SEARCH

SEARCH BY CITATION

Keywords:

  • Heaves;
  • Horse;
  • Genetics;
  • Lung;
  • Major gene;
  • Segregation analysis

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgments
  9. References

Background: Mode of inheritance of equine recurrent airway obstruction (RAO) is unknown.

Hypothesis: Major genes are responsible for RAO.

Animals: Direct offspring of 2 RAO-affected Warmblood stallions (n = 197; n = 163) and a representative sample of Swiss Warmbloods (n = 401).

Methods: One environmental and 4 genetic models (general, mixed inheritance, major gene, and polygene) were tested for Horse Owner Assessed Respiratory Signs Index (1–4, unaffected to severely affected) by segregation analyses of the 2 half-sib sire families, both combined and separately, using prevalences estimated in a representative sample.

Results: In all data sets the mixed inheritance model was most likely to explain the pattern of inheritance. In all 3 datasets the mixed inheritance model did not differ significantly from the general model (P= .62, P= 1.00, and P= .27) but was always better than the major gene model (P < .01) and the polygene model (P < .01). The frequency of the deleterious allele differed considerably between the 2 sire families (P= .23 and P= .06). In both sire families the displacement was large (t= 17.52 and t= 12.24) and the heritability extremely large (h2= 1).

Conclusions and Clinical Relevance: Segregation analyses clearly reveal the presence of a major gene playing a role in RAO. In 1 family, the mode of inheritance was autosomal dominant, whereas in the other family it was autosomal recessive. Although the expression of RAO is influenced by exposure to hay, these findings suggest a strong, complex genetic background for RAO.

Equine chronic lower airway disease in its severe form is known as recurrent airway obstruction (RAO) or heaves and is associated with both genetic and environmental factors.1,2 Affected horses have increased breathing effort, cholinergic bronchospasm, coughing, airway hyperreactivity, and accumulation of neutrophil and mucus in the airways.3,4

There is a genetic predisposition for RAO, demonstrated trough genetic epidemiological investigations in full- and half-sibling groups. There is a familial predisposition for equine chronic lower airway disease in Warmblood and Lipizzan horses in which the risk for developing moderate to severe equine chronic lower airway disease is significantly increased in offspring with 1 or 2 affected parents.1,5–7 The inheritance mode is not simple Mendelian pattern leading to the proposition that RAO is a polygenetic disorder.8

We have previously used standardized questionnaire to investigate direct offspring of 2 RAO-affected Warmblood stallions (sire 1 and 2) based on a Horse Owner Assessed Respiratory Signs Index (HOARSI 1–4, from healthy to severe signs).2 HOARSI is a composite score based on owner-observed coughing frequency, nasal discharge, breathing effort, and performance. The risk of sire 1 and 2 offspring developing RAO (HOARSI 3/4) was increased 4.1- and 5.5-fold, respectively, when compared with an unrelated control group and a group of maternal half-siblings.

There was an association and linkage of microsatellite markers in the ECA13q13 region with HOARSI in the family of 1 sire but not in the family of the 2nd sire.9 These results suggested a genetic background with locus heterogeneity for equine RAO. An alternative explanation for the differing results in the 2 families was that recombination had broken down the association between the paternal haplotype and disease locus.

Based on these findings, we hypothesized that 1 or more major genes are responsible for RAO. The aim of this study was to elucidate the mode of inheritance of the trait HOARSI 3 and 4 (combined grades) as a surrogate marker for susceptibility to RAO. Data from the 2 Warmblood families investigated in previous studies2,9 were therefore analyzed by segregation analysis testing environmental and genetic models. Also allele frequencies and penetrances, transmission probabilities by genotype, dominance effects, displacement for the major gene, and heritability for the polygenic effect were estimated which further define the inheritance characteristics and are useful for subsequent parametric linkage analysis.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgments
  9. References

