A Retrospective Study of Factors Influencing Survival Following Surgery for Gastric Dilatation-Volvulus Syndrome in 306 Dogs

Introduction

Gastric dilatation-volvulus (GDV) is a life-threatening condition in dogs that has historically been associated with high mortality rates. Early studies reported overall mortality rates of 33% to 68%.^1,2 In those studies, splenic and gastric injuries were not examined as independent factors influencing survival. More recent studies demonstrate that overall mortality rates have declined to 15%,^3–5 but they remain higher when splenectomy (32%),⁴ partial gastrectomy (35%),⁴ or both (55%)⁴ are performed.

The influence of the duration of ischemia on tissue viability in dogs with GDV has been well documented.^6,7 The duration and degree of gastric and splenic ischemia in experimental studies of GDV have been shown to directly influence organ viability.^8–10 In clinical studies, the duration of signs of GDV has been thought to be directly related to increased risk of gastric and splenic injury and mortality.^11–13 Based on this scientific evidence, the following techniques were recommended to improve outcomes: rapid decompression of the stomach; prompt fluid resuscitation; rapid transition to surgery for quick and efficient repositioning of the stomach and spleen; and prompt management of damaged tissues.

The primary goal of this study was to update the knowledge of overall and postoperative mortality rates and factors influencing mortality in a large number of dogs with GDV. The influences of several factors on survival, which had not been previously studied, were evaluated. We tested the hypothesis that the time from hospital presentation to surgery (i.e., the time required for diagnostic testing and stabilization) and the times required for anesthesia and surgery all directly influence both the overall and postoperative mortality rates. We also compared the overall and postoperative mortality rates identified in this report with rates in previous studies to determine if there has been improvement.

Materials and Methods

The medical records for all dogs diagnosed with GDV from 2000 to 2004 were retrospectively evaluated. A diagnosis of GDV was based on right lateral abdominal radiographs. Only data from dogs that were anesthetized for surgery and confirmed to have GDV were included in the study.

Medical records were evaluated for the following: signalment; duration of clinical signs prior to presentation (i.e., time from when the owner noticed the dog was sick to presentation at the hospital); interval from presentation to surgery (i.e., the initiation of the skin incision); evidence of preoperative cardiac arrhythmias; duration of anesthesia; duration of surgery; splenic injury treated by splenectomy; gastric injury treated by partial gastrectomy; and postoperative cardiac arrhythmias.

Preoperative treatments, complete blood count results, and serum biochemical profile results were not evaluated in this study. Routine preoperative treatment consisted of needle gastric decompression and administration of 0.9% saline or lactated Ringer’s solution as a bolus of 60 to 90 mL/kg intravenously [IV] over the first hour, followed by 5 mL/kg per hour until the time of anesthetic induction.

Preoperative and postoperative cardiac arrhythmias were recorded when physical evidence of pulse deficits and/or electrocardiogram abnormalities were noted. A dog was considered to have intermittent ventricular arrhythmias when the heart rate was <150 beats per minute and the ventricular ectopic beats made up <50% of the beats in 1 minute. When the heart rate was >150 beats per minute and the ventricular ectopic beats made up >50% of the beats in 1 minute, the dog was considered to have ventricular tachycardia. When ventricular tachycardia was present, the dog was treated with one or more boluses of lidocaine (1 to 8 mg/kg IV), followed by a continuous-rate infusion (CRI) of lidocaine (25 to 75 μg/kg IV per minute). The lidocaine infusion was continued until arrhythmias improved, then it was slowly decreased over a 12-hour period by cutting the CRI in half every 6 hours.

A similar anesthesia protocol was used in all dogs. Premedication was a combination of hydromorphone^a (0.05 mg/kg IV) and diazepam^b (0.2 mg/kg IV). Propofol^c (4 to 6 mg/kg IV) was administered for induction, intubation was performed, and anesthesia was maintained with either isoflurane^d or sevoflurane^e in pure oxygen. Fluids were continued intraoperatively at a rate of 10 to 20 mL/kg per hour. Appropriate fluid rate and fluid composition were adjusted as needed based on each individual case to address areas of concern (e.g., heart rate, blood pressure). Surgery was performed by a Diplomate of the American College of Veterinary Surgeons or a surgery resident. Surgical treatment always included an exploratory laparotomy, gastric decompression by orogastric tube or needle gastrocentesis, repositioning the stomach, and incisional gastropexy.¹⁴

