A variety of surgical treatments and medical therapies are recommended for dogs with extrahepatic congenital portosystemic shunts (CPSS). The objective of this review was to assess the evidence base for the management of extrahepatic CPSS in dogs. An online bibliographic search was performed in November 2010 to identify articles relating to the question “Which of the treatment options for extrahepatic CPSS in dogs offers the best short- and long-term outcomes?” Articles were assigned a level of evidence based on a modified grading system. Thirty-eight articles were included in the review. Thirty-six articles were classified as grade 4 and two as grade 5. The timings and methods of assessment of short- and long-term outcomes varied widely between studies. One prospective study (grade 4a) showed that surgically treated dogs survived significantly longer than medically treated dogs. Four retrospective studies (grade 4b) compared the outcome of two surgical techniques but there were no statistically significant differences between treatment groups in terms of complications or outcome. The review found that the evidence base for the treatment of extrahepatic CPSS is weak. There is a lack of evidence of short- and long-term outcomes to recommend one treatment over another.
Over the last 10 years there has been growing emphasis on the use of evidence-based medicine (EBM) throughout the veterinary profession. There are a number of definitions of EBM but it is often described as “the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients” (Sackett and others 1996). This means that the best available evidence should be identified and integrated with clinical expertise and the needs of the patient in order to make appropriate clinical decisions. Several papers have reported the application of EBM in the small animal veterinary field to subjects as diverse as cruciate disease, atopic dermatitis, obesity, chronic kidney disease and emesis (Aragon and Budsberg 2005, Nuttall and Cole 2007, Roudebush and others 2008, Elwood and others 2010, Roudebush and others 2010). Evidence-based medicine is designed to optimise decision making in a given clinical situation and is, therefore, relevant to all practitioners. Unfortunately the evidence base remains weak for many clinical veterinary scenarios.
The application of EBM relies on five well defined steps (Phillips and others 1998).
- 1Asking focused questions. In order for the question to be of benefit to clinicians and their patients the question must be relevant to the patients’ problems and directed towards searching for relevant and precise answers.
- 2Finding the evidence. This is the process of gathering data relevant to the question but it is important to find the “best” evidence.
- 3Critical appraisal. All evidence is important but not all evidence is equal. The validity, results, impact, power, relevance and applicability of the evidence is carefully assessed. There are a number of systems published in the medical and veterinary literature for grading the quality of evidence in studies (Phillips and others 1998, Olivry and Mueller 2003, Roudebush and others 2004, Aragon and Budsberg 2005, Howick and others 2011).
- 4Making a decision. This involves the integration of the evidence into clinical decision making based on its relevance. The evidence is placed into context with current clinical expertise, the needs of the animal and the preferences of the owner.
- 5Evaluating performance. This is the assessment of the effectiveness of the above process. It also involves appraisal of the outcome of the decision. This can be recorded and in some instances may provide the impetus for further publication or research.
Extrahepatic congenital portosystemic shunts
A variety of treatment options have been recommended for the management of canine extrahepatic congenital portosystemic shunts (CPSS). These include medical management (Watson and Herrtage 1998) and various forms of surgical attenuation; acute suture ligation and gradual occlusion methods including ameroid constrictors, cellophane banding and intravascular techniques (Hunt and Hughes 1999, Leveille and others 2003, Hunt and others 2004, Mehl and others 2005). Surgical attenuation of CPSSs is reported to result in clinical recovery but has also been associated with recurrence of clinical signs due to persistent shunting or the development of multiple acquired portosystemic shunts. Reported outcome measures include clinical assessment, owner assessment, serum bile acid or ammonia tolerance testing and scintigraphy. However, it is unclear which of these represent the most appropriate assessment of outcome. The aim of this review was to examine the evidence for short- and long-term outcomes associated with the management of extrahepatic CPSS in dogs with a view to identifying whether one treatment method can be recommended over another.