HOARSI and Severity Classes

HOARSI 1–4 (healthy, mild, moderate, and severe clinical signs) has previously been described in detail.2 Briefly, horse owners were contacted by phone and only horses with clinical signs that had persisted for at least 2 months were included in the study. All individuals were 5 years or older with at least a 12 months history of hay feeding. A standardized questionnaire was used to gather information on the horses' history of chronic coughing, respiratory distress, increased breathing effort after exercise, and nasal discharge, all graded into categories of frequency and severity of clinical signs. This information was combined into a HOARSI classification. The classification refers to the period when the horses were exhibiting the most severe clinical signs. While HOARSI 1 comprises unaffected individuals (severity class 1), RAO in this study is represented in the severity class 3, which is comprised of HOARSI 3 and 4 individuals.2 Validation on 33 offspring of sire 1 and 36 offspring of sire 2 by comprehensive examination showed that HOARSI 3 and 4 individuals during exacerbation of disease are fully consistent with the RAO-phenotype (Gerber and Laumen, unpublished data). The milder signs of HOARSI 2 are consistent with inflammatory airway disease (IAD).10 Importantly, however, in our study these milder clinical signs were reported in older “occasional coughers” and therefore cannot be compared directly with IAD in racehorses. Earlier rechecks of HOARSI10 and those performed for the present study on over 200 questionnaires revealed a high consistency of categorization with a rate of misclassification of <1%.

Sire Families

The 2 high-prevalence sire families have also been described in detail elsewhere.2 Briefly, 2 Warmblood sires were selected, which showed obvious clinical signs of respiratory distress (nostril flare, increased abdominal lift or increased respiratory rate) and airway obstruction when stabled in stalls with straw bedding and fed hay, and remission of these signs when stabled in a barn complex especially adapted to the requirements of horses with RAO (bedding of dust-free shavings; haylage feeding). Sire family 1 comprised 197 offspring of which 149 were half-sibs, 36 formed 18 pairs of full sibs, and 12 formed 4 trios of full sibs. Of the 197 horses in sire family 1, 89 were considered unaffected and assigned to the severity class 1 (HOARSI 1; 45%), 44 were classified as mildly affected (severity class 2; HOARSI 2; 22%). HOARSI 3 (39 individuals), and HOARSI 4 (25 individuals) were summarized as RAO affected (33%), as described above, and assigned to the severity class 3. Sire family 2 comprised 163 offspring of which 137 were half-sibs, and 26 formed 13 pairs of full sibs. The 2 families were connected by 11 pairs and 5 trios of maternal half-sibs. Of the 163 horses in sire family 2, 64 were considered unaffected and assigned to the severity class 1 (HOARSI 1; 39%), 52 were classified as mildly affected (severity class 2; HOARSI 2; 32%). HOARSI 3 (32 individuals) and HOARSI 4 (15 individuals) were summarized as RAO affected (29%), as described above, and assigned to the severity class 3.

Representative Population Sample to Estimate Prevalences of HOARSI

Prevalences for the trait HOARSI were estimated in a random sample of the Warmblood population registered with the Swiss Association for Horse Sports (Schweizerischer Verband für Pferdesport, SVPS). This population has been described in detail in the doctoral thesis of one of the coauthors.11 In 2006, there were 16,750 horses registered as active in sports. The sample size was 376 horses, with the CustomInsight Survey Random Sample Calculatora with the error level set at 5% and the confidence level at 95%. With the expectation that at least 50% of the owners would provide information (>70% did in a previous study using this questionnaire),2 2 sets of 400 horses were required. To this end, 2 unique sets of random numbers in the range of 2 to 16,751 were generated with the Research Randomizer.b The SVPS provided information on the 16,750 horses in an Excel spreadsheet with a header row and 1 row for each horse and the random numbers were thereby used to select random horses from the spreadsheet. Using the first and the second set, the questionnaire was conducted until the desired sample size was reached. A total of 401 horses were surveyed by this random selection.

Of these 401, 353 were considered clearly unaffected and assigned to the severity class 1 (HOARSI 1), 29 were classified as mildly affected (severity class 2; HOARSI 2). HOARSI 3 (13 individuals) and HOARSI 4 (6 individuals) were summarized as RAO affected, as described above, and assigned to the severity class 3. Thus, the prevalences were 88% for the severity class 1, 7.3% for the severity class 2, and 4.7% for the severity class 3.