Splenectomy (using polydioxanone^f or chromic gut suture^g ligation), partial gastrectomy (sutured, one- or two-layer closure with polydioxanone suture), or both surgical procedures were performed when deemed necessary according to the following subjective criteria. Splenic viability was assessed based on color, compliance, vascular integrity, and/or response to reperfusion. Vascular integrity of the spleen was assessed based on evidence of vascular thrombi, presence of blood flow and/or a pulse in the splenic artery and veins, and evidence of hemorrhage from splenic vessels or parenchyma. Response to reperfusion was assessed by examining the changes in color, compliance, and vascular integrity once the spleen was repositioned following gastric repositioning and decompression. Assessment of gastric viability was based on serosal color, gastric wall texture, and vascular integrity. The vascular integrity of the stomach was assessed by examining blood flow in surface vessels, blanching and reperfusion with digital pressure, and bleeding at the cut edge in cases of resection. Extensive gastric injury was considered to have occurred when it appeared that >60% of the stomach was irreversibly damaged and the cardia region was extensively involved.

Postoperatively, dogs were maintained on crystalloid fluid therapy at a rate of 2 to 4 mL/kg per hour, and pain management was based on doctor preference and included either hydromorphone^a (0.05 mg/kg as needed) or fentanyl^h (2.5 to 5 μm/kg per hour). Types of postoperative fluids and pain management medications were not evaluated for effect on postoperative mortality.

The overall mortality rate was defined as the ratio of the number of dogs that died or were euthanized either intraoperatively or postoperatively to all dogs that were anesthetized for surgery. The postoperative mortality rate was defined as the ratio of dogs that died or were euthanized during the postoperative period to all dogs that survived anesthesia. The postoperative period was defined as the time from the end of anesthesia to hospital discharge.

A statistical consulting serviceⁱ and software package^j were used to evaluate data. Logistic regression models were used to evaluate the effects of breed, sex, age, and the presence of any cardiac arrhythmia on postoperative mortality. Logistic regression models were also used to evaluate the effects of duration of clinical signs, interval until surgery, anesthesia time, surgery time, and type of surgery performed (e.g., incisional gastropexy alone, incisional gastropexy with splenectomy or incisional gastropexy with partial gastrectomy, or an incisional gastropexy with both splenectomy and partial gastrectomy) on postoperative mortality rate. Hosmer-Lemeshow goodness-of-fit tests were performed for evaluating the logistic regressions. The overall and postoperative mortality rates of the dogs in this study were compared with the more recent studies by Brockman et al³ and Brourman et al,⁴ in which a one-sided binomial proportion test was used when the data were presented in a fashion allowing clear comparison. The lesser percentage between the two studies was used when comparing for significance. Findings were considered significant when P<0.05.

Results

Of the dogs diagnosed with GDV between 2000 and 2004, 306 met the inclusion criteria for the study. The overall mortality rate was 10% (30 of 306 dogs), and the postoperative mortality rate was 6.1% (18 of 294 dogs), as illustrated in Figure 1⇓. Of the 30 dogs that died or were euthanized, 12 died intraoperatively. Four of the 12 died spontaneously, and the other eight were euthanized intraoperatively based on the surgeon’s assessment of extensive gastric damage.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1—

Overall and postoperative mortality rates in all dogs with GDV that had incisional gastropexy alone (GP), incisional gastropexy and partial gastrectomy (GP/GX), incisional gastropexy and splenectomy (GP/SX), or all three surgical procedures (GP/GX/SX). Asterisk (*) indicates a significant (P<0.05) association with mortality.

Thirty-eight breeds were represented in this study. The most common breeds included German shepherd dogs (n=40, 13.1%), Great Danes (n=30, 9.8%), standard poodles (n=24, 7.8%), Labrador retrievers (n=19, 6.2%), large mixed-breed dogs (n=18, 5.9%), golden retrievers (n=17, 5.6%), Akitas (n=17, 5.6%), Doberman pinschers (n=14, 4.6%), and chow chows (n=13, 4.2%). Breed did not affect the overall mortality rate in this study (P=0.168). In addition, age and sex did not significantly affect the overall mortality (P=0.339 and P=0.447, respectively), as described in Table 1⇓. The mean age of dogs was 8±3.54 years (range 7 months to 16 years). One hundred twenty-three (40%) dogs were neutered males, 99 (32.7%) were spayed females, 58 (19.0%) were intact males, and 25 (8.2%) were intact females.

The mean duration of clinical signs was 7.4±3.44 hours (range 2 to 12 hours, median 6 hours). As shown in Table 1⇑, duration of clinical signs did not affect overall mortality (P=0.417). The mean time from presentation to surgery was 75±26.6 minutes (range 15 to 300 minutes; median 60 minutes). A significant inverse relationship was found between time from presentation to surgery and the overall mortality rate (P=0.035; odds ratio 0.93; 95% confidence interval 0.86 to 0.99). In other words, an increased time from presentation to surgery was associated with a lower overall mortality rate.

The only arrhythmias recorded in this cohort of dogs were of ventricular origin. Preoperative cardiac arrhythmias were recorded in 11% (16 of 147) of dogs presented for GDV. Twelve dogs had intermittent ventricular arrhythmias, and four dogs had ventricular tachycardia. Of the dogs with preoperative intermittent ventricular arrhythmias, four died. No deaths occurred in dogs diagnosed with preoperative ventricular tachycardia. Preoperative intermittent ventricular arrhythmia was associated with a significantly higher overall mortality rate (P=0.035), but preoperative ventricular tachycardia was not associated with an increased overall mortality rate (P=0.088). Forty-eight percent of dogs with preoperative cardiac arrhythmias subsequently underwent splenectomy and/or partial gastrectomy, whereas only 27% of dogs without preoperative cardiac arrhythmias required splenectomy and/or partial gastrectomy; however, this difference was not significant (P=0.056).

Postoperative cardiac arrhythmias were present in 133 of 172 dogs. Of these 133 dogs, 105 had intermittent ventricular arrhythmias and 28 had ventricular tachycardia. Postoperatively, one of the dogs with intermittent ventricular arrhythmias died (P=0.47), and four of the dogs with ventricular tachycardia died (P=0.035), suggesting that postoperative ventricular tachycardia was associated with an increased postoperative mortality rate [Table 1⇑].

The mean duration of anesthesia and mean duration of surgery were 67±20.44 minutes (range 20 to 240 minutes; median 60 minutes) and 48±30 minutes (range 20 to 240 minutes; median 30 minutes), respectively. No association was seen between the length of time for either parameter and postoperative mortality rate [Table 1⇑].

Incisional gastropexy was performed in all 294 dogs that survived surgery. Dogs that had gastropexy alone (n=212) had a postoperative mortality rate of 3% (n=6). Splenectomy was performed in 61 (21%) of the 294 dogs. Dogs in which splenectomy was the only additional surgical procedure performed (34 of 61 dogs) had a significantly higher postoperative mortality rate (15%; P=0.008) than dogs that did not require splenectomy. Partial gastrectomy was performed in 58 (20%) of 294 dogs. The postoperative mortality rate (9%; P=0.496) for dogs in which partial gastrectomy was the only additional surgical procedure performed (n=23) was not significantly higher than the rate for dogs that did not have a partial gastrectomy. Dogs that had both splenectomy and partial gastrectomy (n=25) had a significantly higher (P<0.0001) postoperative mortality rate of 20% compared to dogs that did not have both splenectomy and partial gastrectomy [Figure 1⇑].

The overall mortality rate of 10% cited in this present study was significantly lower than the overall mortality rates from the data reported in recent previous studies [Table 2⇓].^3,4

Discussion

The overall mortality rate (10%) for the 306 dogs taken to surgery during the 5-year period was improved compared to the overall mortality rates reported in the more recent studies by Brockman et al³ (15%, P=0.001), Beck et al⁵ (16%), and Brourman et al⁴ (18%). Consistent with previous reports, this study found that preoperative cardiac arrhythmias were associated with a higher overall mortality rate, and splenectomy and splenectomy combined with partial gastrectomy were associated with higher postoperative mortality rates. Factors such as breed, age, sex, duration of clinical signs, and partial gastrectomy alone were not associated with increased mortality rates. The breeds treated in this present study are consistent with breeds described previously.¹⁵

Gastric necrosis, splenic injury, or both—treated by partial gastrectomy and/or splenectomy—have historically been associated with postoperative mortality rates ranging from 30% to 68%.^3,4,13,16 The postoperative mortality rates in this present study for partial gastrectomy (9%), splenectomy (15%), and both splenectomy and partial gastrectomy (20%) are significantly lower than those reported by Brourman et al⁴ and Brockman et al,³ where the postoperative mortality rates were never <31% for these conditions [Table 2⇑].

In this study, an increased time from presentation to surgery was associated with a lower mortality rate. This relationship was an unexpected finding, as it is widely accepted that rapid transition to surgery is beneficial in order to minimize the duration of ischemia and degree of organ injury.^6,7 This unexpected finding could be explained by the following hypothesis: dogs with GDV that were bright and alert on presentation may have been triaged and transitioned to surgery more slowly and were given complete fluid resuscitation compared to dogs that were recumbent and depressed at the time of presentation and therefore may have received more rapid transition to surgery without the completed full-fluid resuscitation. Despite these findings, we are not advocating waiting an extreme amount of time from presentation to surgery, as duration of splenic and gastric ischemia has been shown to directly influence organ viability.^8–10 Nonetheless, it is reasonable to suggest that, if possible, complete fluid resuscitation should be achieved before surgical intervention, while still performing surgery in a timely manner. Additional investigations to determine whether all dogs with GDV should be rapidly transitioned to surgery are warranted.

The mean anesthesia time of 67 minutes in this study was lower than the 156 minutes noted in the only previous study to report anesthesia time.³ The length of anesthesia and surgery did not influence the postoperative mortality rate in this study. No previous reports mention surgery time, and the mean time of 48 minutes reported here should be compared in future reports to determine the influence of surgery time on overall and postoperative mortality rates in dogs with GDV.

The postoperative mortality rate was higher for dogs treated with splenectomy than the rate for dogs treated with partial gastrectomy or incisional gastropexy alone. Recent reports evaluated both radiographic findings and serological data as predictors of gastric necrosis.^17–19 The identified higher postoperative mortality rates in this report suggest that development of serum biochemical or other markers of splenic damage may be useful in predicting overall mortality in dogs with GDV.

Because of the retrospective nature of this study, full characterization of the cardiac arrhythmias was not possible. Brourman et al⁴ reported a significantly increased overall mortality rate (38%) when preoperative cardiac arrhythmias were detected, and Brockman et al³ reported no significant effect of cardiac arrhythmias on postoperative mortality rate. Consistent with the study by Brourman et al,⁴ the 11% of dogs in our study with preoperative cardiac arrhythmias had a significantly higher overall mortality rate than dogs without a preoperative cardiac arrhythmia. Interestingly, preoperative intermittent ventricular arrhythmias were associated with a significantly higher overall mortality rate, whereas ventricular tachycardia (usually considered a more serious condition) was not. Postoperative intermittent ventricular arrhythmias were not shown to significantly affect postoperative mortality, but postoperative ventricular tachycardia compared to no arrhythmia and intermittent arrhythmia (combined) did result in a significant increase in postoperative mortality.

The significance and treatment of cardiac arrhythmias in dogs with GDV have been controversial subjects.^1,3–5 Guidelines regarding the point at which treatment should be initiated and the method of treatment for ventricular arrhythmias were established 15 years ago based on consultations with a cardiologist. The established guidelines do not take into account multifocal ectopia or the R-on-T phenomenon that might also indicate a need for treatment.

A statistical limitation in this report involved the low number of overall and postoperative mortality rates. The bottom line is that mortality rates were low in this report, so larger sample sizes would be required to better perform statistical tests comparing mortality rates among groups. The other limitations of this study relate to its retrospective nature and the fact that the records were not complete enough to fully evaluate all parameters of therapy. What we can say is that management of GDV cases is very standardized in our clinic with regard to fluid therapy and pain management, and it reflects what was previously described in the Materials and Methods section. We understand that pre-operative, intraoperative, and postoperative management of a dog with GDV can greatly influence survival, and further investigation into these areas is needed.

Conclusion

The 10% overall mortality rate for dogs with GDV is significantly better than rates reported in recent studies. The postoperative mortality rates for dogs having splenectomy and partial gastrectomy were also lower than rates in previous reports. Splenic and gastric damage continues to be associated with higher postoperative mortality rates, with splenectomy having the greatest influence on postoperative mortality rate. Consistent with some previous studies, preoperative and postoperative cardiac arrhythmias were found to be associated with significantly increased overall and postoperative mortality rates. The hypothesis that time from hospital presentation to surgery and the duration of anesthesia and surgery directly affect overall mortality rates was not supported by the data included in this study. Instead, the time from presentation to surgery was inversely associated with the overall mortality rate. The fact that the anesthesia times were less than half of those previously reported may suggest that anesthesia and surgery times play a role in improved survival. Further research is warranted. In the meantime, these results should not be over interpreted: GDV cases should be treated surgically as soon as reasonably possible.