Final report on the Bloat survey conducted by AHT/KC

Received by Joint Breed Clubs' Health committee

 

Animal data.

Surveys were completed for 1911 unique animals, from 1091 litters (mean 1.75 animals per litter), with 412 unique sires (mean 4.64 animals per sire) and 848 unique dams (mean 2.25 animals per dam). The breakdown of animals per litter is shown in table 1.

 

# in litter

count

%

1

595

54.54%

2

293

26.86%

3

118

10.82%

4

63

5.77%

5

14

1.28%

6

4

0.37%

7

3

0.27%

8

0

0.00%

9

1

0.09%

Table 1. Distribution of animals per litter in survey.

 

Of the 1911 unique animals for which surveys were completed, 1046 were female (54.74%) and 865 were male (45.26%). Table 2 shows cross-tabulation of neuter status with sex.

 

unknown

entire

neutered

 

13 (1.2%)

733 (70.0%)

300 (28.7%)

females

7 (0.8%)

667 (77.1%)

191 (22.1%)

males

Table 2. Cross-tabulation of neuter status with sex.

 

There was a large range of year of birth (yob), from 1991 to 2013 (the year of survey), with 93% of respondent animals born between 2000 and 2011.

 

Bloat data.

1657 of 1911 animals were reported as never having experienced an episode of bloat (86.71% of animals). 254 animals (13.29%) were reported as having experienced at least one episode of bloat at the time of survey, with details on 472 incidents of bloat supplied. Table 3 shows the distribution of animals and incidents by number of episodes reported.

Of the 472 reported incidents of bloat, just over half resulted in surgery or the death of the animal (50.85%), while 230 episodes were reported to have been resolved spontaneously, by management or medication (48.73%), see table 4. 

 

episodes

animals

incidents

1

165

165

2

39

78

3

8

24

4

5

20

5+

37

185

total

254

472

Table 3. Distribution of number of reported episodes of bloat per animal.

 

Treatment / outcome

Count

Percent

Don’t know

2

0.42%

Resolved itself

83

17.58%

Managed through diet/lifestyle changes…

92

19.49%

Managed with medication

55

11.65%

Surgery

194

41.10%

Died or put to sleep

46

9.75%

Table 4. Distribution of treatment/outcome of bloat epsidodes.

 

Data on multiple cases of bloat per animal were collapsed to generate a data field indicating whether the animal had ever experienced an episode of bloat (0=no, 1=yes) with an accompanying field specifying either age in days at time of the survey (where bloat field = 0), or age in days at earliest episode (where bloat field =1).  This was repeated for bloat where treatment/outcome was specified as surgery or death/put to sleep.

Examination of the prevalence of animals experiencing bloat by yob revealed bias, with the prevalence tending to be higher in earlier yob (table 5), over the majority of the data (yob 2000-12). This bias is consistent with bloat having being more commonly observed in older animals.

The minimum age at survey for animals not having experienced bloat was 176 days, and the maximum was 5454 days (14.94 years). The mean (μ) and standard deviation (σ) age of animals at survey which had not experienced an episode of bloat was: μ= 2218.46 days, σ= 1161.05 days (6.08 years and 3.18 years respectively). The minimum age supplied for an animal experiencing its earliest episode of bloat was 61 days, and the maximum age for an animal experiencing its first episode was 4593 days (12.58 years).  The mean and standard deviation of age at earliest episode were: μ= 1755.30 days, σ= 1116.38 days (4.81 years and 3.06 years respectively).

The minimum age at survey for animals not having experienced bloat resulting in surgery or death was 176 days, and the maximum was 5454 days (14.94 years), with μ= 2222.82 days, σ= 1164.11 days (6.09 years and 3.19 years respectively).The minimum age supplied for an animal experiencing its earliest episode of bloat resulting in surgery or death was 152 days, and the maximum earliest age reported was 4593 days (12.58 years), with μ= 1959.75 days, σ= 1097.18 days (5.37 years and 3.01 years respectively). 

 

 

yob

Bloat with all treatments & outcomes

Bloat resulting in surgery or death

 

no bloat

bloat

prevalence

no bloat

bloat

prevalence

unknown

2

1

33.33%

2

1

33.33%

1991

0

2

100.00%

0

2

100.00%

1992

2

1

33.33%

3

0

0.00%

1993

1

0

0.00%

1

0

0.00%

1994

0

1

100.00%

0

1

100.00%

1995

1

1

50.00%

1

1

50.00%

1996

1

3

75.00%

3

1

25.00%

1997

2

3

60.00%

2

3

60.00%

1998

1

3

75.00%

2

2

50.00%

1999

6

2

25.00%

6

2

25.00%

2000

62

14

18.42%

62

14

18.42%

2001

79

20

20.20%

81

18

18.18%

2002

90

26

22.41%

93

23

19.83%

2003

80

27

25.23%

85

22

20.56%

2004

114

25

17.99%

120

19

13.67%

2005

150

22

12.79%

153

19

11.05%

2006

143

27

15.88%

145

25

14.71%

2007

155

17

9.88%

159

13

7.56%

2008

170

12

6.59%

173

9

4.95%

2009

148

20

11.90%

150

18

10.71%

2010

179

15

7.73%

186

8

4.12%

2011

174

6

3.33%

178

2

1.11%

2012

92

6

6.12%

95

3

3.06%

2013

5

0

0.00%

5

0

0.00%

  Table 5. Number and prevalence of animals experiencing bloat by yob

 

Genetic analysis.