Segregation Analysis

The segregation analyses were performed on the 2 sire families, combined and separately, using severity classes and prevalences as described above. Likelihoods were computed with the Pedigree Analysis Package (PAP)12 and maximized with NPSOL.13 The analyses were based on the assumption that a possible major gene would be inherited in an autosomal way, as there was no evidence for an alternative. Within this framework, a maximum of 7 parameters was estimated. Of these 7 parameters 6 were used to characterize a major gene. They were the allele frequency in the pedigree (p), the transmission probabilities (tp) for the 3 genotypes AA, Aa, and aa to transmit the allele A to an offspring, the dominance effect (d), which places the heterozygotes relative to the homozygotes, and the displacement (t) which is the distance between the 2 homozygous genotypes in standard deviations. The 7th parameter, the heritability (h2), characterized the polygenic component. To investigate the possible presence of a major gene 5 models had to be evaluated. The models were compared in a hierarchical way so that 1 model estimated the same parameters as the other model, the submodel, but in addition at least 1 more parameter. Chi-square tests were used to compare the -2 ln likelihoods of a model and its submodel, the degree of freedom being the difference in the number of estimated parameters. In this comparison, the model with the smallest -2 ln likelihood value is the best model. If the test statistics show no difference between 2 models the submodel, which by definition has fewer parameters, is superior, because the additional estimated parameters do not contribute to the explanation of the data. In a first step the question that was addressed was whether the phenotype had a genetic background or not. For this purpose a general model was compared with an environmental model. The general model had no restrictions with respect to the transmission probabilities of the major gene, which means that all 7 parameters were estimated. In the environmental model the transmission probabilities were set equal to the allele frequency to cancel out the major gene effect and the heritability was fixed at zero. As the general model turned out to be better than environmental model, the next step was to compare the general model with a mixed inheritance model. In the mixed inheritance model the transmission probabilities were fixed to be Mendelian, otherwise it was the same as the general model. In this comparison the mixed inheritance model was better than the general model, so that finally a major gene model and a poygenic model were compared both with the mixed inheritance model. The major gene model was the same as the mixed inheritance model but the heritability was fixed at zero. In the polygene model only the heritability was estimated. These final comparisons allowed us to decide whether a major gene alone, a polygenic component alone, or a major gene together with a polygenic component explained our data best. In all models affection probabilities per genotype were calculated based on the parameter estimates and the prevalences.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgments
  9. References

In our analyses the general model was always better than the environmental model (P < .01), confirming a genetic background of the disease. In all 3 datasets the mixed inheritance model did not differ significantly from the general model (P= .62, P= 1.00, and P= .27) but was always better than the major gene model (P < .01) and the polygene model (P < .01) (Table 1). The parameter estimates for the major gene in the mixed inheritance model were strikingly different between the 2 sire families (Table 2). In the family of sire 1, the major gene was inherited in an autosomal recessive way (d= 0.00), while in the family of sire 2, the major gene was inherited in an autosomal dominant way (d= 1.00). The frequency of the deleterious allele also differed considerably between the 2 sire families with P= .23 and P= .06, respectively. In both sire families the displacement was large (t= 17.52 and t= 12.24, respectively) and the heritability extremely large with h2= 1 in both. In sire family 1, with the recessive inheritance mode, 10% of the offspring with genotype aa were in severity class 2 and 90% in severity class 3. However, only 7% of the offspring with genotypes AA and Aa were assigned to severity class 2 and none to severity class 3. In sire family 2, with the dominant inheritance mode, all individuals with Aa or aa genotypes were affected, with 60% in severity class 2 and 40% in severity class 3, and all offspring with genotype AA were unaffected. The analyses of the combined sire families led to similar parameter estimates as in sire family 1.

Table 1.   The -2 ln likelihoods for the different models of the segregation analyses using the 3 different datasets, family of sire 1, family of sire 2, and the 2 families combined.
ModelSire 1 and Sire 2Sire 1Sire 2
-2 ln LikelihoodsP Value-2 ln LikelihoodsP Value-2 ln LikelihoodsP Value
  1. The probabilities (P value) refer to hierarchical comparison of the models, where a smaller likelihood is better than a larger likelihood. If in the comparison both likelihoods are the same, the model estimating fewer parameters is better.