Materials and methods
To establish the evidence base for the treatment of extrahepatic CPSS in dogs, the following question was composed:
“Which of the treatment options for extrahepatic CPSS in dogs offers the best short- and long-term outcomes?”An online bibliographic search was performed in November 2010 for articles relating to the treatment of extrahepatic CPSS in dogs. The search utilised the PubMed (http://www.pubmed.gov/), ISI Web of Science (http://wok.mimas.ac.uk/) and CAB direct (http://www.cabdirect.org/) databases. Databases were searched using the following terms: (portosystemic shunt OR portocaval shunt OR portovascular anomaly OR portosystemic communication OR extrahepatic vascular anomaly) AND (dog OR canine OR case) AND (treatment OR outcome OR mortality OR morbidity OR complications).
Analysis was restricted to the English language veterinary literature reporting information on the treatment of extrahepatic CPSS in dogs. The abstracts were reviewed for relevance to the question. Articles were excluded if they were experimental studies, described only acquired PSS or intrahepatic CPSS, were case reports or small case series (less than nine dogs) or did not provide any detail regarding outcome (e.g. articles describing diagnostic tests, and so on).
Articles that fitted these criteria were reviewed by the primary author and were assigned a level of evidence based on a modification of a grading system published by the Oxford Centre for Evidence-Based Medicine (Table 1). This grading system was simplified to remove some categories which did not readily apply to the available veterinary literature. In addition, Level 4 evidence was further sub-divided to enable greater differentiation among the evidence currently available. This allowed a distinction to be made between lower quality prospective cohort studies, retrospective cohort studies, case series specifically dealing with outcome, and other case series. As most of the articles fell into Level 4 evidence this enabled more detailed interpretation of the available evidence.
Table 1. Levels of evidence
|1a||Systematic review of randomised controlled trials (RCT)|
|1b||Individual RCT (with narrow confidence interval)|
|2a||Systematic review of cohort studies*|
|2b||Individual cohort study (including low quality RCT)|
|3a||Systematic review of case–control studies†|
|3b||Individual case–control study|
|4a||Lower quality prospective cohort/case–control study – concerns regarding definition of comparison groups and / or objective (preferably blinded) nature of assessment and /or consideration of confounding factors and / or adequacy of follow up|
|4b||Retrospective cohort/case–control study|
|4c||Case series – describing outcome for one treatment method with no control group|
|4d||Case series – describing novel aspect of management and providing some information regarding outcome|
|4e||Lower quality case series – concerns regarding study design and/or ability to interpret information|
|5||Expert opinion without explicit critical appraisal, or based on physiology, bench research or “first principles”|
The following data was recorded from each article: the type of study described (such as case series, cohort study, etc.); the size of the study population; the treatment used; the peri-operative complication and mortality rate (short-term outcome); details of long-term outcome including duration of follow-up and method of assessment. It was anticipated that the nature of the outcomes reported would vary between articles, in particular the nature and duration of long-term outcomes. Therefore, the way in which outcome was assessed, the timings and the result was recorded for each article.
Thirty-eight articles were identified as providing relevant information for answering the question. Thirty-six of the articles were classified as grade 4. None of the articles were graded higher than 4 (Table 2).
Table 2. Summary of grading of evidence for the treatment of extrahepatic CPSS in dogs
Direct comparison of different treatments
Seven articles were identified that compared the outcome of two different treatments (Tobias and others 1998, Tisdall and others 2000, Murphy and others 2001, Faverzani and others 2003, Hurn and Edwards 2003, Winkler and others 2003, Greenhalgh and others 2010).
The purpose of two of these articles was to compare medical and surgical treatments. The first article was very brief with insufficient information; it was classified grade 5 (Faverzani and others 2003). The second article described a prospective cohort study, comparing medical management with surgical treatment in a relatively large number of cases (99 surgical and 27 medical) (Greenhalgh and others 2010). The dogs were followed up for between 15 and 1,807 days and outcome was assessed on the basis of mortality alone. This study found that the probability of survival was significantly lower in medically treated dogs compared with surgically treated dogs. For medically managed dogs 13/27 (29·6%) died during the study period as a result of CPSS-related causes compared with 10/99 (10·1%) of surgically treated dogs (overall mortality for medically treated dogs was 48·1% and for surgically treated dogs was 12·1%). Four of the surgically treated dogs died during the postoperative period (up to seven days post surgery). This study was prospective but not blinded, there was potential for selection bias and the surgical treatments varied between dogs; thus the study was graded 4a.