A range of preliminary mixed models were run aiming to estimate the additive genetic variation in developing bloat. The proportion of phenotypic (measurable/observable) variation comprised of additive genetic variation is known as the heritability and describes the influence of genetics on the trait and how it will respond to selection.

Estimates of heritability from these preliminary models ranged from approximately 0.04 to 0.2, varying with the data and model used (data are routinely trimmed to minimise the number of classes of effect needing to be estimated; for example data may be limited to specific yob to avoid requiring the model to estimate effects of yob classes where data is scarce, such as 1991/2/3 etc, see table 5). Binary data (0/1), as used here, suffer a loss of precision in risk description in comparison to more continuous data (for example hip score ranges from 0 to 106), and this tends to lower the estimate of heritability (methods of adjustment are available). Thus, normally we might expect the true heritability to be higher than the estimates quoted above.

However, it was noted that the estimated effects of age from the analysis were extremely large, particularly when yob was also included in the model, and ran counter to previous reports (i.e. the models estimated that risk of bloat decreased as animals got older).  This is a worrying observation, and is consistent with another bias in this data; that there appeared to be a higher prevalence of bloat in younger animals than in older animals. While we may surmise at the cause of this bias, there is no information on what actually causes it.

Therefore, while we have detected a signature of incidence compatible with genetics (estimated in the form of heritability), regrettably we do not think we can claim beyond doubt to have established that there is genetic variation in the likelihood of developing bloat. There are significant biases in the current dataset and the issue is whether these biases would either influence or produce the signature detected as an artefact. Although we feel this is unlikely, the presence of the biases described unfortunately prevent us from claiming to have determined risk of bloat is heritable in the Irish Setter.

 

Further work.

A further method of data analysis is currently being explored which may be particularly suited to data such as was collected from the survey. Survival analysis models the survival function from data of the age at which animals succumb to disease or death (and also the hazard function as the risk of disease at a particular age), and is a flexible way to incorporate censored data (i.e. animals unaffected or still alive at time of survey). Survival Kit software has been tailored towards animal breeders and allows the estimation of random effects (i.e. genetic variation). Work on using Survival Kit software and running survival analysis is being undertaken.

 While we are unable to state categorically that bloat is heritable in the Irish Setter, the evidence does appear to indicate that genetics could play an important role. Given the prevalence of bloat in this breed, and the concerns of breeders and owners, it may be useful to attempt to ‘extend’ the survey, by building and maintaining a database of details of dogs affected and unaffected by bloat. As the amount of data increases, it becomes easier to reduce the effects of bias seen in the current data set, for example if enough data were available analysis could focus solely on older dogs, reducing the influence of age (and providing a more reliable indicator of ‘lifetime’ risk). Assistance in this undertaking may be available from the Kennel Club.

Breeders may also wish to explore the idea of taking and storing DNA samples from both affected and unaffected dogs alongside a database of information as suggested above. While it is highly unlikely that a mutant variant of a single gene is solely responsible for the development of bloat, genetics could certainly have an important role, as stated above. As such, taking and storing DNA samples may bequeath the breed in future with an extremely useful resource to begin to track down regions of the DNA and maybe even genes which are responsible for increasing or decreasing the risk of developing bloat.  Because there are almost certainly multiple risk factors for bloat, some of which might be genetic, it will be necessary to collect DNA from large numbers (many hundreds) of dogs that have suffered at least one attack of bloat as well as DNA from a similar number of old dogs that have never had bloat before a genetic study is likely to be successful.

 

January 2017

...........................................

INTERIM REPORT

KC/AHT Bloat Survey.

The AHT/KC survey of Irish setters last year, with financial support of the JISBC, was aimed primarily at gathering evidence to prove that bloat is heritable. If proven, DNA samples could then be evaluated for possible genetic markers. However, it was always understood that it was very unlikely to be a simple inheritance (i.e. single gene mutation) pattern. Whilst there may be genetic susceptibility, environmental factors are likely to play a significant role.

Regrettably, although the results raised a suspicion that there is an inherited component to the condition, confounding factors prevented the results reaching statistical significance and therefore proof. The number of respondents and the fact that some of the dogs alive at the time of the survey may go on to develop bloat later in life confounded any clear conclusion.

The recommendation must therefore be that the breed prospectively collects DNA samples from dogs that suffer bloat. Whilst this outcome is disappointing, it is recommended that we work with the AHT to collect DNA samples and pedigrees of dogs when they suffer bloat confirmed by surgery or post mortem.

As you will be aware, the survey also gathered information on a number of other conditions. This data has not yet been analysed, but I have asked Tom Lewis (KC, formerly AHT) to analyse results pertaining to megaoesophagus and epilepsy, where it is more likely that heritability can be proven.

Ed Hall

Chairman, Joint Irish Setter Breed Clubs Health Committee

22/3/2015

 

Tags: