Does occurrence of ventricular arrhythmia reduce the survival rate in dogs with gastric dilatation volvulus (GDV)?

[...]

Appraisal, application and reflection

Relevant to this Knowledge Summary, four papers were found which investigated the effects of ventricular arrhythmias on the mortality rates of dogs with GDV. Of these, three were retrospective, observational, case control studies (Brourman et al., 1996; Green et al., 2012; and Mackenzie et al., 2010) and one was a prospective observational study (Aona et al., 2017).

The first retrospective observational study (Green et al. 2012) reviewed the presence of radiographic abnormalities and their effects on mortality in dogs with GDV. Logistic regression analysis was used to evaluate the risk that presenting clinical signs (including presence of ventricular arrythmia) may have on chance of survival to discharge.

Only 2/16 (13%) of the dogs that died or were euthanised during their hospitalisation were diagnosed with ventricular arrhythmias and it was concluded that the presence of ventricular arrhythmia was not strongly associated with increased patient mortality rates. Neither P value or confidence intervals (CI) were stated for ventricular arrhythmia, and it was unknown if they were simply not calculated or that they were not found to be of statistical relevance (e.g. P < 0.05 or CI > 95%). Calculating these values would give a better indication of the strength of evidence that the presence of ventricular arrhythmia may affect survival to discharge.

Ventricular tachycardia can be caused by both extra cardiac pathologies as well as primary heart disease, which can often be seen on thoracic radiographs (Rishniw, 2020). In the paper by Green et al. (2012), 6/101 (6%) of the cases included in the study were radiographed upon admission due to suspected cardiac disease but it was not stated what proportion, if any, of these dogs presented with or developed ventricular tachycardia during hospitalisation. There may have been bias present, where only dogs with suspected cardiac disease were fully examined for the presence of cardiac arrhythmias, meaning that other dogs may have developed arrhythmias that went undiagnosed. It was also not stated how a diagnosis of ventricular arrhythmia was made in the 6% of dogs that presented with the condition. However, the paper itself was detailed and well-presented, and the results of the study were clear and logical.

Mackenzie et al. (2010), a retrospective, observational study, evaluated the factors significantly affecting overall and postoperative mortality rates in dogs anaesthetised for surgery to correct GDV. The study found that preoperative cardiac arrhythmias of any origin; as well as postoperative ventricular tachycardia (not concurrently) were associated with a higher overall mortality rate among the 306 cases that were evaluated. Statistical analysis was not performed on the collected data, meaning that the strength of association between ventricular arrhythmia and mortality rate was not able to be properly evaluated.

Furthermore, incomplete medical records and the retrospective nature of the study may mean that some aspects of the included cases could not be properly analysed. Additionally, there was confusion as to the origin of some of the figures presented in the results section of the paper as well as the conclusion that was drawn with regards to the incidence of preoperative intermittent ventricular arrhythmias. It was also stated that preoperative IVA’s were associated with ‘much higher mortality rates,’ but only provided a P value to support this statement, rather than the number of dogs with preoperative IVA that died.

The third retrospective observational study that was evaluated (Brourman et al., 1996) looked at factors affecting mortality in the perioperative period of dogs admitted to either a university hospital or a private veterinary clinic with GDV. Only those cases that underwent surgical correction (which predominantly involved decompression and repositioning of the stomach) were included in the study. Fisher’s Exact test was used to determine the strength of association between patient factors and mortality, with P < 0.05 indicating that a significant association existed.

Of the included cases, 92/137 (67%) were found to have developed a cardiac arrhythmia during their hospitalisation; of which 20/92 (22%) died during the perioperative period. The mortality rate for those dogs that developed preoperative cardiac arrhythmias was found to be 35/92 (38%), more than double that of the mortality rate of dogs without preoperative cardiac arrhythmias 18/121 (15%). Preoperative cardiac arrhythmias were found to have a P value of greater than 0.05, indicating a considerable association with patient mortality. The results did not differ significantly between the two study centers. Although it would have been a useful tool for evaluating the strength of evidence presented by this paper. Confidence intervals were not calculated in this study.

Both hematology and biochemistry parameters were monitored for every dog involved in the study, but none were found to have any impact on patient mortality or the chances of developing cardiac arrhythmias.

While all instances of cardiac arrhythmia were diagnosed through clinical exam and electrocardiography, there was no further investigation conducted into the origin or categories of arrhythmia present. Additional clinical research could aid in determining which types of cardiac arrhythmia are significantly associated with increased mortality in dogs with GDV. 

The remaining paper, a prospective, observational study (Aona et al., 2017) was the most recent study conducted on the subject; and evaluated electrocardiography and cardiac biomarker concentrations taken from dogs admitted to a university hospital for suspected GDV from May 2011 to October 2012. Dogs that died prior to surgery or were euthanised were excluded from the study. Ventricular arrhythmias were identified and graded using a previously published scale of 0–4, with 0 representing the absence of VPCs (ventricular premature complexes) and 4 being ventricular tachycardia. Of the 15 dogs that developed ventricular arrhythmias, nine were categorised as Grade 1 and six were categorised as Grade 4.

A strong association was found between increased concentrations of cTn1 and the likelihood of a patient developing a Grade 1 or 4 ventricular arrhythmia. The P values for dogs with Grade 1 and Grade 4 were found to be P < 0.001 and P = 0.002 respectively, showing a strong correlation between these data sets.

However, arrhythmia grade was not found to be a significant contributing factor in patient survival following surgical correction of GDV. It is unknown whether P values were calculated to show the strength of correlation between presence of ventricular arrhythmia and mortality rate. Doing so may have helped to statistically show the strength of evidence of the study in relation to the PICO question.

Categorical data in more than two categories (arrhythmia grade) was compared using both chi-squared and Friedman tests. Biomarker concentrations were classed as continuous variables and compared using Pearson’s and Spearman’s rank correlations.

While the design of the study was well laid out and thorough, the sample size was not particularly large, and as such may not have provided results that were as accurate as they could have been. Additionally, it was not stated whether the presence of ventricular arrhythmia alone was considered a major factor in overall patient mortality and survival post discharge.

Large breed, deep-chested dogs have been found to be statistically more likely to develop GDV and indeed, in the papers where breed was recorded, it was these dogs that were overrepresented. It should be noted that smaller dogs may still develop GDV, although not with the same frequency (PDSA, 2020).

In conclusion, while some of the evaluated papers provided evidence that may suggest ventricular tachycardia has an association with increased patient mortality in dogs with GDV, there are several weaknesses in the methods and results that prevent conclusive assumptions from being made. Primarily, the cases included in the appraised studies, with the exception of Brourman et al. (1996), were collected from a single hospital. For the conclusion of this Knowledge Summary to be applicable to small animal general practice in the UK, it would be preferable that this be drawn from the largest population and geographical area possible.

All but one of the papers (Mackenzie et al., 2010) used 137 dogs or fewer, with only 22 dogs being included in the study conducted by Aona et al. (2017). As such, any confounding factors present are more likely to skew the results obtained by the study, and it is significantly more difficult to extrapolate to the general population on which the outcome of the study was to be applied.

It could also be assumed, as a general rule, that veterinary university hospitals are better equipped than most small animal clinics, meaning that the dogs may be more likely to survive to discharge at these hospitals, irrespective of their cardiac status.

There is some evidence to suggest that the presence of preoperative ventricular arrhythmias may be associated with reduced survival rate in dogs presented for GDV, but further research on the subject is required to definitively prove this hypothesis. A conclusive answer to this PICO question could allow clinicians in practice to better predict survival rates for patients and provide more accurate information to owners regarding the chances of success for continued treatment of their animal.

Further research on the subject may include larger, multi-centre observational studies with a focus on diagnosing and monitoring cardiac arrythmias in dogs diagnosed with GDV. A larger number of cases included would allow for more accurate and reliable results, while a multi-center study would be more reflective of the wider population, rather than just that of a single centre study. Further, in-depth statistical analysis may also be of use in determining strength of correlation between ventricular arrhythmias and mortality rates in dogs with GDV.

[…]