diagnosis of pulmonary fibrosis, 2 of which had a positive PCR test on BALF for EHV‐5, as noted above.

Hemosiderophages were observed in the BALF of 50 (55%) horses. Of these horses, 31 (62%) had a score of 1 (<20% alveolar macrophages), 10 (11%) had a score of 2 (20%‐50% alveolar macrophages), and 9 (10%) had a score of 3 (>50% of alveolar macrophages). In horses with mEA, 71.8% had no hemosiderophages observed on PPB, and 12.8% had grade 1 (mild), whereas equal frequencies (7.7%) of grade 2 (moderate) and grade 3 (marked) were noted. In horses with sEA, 19% had no hemosiderophages observed, with 59.5% having grade 1, 14.3% grade 2, and 7.1% grade 3.

Results of the univariate analyses to determine a potential association between hemosiderophages and examination findings (BALF cytology, history and clinical examination) are presented in Table 2. Horses with a diagnosis of sEA were significantly more likely to have hemosiderophages in the BALF than those with mEA (P < .001). In addition, horses in the hemosiderophage‐positive group had higher respiratory rates and were more likely to have a presenting complaint of tachypnea (Table 2). Although geldings were over‐represented compared to mares in the study population, no statistical association of sex with BALF hemosiderophages was found. (P = .55).

TABLE 2.

Patient characteristics (median and interquartile range or frequency and percentage).

	HSD‐POS	HSD‐NEG	Significance (P value)
	Median (interquartile range)	Median (interquartile range)
Age (years)	17.0 (15.0‐20.0)	15.0 (11.5‐18.0)	.135
Presenting respiratory rate (bmp)	24 (17‐24)	18 (8‐30)	.004*
Baseline R RS (cm H2O/L/s)	0.52 (0.40‐0.68)	0.54 (0.29‐1.80)	.826
PC75 R RS (H2O/L/s)	4.0 (2.0‐8.0)	6.0 (4.0‐14.0)	.382
BAL neutrophils (%)	17.5 (5.5‐45.0)	6.0 (4.0‐24.0)	.102
BAL mast cells (%)	1.5 (1.0‐3.9)	2.5 (1.1‐4.5)	.276

	HSD‐POS	HSD‐NEG	Significance (P value)
	Frequency (percentage)	Frequency (percentage)
Diagnosis—severe equine asthma	34 (68.0)	8 (19.5)	<.001* (hemosiderophages significantly more frequent)
Complaint of cough	39 (78.0)	36 (87.8)	.205
Complaint of exercise intolerance	44 (88.0)	34 (82.9)	.348
Complaint of nasal discharge	7 (14.0)	8 (19.5)	.382
Complaint of tachypnea (historical)	21 (42.0)	8 (19.5)	.029*

^*Significant difference on univariate analysis (P < .05).

All examined clinical, historical and BAL variables were included in a forward stepwise logistic regression model, based on standard criteria of entry and removal. Diagnosis of “other cardiopulmonary disease” and PC75 R _RS applied to fewer than 70% of horses, and thus these variables were not incorporated into logistic regression analysis. Only diagnosis of sEA, with an OR of 11.1 (95% CI, 3.2‐38.5; P < .001) retained significance as an independent predictor of hemosiderophages on BALF cytology. Using Nalgekerke R, 32% of the variability in the dependent variable (presence of hemosiderophages on BALF cytology) could be explained by a diagnosis of sEA.

4. DISCUSSION

We retrospectively evaluated the presence of hemosiderophages in BALF in a population of sedentary or lightly active horses referred for suspected respiratory disease. We identified the presence of hemosiderophages in the BAL cytology of a population of horses without history of intense exercise. In this population of horses at the same low level of sport activity, those with sEA were significantly more likely to have pulmonary hemosiderosis compared to those with mEA. Although clinical complaint of tachypnea and high respiratory rate on presentation were associated with hemosiderophages in BALF cytology in the univariate analyses, these variables did not retain significance in the logistic regression model, indicating an interaction between these clinical findings and the diagnosis of sEA.

The presence of hemosiderophages in respiratory secretions in the horse is commonly considered sensitive and specific for EIPH, with hemosiderophages estimated to be found in BALF cytology of >90% of racing Thoroughbreds performing short intense work ¹⁹ and in 45% of elite endurance horses performing at a slower pace for longer times. ²⁰ Pulmonary hemosiderosis associated with alveolar hemorrhage also has been associated with pulmonary hypertension associated with chronic lung inflammation, specifically pulmonary veno‐occlusive disease, generalized vasculitis and coagulation disorders in humans. ²¹ , ²² , ²³ The presence of hemosiderophages in BALF cytology has been shown to be a sensitive and specific marker for left ventricular dysfunction in humans, ²⁴ , ²⁵ and high numbers of alveolar hemosiderophages have been found in human patients with congestive heart failure. ²⁶ Studies in small animals, ⁸ mice, ²⁷ and horses ²⁸ have shown that damage to the pulmonary microcirculation and blood gas permeability related to pulmonary hypertension have been closely related to alveolar bleeding and consequent hemosiderosis.

Apart from its close link to EIPH and alveolar hemorrhage in horses, hemosiderophages in BALF cytology have been described in relation to lower airway inflammation and free radical accumulation in other species. ¹ Although hemosiderophages commonly are attributed to alveolar hemorrhage, the PPB reaction identifies any ferric (Fe³⁺) iron and can thus more broadly be seen as evidence of any failure of iron homeostasis in the lung. ¹ Hemosiderin functions to limit oxidative damage caused by iron‐induced free radicals in the lung, ²⁹ and has been found to be closely associated with pathology involving pulmonary oxidative damage in steel workers, smokers, individuals with exposure to lung pollution, ⁵ chronic obstructive pulmonary disease (COPD), those with lung cancer, ⁶ , ³⁰ and cystic fibrosis. ³¹ Individuals with COPD have increased concentrations of the inflammatory cytokine interleukin‐6 that initiates the process of iron sequestration in pulmonary macrophages and consequently weakens macrophage immune function. ³² , ³³ Case studies in humans have found that conditions such as idiopathic pulmonary fibrosis and bronchiolitis obliterans, in which the lower airways develop severe inflammation secondary to previous injury, can result in pulmonary hemosiderosis. ³⁴ , ³⁵ Moreover, increased percentages of hemosiderophages have been associated with higher risk of infection. ³⁶

In our study, sEA was found to be a risk factor for alveolar hemosiderosis. Ours is the first study to establish a link between a diagnosis of sEA and pulmonary hemorrhage. Horses with sEA have similar pulmonary inflammation and remodeling ³⁷ that is described in other species with pulmonary hemosiderosis. Moreover, sEA is characterized by increased intrapleural pressures and in some cases, pulmonary hypertension ³⁸ that may result in alveolar hemorrhage. Thus, it is not surprising that hemosiderophages are found in the BALF of horses afflicted by sEA. Interestingly, no cytologic markers of lower airway inflammation (e.g., neutrophils, mast cells) were associated with the presence of hemosiderophages in the BALF (Table 2). This finding may reflect the relatively long lifespan of hemosiderophages, because the time required to clear the equine lung of these cells has been reported to be from 2 weeks to >1 year, ³⁹ , ⁴⁰ during which time severe inflammation in the lung may have subsided because of environmental management or pharmacological treatment.

Our study had some limitations. The horses were drawn from a larger population of horses presented to our hospital with a presumptive diagnosis of lower respiratory disease (most commonly EA) and therefore healthy horses were not included. Indeed, almost all healthy horses seen by our hospital are functioning athletes, making it difficult to find a population of healthy sedentary horses. Sedentary horses could be evaluated in the future by studies of BALF cytology in nonhospital populations. In addition, Standardbreds and Thoroughbreds were excluded to avoid the influence of alveolar hemorrhage episodes arising from irreversible fibrotic remodeling because of chronic EIPH. ⁴⁰ The primary information examined in our study included owner history, final diagnosis, signalment, lung function testing, radiographs, and quantitative information from the BALF cytology. A more detailed examination of the morphology of hemosiderin contained within alveolar macrophages would be informative with respect to the timeframe of the hemorrhagic event. In addition, pulmonary hypertension has been described in horses with sEA, ⁴¹ and recently sEA was found to be associated with remodeling of the pulmonary arteries with an increase in smooth muscle mass that may contribute to pulmonary hypertension. ⁴² Considering these findings, it would have been useful to assess pulmonary arterial pressure ultrasonographically ⁴³ or to determine pulmonary arterial wedge pressure using cardiac catheterization, given that no previous correlation has been found between pulmonary vascular pressures (as estimated from pulmonary arterial wedge pressure) in bleeding versus nonbleeding Thoroughbreds. ⁴⁴ In addition, the possibility that hemosiderosis may be a marker of infection, as suggested in human patients with COPD, ³³ cannot be ruled out and complete bacteriology was not performed in our cases. It is further possible that pulmonary hemosiderosis stems from systemic iron excess, a rare condition in horses, ¹ , ⁴⁵ which was not considered clinically in any of our cases. Because the PPB reaction results in intracellular ferric (Fe³⁺) iron being identified with blue color without specificity as to its source, the possibility of ferruginous bodies resulting from iron‐containing protein coating inhaled inorganic particulates cannot be ruled out. ¹

We identified not only the presence of pulmonary hemosiderosis in a population of nonathletic horses presented to our hospital for suspected respiratory disorders, but also showed that horses with sEA are markedly more likely than those with mEA to have hemosiderophages in the BALF cytology. Additional studies should be directed at determining whether hemosiderophages in this population of sedentary horses with airway inflammation are a consequence of low‐grade alveolar hemorrhage, iron dyshomeostasis, or a combination of these.

CONFLICT OF INTEREST DECLARATION

Authors declare no conflict of interest.

OFF‐LABEL ANTIMICROBIAL DECLARATION

Authors declare no off‐label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

Authors declare no IACUC or other approval was needed.

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.

Mahalingam‐Dhingra A, Bedenice D, Mazan MR. Bronchoalveolar lavage hemosiderosis in lightly active or sedentary horses. J Vet Intern Med. 2023;37(3):1243‐1249. doi: 10.1111/jvim.16692