General518.03< .01286.85< .01238.44< .01
Environmental1137.07 595.36 541.71 
General518.03.62286.851.00238.44.27
Mixed inheritance519.81 286.85 242.33 
Mixed inheritance519.81< .01286.85< .01242.33< .01
Major gene713.13 390.99 327.37 
Mixed inheritance519.81< .01286.85< .01242.33< .01
Polygene857.05 455.65 401.78 
Table 2.   Frequency of the deleterious allele (p), affection probabilities by genotype (AA, Aa, and aa) for the severity classes 2 and 3, dominance effect (d), displacement (t), and heritability (h2) for the mixed model of the 3 datasets.
 PSeverity Class 2Severity Class 3dth2
AAAaaaAAAaaa
Sire 1 and sire 2.220.080.080.000.000.001.000.009.551.00
Sire 1.230.070.070.100.000.000.900.0017.521.00
Sire 2.060.000.600.600.000.400.401.0012.241.00

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgments
  9. References

Results of the present study suggest the presence of 1 or more major genes playing a role in the expression of RAO. For both sire families the mixed inheritance model appeared most likely, although the parameter estimates for the major gene differed significantly. In the family of sire 1 the major gene seems to be inherited in an autosomal recessive way, whereas in the family of sire 2 the major gene is likely to be inherited in an autosomal dominant way. A plausible explanation for these findings is genetic heterogeneity for the disease. The results thus suggest that the major gene in the 2 families is not the same, and that the gene involved in the expression of the disease in sire family 1 is possibly not involved in the expression of the phenotypically same disease in sire family 2 and vice versa.

By examining pedigrees by hand, previous investigators had observed that the inheritance mode of RAO did not appear to follow a simple Mendelian pattern,1 and they later proposed that the disease was polygenic.8 All data sets also indicated autosomal inheritance, which is in accordance with a similar risk conferred by either an affected dam or an affected sire1 and the absence of a sex predilection for RAO in general.1,14–16 Overall, our study supports a strong, but complex genetic background with locus heterogeneity for RAO.

The evidence for genetic heterogeneity is interesting in the light of our previous findings in the same 2 sire families. There is an association and linkage of microsatellite markers in the ECA13q13 region, which harbors the IL4R gene, with HOARSI in the family of sire 1 but not in the family of the second sire.9 One explanation for the different results in the 2 families would be that recombination has broken down the association between the paternal haplotype and disease locus. However, the present study supports genetic heterogeneity as the likely explanation. In this case, we would expect to find an association with a different candidate gene, presumably located elsewhere, in the offspring of sire 2. Thus, the results of the present and previous9 studies could help explain the contradictory findings regarding the immunological background of RAO, and in particular concerning the Th1 versus Th2 predominance, which some investigators have proposed,17–19 but others have questioned.20 Our findings suggest that results of immunological studies in equine RAO could depend on the genetic background of the animals used, an influential factor well documented in experiments using rodents.21–23

The large displacements of >6 on the liability scale in both sire families indicate that the distributions of the homozygous genotypes match well with the severity classes. This is reflected in the affection probabilities (Table 2). In the offspring of sire 1 93% of AA are unaffected, only 7% are mildly affected (HOARSI 2, severity class 2) and none are moderately to severely affected (HOARSI 3 and 4 combined, severity class 3, corresponding to RAO).2 Conversely, 90% of aa fall in the severity class 3, only 10% in severity class 2 and none are unaffected. Among sire 2 offspring, all of the AA are unaffected, while all of Aa and aa, corresponding to the dominant mode of inheritance, are affected either in severity class 2 (60%) or 3 (40%).

The parameter estimates in the combined families differed only slightly from the estimates in sire family 1 but strongly from the estimates in sire family 2 (Table 2). This could be explained by the larger overall number of offspring together with the almost 2-fold larger number of full sibs in sire family 1. More importantly, these results clearly demonstrate that in the case of genetic heterogeneity a joint analysis of families will lead to incorrect parameter estimates, which will affect subsequent parametric linkage analyses.

Using data from more than 2 families would strengthen the conclusion at the population level. Another limitation of our data sets is the missing phenotypes of the dams. Originally, we had attempted to also contact the owners of the dams, but most of them were impossible to locate and those breeders who we were able to contact were reluctant or unwilling to provide information about the disease status of their broodmares. It was disappointing, but perhaps understandable, to observe how the reluctance of the producers (breeders) of affected animals contrasted with the excellent compliance of the end users (owners). Also, we have not specifically investigated to what extent the sires' ancestors are represented in the dam lines, because a recent study23 had revealed that the Swiss Warmblood are a relatively heterogenous breed. Overall, the present family material, consisting of numerous offspring by affected sires, is quite unique and to our knowledge this is the first study investigating the inheritance mode of equine RAO.

The extremely large heritabilities estimated in both families support the hypothesis of a complex genetic background to this disease, in that 2 or more major genes could be responsible. Furthermore, HOARSI 3/4 prevalences were about 6 times higher in the 2 half-sib families than those estimated in the population sample. This confirms previous findings1,2,5–7 indicating a genetic background to RAO and clearly shows the inappropriateness of an all-environmental hypothesis.

On the other hand, the very high estimates of heritability could at first seem surprising, because it is well documented that RAO is strongly influenced by environmental factors, particularly hay feeding.2,3 It has long been documented that hay feeding is the major risk factor25 and that eliminating hay from the diet is the mainstay of RAO therapy.26 In the present study, however, the effects of hay feeding on the manifestation of RAO were not taken into account, because a history of at least 12 months of hay feeding was part of the inclusion criteria for all horses. Thus, the results of the present analysis can only be interpreted to mean that in this population, where all horses are fed hay by definition, other environmental factors play only a minor role in the expression of the disease, if any. Hay feeding appears to be an overriding “dominant” risk factor. In contrast, other reports have proposed environmental factors such as straw bedding, housing, and time horses spend outdoors3,15,24–27 as risk factors for lower airway disease in horses. In our previous study on the same 2 sire families, where some horses had been fed hay and others not, however, only hay feeding, but none of these latter risks factors had shown a significant effect on RAO.2

Conclusion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgments
  9. References

In conclusion, the segregation analyses clearly show the presence of a major gene playing a role in the expression of the disease. The results further suggest that the major genes may differ between families: in one family the inheritance mode appears to be autosomal recessive, whereas in the other it is autosomal dominant. These findings suggest a complex genetic background for RAO. The different parameter estimates are important for subsequent parametric linkage analyses. Furthermore, the results show that when genetic heterogeneity is present, combined segregation analyses can lead to erroneous results, and subsequent linkage studies should also include separate analyses for different families segregating the disease.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgments
  9. References

We thank U. Jost, N. Hasler, E. Laumen and all horse owners for their help. This study was supported by Vetsuisse and DKV grants, the Berne Equine Lung Research Group, and the Swiss National Science Foundation grant number 310000-116502.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgments
  9. References
  • 1
    Marti E, Gerber H, Essich G, et al. The genetic basis of equine allergic diseases. 1. Chronic hypersensitivity bronchitis. Equine Vet J 1991;23:457460.
  • 2
    Ramseyer A, Gaillard C, Burger D, Straub R, et al. Effects of genetic and environmental factors on Horse Owner Assessed Respiratory Signs Index (HOARSI). J Vet Intern Med 2007;21:149156.
  • 3
    Robinson NE, Derksen FJ, Olszewski MA, et al. The pathogenesis of chronic obstructive pulmonary disease of horses. Br Vet J 1996;152:283286.
  • 4
    Gerber V, Straub R, Marti E, Hauptman J, et al. Endoscopic scoring of mucus quantity and quality: Observer and horse variance and relationship to inflammation, mucus viscoelasticity and volume. Equine Vet J 2004;36:576582.
  • 5
    Schaeper W. Untersuchungen über die Erblichkeit und das Wesen des Lungendampfes beim Pferd Investigation into the nature and heritability of heaves in the horse. Tierärztliche Rundschau 1939;31:595599.
  • 6
    Koch P. Heredity of chronic alveolar emphysema of the lungs in horses. Deutsche Tierärztliche Wochenschrift 1957;64:485486.
  • 7
    Gerber H. The genetic basis of some equine diseases (Sir Frederick Hobday Lecture). Equine Vet J 1989;21:244248.
  • 8
    Marti E, Harwood L. Genetics of human asthma and equine recurrent airway obstruction: Can these two diseases be compared?. Pferdeheilkunde 2002;18:587.
  • 9
    Jost U, Klukowska-Rötzler J, Dolf G, et al. Association of microsatellite markers near interleukin-4 receptor α chain gene (IL4Rα) with chronic lower airway disease—locus heterogeneity in two high-prevalence horse families. Equine Vet J 2007;39:236241.
  • 10
    Anon. International Workshop on Inflammatory Airway Disease: Defining the Syndrome. Proceedings of Workshop in Boston, October 2002. Eq. Vet. Educ. 2002: 61–63.
  • 11
    Baleri D. Warmblut-Sportpferdepopulation in der Schweiz: Verwendungszweck, Haltung, Fütterung, Gesundheitsstatus und Gesundheitsmanagement Warmblood sporthorse population in Switzerland: Managementsystems, Feeding practice, Healthstatus and Healthmanagement]. Inaugural-Dissertation, University of Berne, 2008.
  • 12
    Hasstedt SJ. PAP: Pedigree Analysis Package (Rev. 4.0). Salt Lake City, UT: Department of Human Genetics, University of Utah; 1997.
  • 13
    Gill PE, Murray W, Saunders MA, et al. User's Guide for NPSOL (Version 4.0): A Fortran Package for Nonlinear Programming. Stanford: STANFORD UNIV CA SYSTEMS OPTIMIZATION LAB; 1986.
  • 14
    Gerber H. Chronic pulmonary disease in the horse. Equine Vet J 1973;5:2633.
  • 15
    McPherson EA, Lawson GH, Murphy JR, et al. Chronic obstructive pulmonary disease (COPD): Factors influencing the occurrence. Equine Vet J 1979;11:167171.
  • 16
    Dixon PM, Railton DI, McGorum BC. Equine pulmonary disease: A case control study of 300 referred cases. Part 2: details of animals and historical and clinical findings. Equine Vet J 1995;27:422427.
  • 17
    Lavoie JP, Maghni K, Desnoyers M, et al. Neutrophilic airway inflammation in horses with heaves is characterized by a Th2-type cytokine profile. Am J Respir Crit Care Med 2001;164:14101413.
  • 18
    Cordeau ME, Joubert P, Dewachi O, et al. IL-4, IL-5 and IFN-gamma mRNA expression in pulmonary lymphocytes in equine heaves. Vet Immunol Immunopathol 2004;97:8796.
  • 19
    Horohov DW, Beadle RE, Mouch S, et al. Temporal regulation of cytokine mRNA expression in equine recurrent airway obstruction. Vet Immunol Immunopathol 2005;108:237245.
  • 20
    Ainsworth DM, Grunig G, Matychak MB, et al. Recurrent airway obstruction (RAO) in horses is characterized by IFN-gamma and IL-8 production in bronchoalveolar lavage cells. Vet Immunol Immunopathol 2003;96:8391.
  • 21
    Nygaard UC, Aase A, Løvik M. The allergy adjuvant effect of particles—genetic factors influence antibody and cytokine responses. BMC Immunol 2005;21:611.
  • 22
    Yamashita Y, Charles N, Furumoto Y, et al. Cutting edge: Genetic variation influences Fc epsilonRI-induced mast cell activation and allergic responses. J Immunol 2007;179:740743.
  • 23
    Giavina-Bianchi P, Kalil J, Rizzo LV. Development of an animal model for allergic conjunctivitis: Influence of genetic factors and allergen concentration on immune response. Acta Ophthalmol 2008;86:670675.
  • 24
    Glowatzki-Mullis ML, Muntwyler J, Pfister W, et al. Genetic diversity among horse populations with a special focus on the Franches–Montagnes breed. Anim Genet 2006;37:3339.
  • 25
    Lowell FC. Observation on heaves. An asthma like syndrome in the horse. J Allergy 1964;35:322330.
  • 26
    Schatzmann U, Straub R, Gerber H, et al. Elimination of hay and straw as a therapy for chronic lung diseases in the horse. Tierarztl Prax 1974;2:207214.
  • 27
    Vandenput S, Votion D, Duvivier D, et al. Effect of a set stabled environmental control on pulmonary function and airway reactivity of COPD affected horses. Vet J 1998;155:189195.