Four articles compared suture ligation with ameroid constrictors (Tobias and others 1998, Murphy and others 2001, Hurn and Edwards 2003, Winkler and others 2003). The first article was a review which described some details of the outcome for dogs treated with suture ligation and dogs treated with ameroid constrictors (Tobias and others 1998). As this information was not presented as a scientific study and was incomplete the article was graded 5. The second article was a retrospective study comparing the short- and long-term outcome of suture ligation in 12 dogs with ameroid constrictor placement in 10 dogs (Murphy and others 2001). Short-term outcome was defined as peri--operative (prior to discharge) complications and mortality. There was no mortality in either group and although there were more postoperative complications in the suture group (50% versus 20%) this difference was not statistically significant. Complications included ascites, hypoglycaemia and neurological abnormalities. This study did find that surgery time was statistically significantly shorter for the ameroid group. Long-term outcome was assessed using a clinical rating score for five of the suture ligation dogs at one to 50 months and for nine of the ameroid dogs at 2·5 to 50 months. In the suture group two dogs were normal, two improved and one suffered recurrence of clinical signs. In the ameroid group seven dogs were considered normal and two improved. Despite a better clinical rating score for the ameroid group this difference was not statistically significant. Long-term outcome was also assessed by scintigraphy between 2·5 and 32 months and bile acid testing between one and 32 months. There was no statistically significant difference in abnormal shunt fraction or increased bile acids between the two groups. This study was graded 4b. The third article compared the retrospective short- and long-term outcomes between suture ligation in 29 dogs and ameroid constrictor placement in 16 dogs (Winkler and others 2003). Short-term outcome was assessed by peri-operative mortality for dogs that died or were euthanized within one week of surgery. There was no statistically significant difference in the peri-operative mortality rate between the two groups (3/29, 10·3% for the suture group and 1/16, 6·3% for the ameroid group). The study did not report any additional information regarding peri-operative complications. Long-term follow-up ranged from six weeks to seven years in the suture group and from two months to four years in the ameroid group and was based on survival, the presence of abnormal clinical signs, bile acid testing and evidence of persistent shunting or the development of multiple acquired shunts (MAS). Of the 24 dogs available for follow-up in the suture group two developed MAS, four developed evidence of persistent shunting and one recurrent clinical signs. Of the 15 dogs available for follow-up in the ameroid group two developed MAS and one recurrent clinical signs. One dog from each group died of CPSS-related causes. There was no statistically significant difference in long-term complication rate between the two groups. Bile acids were measured at least six weeks after surgery in 11 dogs from the suture group and in four from the ameroid group. All postoperative bile acid values were abnormal. This study was graded 4b. The fourth article was a retrospective study comparing the short-term outcomes (complications and mortality up to 14 days after surgery) of 20 dogs with suture ligation and 10 dogs with ameroid constrictor placement (Hurn and Edwards 2003). One or more intra-operative and/or postoperative complications occurred in five suture ligation dogs (25%) and five ameroid constrictor dogs (50%). Complications were varied and included small intestinal hypermotility, portal hypertension, hypothermia, hypoglycaemia, vomiting and seizures. Mortality was 3/20 (15%) for the suture group and 1/10 (10%) for the ameroid group. There was no statistically significant difference in mortality or complication rates between the two groups. Similar to Murphy and others (2001), this study found that surgery time was statistically significantly shorter for the ameroid group. This article was graded 4b.
Only one article compared suture attenuation with cellophane banding (Tisdall and others 2000). This retrospective study reported the short-term neurological complications occurring in 54 dogs treated with suture attenuation compared with 35 dogs treated with cellophane banding. Neurological complications occurred within six days of surgery in 7/54 (13%) of the suture group compared with 4/35 (11·4%) of the cellophane group. Two of the dogs in the suture group died or were euthanased as a result of their neurological complications. There was no statistically significant difference in neurological complications or mortality between the two groups. However, this study did not report details of other complications or the long-term outcome of this group of dogs. The article was graded 4b although it only provided information on one aspect of outcome.
Case series reporting outcome for one treatment
Many of the articles included in this review are simple retrospective (or occasionally prospective) case series reporting the short- and/or long-term outcomes of a given treatment for CPSS.
Eleven articles reported case series of dogs treated with suture ligation which specifically looked at outcome. Eight of these articles were graded 4c (Mathews and Gofton 1988, Lawrence and others 1992, Bostwick and Twedt 1995, Hottinger and others 1995, Komtebedde and others 1995, Smith and others 1995, Hunt and Hughes 1999, Kummeling and others 2004). The peri-operative mortality rate in these studies ranged from 2 to 28·8%. The duration and method of assessment of long-term follow-up varied considerably (Table 3). The remaining three suture attenuation articles were given a grade of 4e. Two of these studies described the outcome of suture attenuation in dogs with CPSS but included both extrahepatic and intrahepatic shunts and it was not possible to differentiate between the two in the results (Johnson and others 1987, Swalec and Smeak 1990). A third article did not provide adequate information to allow full interpretation of the results (Wolschrijn and others 2000).
Table 3. Summary of articles reporting short-term (peri-operative mortality) and/or long-term outcomes of suture attenuation of extrahepatic CPSS
Two articles reported case series of dogs treated with ameroid constrictors, including outcome (Vogt and others 1996, Mehl and others 2005). Peri-operative mortality rates were 7·1% and 16·7%. Both studies were graded 4c. Four articles reported case series of dogs treated with cellophane banding (Youmans and Hunt 1998, Hunt and others 2004, Frankel and others 2006, Landon and others 2008). Peri-operative mortality ranged from 0 to 9·4%. All four of these studies were graded 4c. A summary of the content of these six studies, including long-term outcome, is presented in Table 4.
Table 4. Summary of articles reporting short-term (peri-operative mortality) and/or long-term outcomes of ameroid constrictor placement and cellophane banding of extrahepatic CPSS
One article presented a retrospective case series of medically managed dogs, including nine with extrahepatic shunts (Watson and Herrtage 1998). Long-term outcome was assessed on the basis of survival. Two dogs were euthanased at six and 11 months post surgery for shunt-related reasons and one was euthanased at 28 months for a non-clinical reason. The remaining dogs were still alive 64 to 95 months following diagnosis. This study was graded 4c.
Other case series
Thirteen other articles were identified that, whilst reporting case series of dogs treated for CPSS, focused on a specific aspect of management other than outcome.
Two of these articles described combined results from dogs managed with more than one attenuation technique and were graded 4e (Parker and others 2008, Worley and Holt 2008).
The other 11 articles described dogs treated with suture attenuation (Tisdall and others 1994, Harvey and Erb 1998, Meyer and others 1999, Burton and White 2001, Szatmari and others 2004a,b, Kummeling and others 2006, 2010, Lee and others 2006, Doran and others 2008, Cariou and others 2009). Two of the articles were graded 4e as they reported a combined group of extrahepatic and intrahepatic shunts and it was not possible to separate the two groups in the results (Meyer and others 1999, Lee and others 2006). The remaining articles do not provide as high a level of evidence as those specifically evaluating outcome and were therefore graded 4d. For those articles which reported peri-operative mortality for extrahepatic CPSS the mortality ranged from 0 to 16·7% (Tisdall and others 1994, Harvey and Erb 1998, Szatmari and others 2004b, Kummeling and others 2006, 2010, Doran and others 2008, Cariou and others 2009).
This review found that the evidence base for the treatment of extrahepatic CPSS is very weak. No grade 1, 2 or 3 studies were identified which is unfortunate as these provide the best information for clinicians to make an informed choice regarding treatment. Most of the studies included in this review were retrospective cohort studies or retrospective case series and were graded 4 or below. Thus there is a lack of good evidence to support the use of one treatment over another.
The articles included in this study cover a long time period from 1987 to 2010. The most recent studies comparing two surgical treatments were published in 2003 (Hurn and Edwards 2003, Winkler and others 2003). Only one large case series of cellophane banding and one large case series of ameroid constrictors have been published to date (Hunt and others 2004, Mehl and others 2005). There has been a shift away from reporting studies specifically investigating outcome over the last five years. This could suggest that authors believe that sufficient evidence has been gathered in the existing literature to allow distinction between treatment options. The strength of any study which compares two treatments is dependent on whether patients were randomly allocated to treatment groups, whether the assessors were blinded, whether the treatment groups were similar, whether the follow-up was sufficient and whether groups were managed equally apart from the treatment. Therefore, when attempting to choose between treatments for CPSS ideally the decision would be based on evidence from prospective randomised controlled trials, comparing two or more treatments. Unfortunately such trials do not exist; decision making relies on prospective or retrospective cohort studies or case series. This review highlights that there is actually still a need for randomised prospective cohort studies comparing treatment options. Unfortunately, randomised prospective studies are not easily achieved. Issues of time, cost, case recruitment/numbers and owner/clinician compliance are all significant but not insurmountable obstacles.
This review identified a wide variation in how outcome for CPSS treatment is assessed and this represents a significant obstacle both to the comparison of existing articles and for the planning of new studies. Possible outcome measures for CPSS include short-term information such as peri-operative complications and mortality and long-term data focusing on recurrence of clinical signs, quality of life (QOL) and survival. This review found very little consistency between articles not only in the method used to assess outcome but also in the time periods concerned. There was wide variation between the time period that articles used to describe peri or postoperative mortality, ranging from days to weeks or not being specified (Hottinger and others 1995, Winkler and others 2003, Hunt and others 2004, Kummeling and others 2004). Long-term follow-up varies considerably between articles, both in duration and also in methods of assessment, i.e. clinical assessment, biochemical testing, ultrasonography, scintigraphy. The lack of consistency between studies makes direct comparison unhelpful. In order to allow fair comparisons, ideally all studies would use the same validated outcome measure. The inconsistency in length of follow-up may be particularly significant in dogs with CPSS as treatment is typically performed very early in their life. In addition some studies have shown that recurrence of clinical signs can occur three to six years after seemingly successful surgery (Mathews and Gofton 1988, Hottinger and others 1995, Komtebedde and others 1995, Kummeling and others 2004). Thus a long and consistent follow-up period is essential to identify any differences between treatments.
Short-term complications and mortality are important and relatively easy to quantify. Long-term outcome remains much more difficult to assess. Long-term outcome is also more difficult to obtain as it requires that the clinician maintain contact with the owner and the dog over time. Objective outcome measures include ammonia tolerance testing, dynamic bile acid testing, ultrasound, scintigraphy, contrast computed tomography and magnetic resonance imaging. These methods provide specific information, particularly on the presence of persistent or recurrent shunting. However, there is little information which can be used to relate these findings to the final outcome for the individual with some dogs having apparently good clinical outcomes with persistent shunting and many papers reporting persistently increased bile acids or continued shunting on scintigraphy in dogs that have made good recoveries (Komtebedde and others 1995, Hunt and Hughes 1999, Mehl and others 2005). A validated quality of life assessment index would be a useful additional tool to enable clinicians to determine a patient focussed outcome which may be more useful than a technically focused one. Such tools have been developed for other canine conditions including cardiac disease, chronic pain, spinal cord injury and cancer pain (Freeman and others 2005, Yazbek and Fantoni 2005, Wiseman-Orr and others 2006, Levine and others 2008). Use of a consistent outcome measure for CPSS would facilitate comparison between papers and techniques.
When considering management options for dogs with extrahepatic CPSSs ideally the treatment with the best short-term (mortality rate) and long-term outcomes would be selected. The review identified five articles which directly compared two treatment options and were considered to provide reasonable evidence. The paper by Greenhalgh and others (2010) provides the strongest evidence available and reports a reasonable number of dogs. This evidence would support the use of surgical management over medical treatment although it is still weak (grade 4a). Three grade 4b papers compared the outcomes between suture ligation and ameroid constrictor placement. This evidence is weak and based on small numbers; however, this is the most robust evidence that is available for choosing between surgical treatment options. Although these articles did identify differences in several different outcome measures between dogs treated with suture attenuation and ameroid constrictors these were not statistically significant. These studies included relatively small numbers of dogs in each group. It is entirely possible that there is a difference in terms of complications between the two treatments but there were insufficient numbers to detect it (a type II error). Differences between the treatment groups may be relatively small. In such instances the sample size required to detect a statistically significant difference is relatively large, which may be challenging to achieve in clinical veterinary research. When planning a study, a power calculation should be used to assess the number of participants required to detect a specified difference between groups and thereby ensure that worthwhile studies are undertaken. Consideration of this calculation should be included in the materials and methods of a study. None of the studies reviewed in this article reported having performed such a calculation. Two studies did report that ameroid constrictor placement was statistically significantly faster than suture attenuation. However, the importance or significance of this has not been investigated further although this may suggest that ameroid constrictor placement is easier and possibly cheaper. It can be concluded that there is not sufficient evidence to recommend the use of suture ligation over ameroid constrictor placement and vice versa on the basis of outcome. One grade 4b study compared the incidence of neurological complications between dogs treated with suture attenuation and cellophane banding. This found no significant difference between the two treatments but only focused on one aspect of outcome. No studies have compared ameroid constrictor placement with cellophane banding.
The majority of articles in this review were case series with no comparison or control group and thus provide weak evidence. Whilst these studies can justify the use of a given technique they do not provide evidence for whether one treatment is superior to another. Although the mortality rate is greater for some of the suture attenuation papers compared with the ameroid or cellophane band studies, for some the mortality rates are very similar. It is unknown whether any apparent differences in outcome are a direct result of the technique, are due to other factors or are a result of chance. Surgery is highly operator dependent and there can be a steep learning curve associated with a given technique. These factors should be taken into account when interpreting evidence because a given technique may not have the same results at a different institution. These case series also span a long time period (from 1987 to 2008). Other factors such as anaesthetic and postoperative care and greater familiarity with case management may have affected the results during this time period although this remains unproven. Whilst the evidence does not support the use of one technique over another based on outcome, there may be other compelling and valid reasons to use one option in a given situation, such as financial considerations, clinician familiarity or treatment availability.
A subjective grading system was used to categorise the articles included in this review. It is important to note that this grading system is not precise and some readers may disagree with the results. However, allowing for differences in opinion regarding the precise grade of the articles included, the overall results of this review would still be the same; all of the evidence would fall within grades 4 and 5. The articles were graded according to their evidence base related to outcome. Some articles provide a higher or lower level of evidence for another aspect of CPSS management. The grading process is not a criticism of specific articles. All of the articles included in this review contain important or useful information and make a valuable contribution to the literature on CPSS.
The review did not include any articles describing the use of intravascular occlusion techniques. This was unfortunate as intravascular treatments are likely to become more common place in the future, particularly for intrahepatic shunts. Several small case series have described the use of intravascular coils in dogs with extrahepatic CPSS but these did not include sufficient cases to meet the inclusion criteria (Leveille and others 2003, Bussadori and others 2008, Hogan and others 2010).
The veterinary profession should continue to strive to adopt an evidence-based approach in order to optimise patient care. It is clear that for the treatment of extrahepatic CPSS (as in many other areas of veterinary medicine) there is a lack of convincing evidence to recommend any one treatment over another. Therefore veterinary surgeons should integrate the existing knowledge with clinical experience and the needs of the animal to recommend the best treatment. In practice this means that clinicians should continue to use the technique that they are most comfortable with. However, veterinary surgeons should endeavour to improve the quality of research and hence the evidence base available. There are still many unanswered questions regarding the management of dogs with CPSS. Large randomised prospective studies are needed to compare existing and new treatments in order to determine which are associated with the best outcome for dogs. Further investigation is also needed into developing consistent and validated outcome measures, possibly including a validated quality of life assessment tool.
The authors are very grateful to the Kennel Club Charitable Trust who generously provided a grant to support MST’s PhD studies.
Conflict of interest
None of the authors of this article has a financial or personal relationship with other people or organisations that could inappropriately influence or bias the content of the paper.