In congenital portosystemic shunts (CPSS), portal blood is shunted around the liver and directly into the systemic circulation, resulting in macroscopic and microscopic liver hypoplasia.1–3 In many dogs with CPSS, hepatic function is completely restored after surgical attenuation of the shunt. However, portosystemic shunting, with or without clinical disease, persists in 10–20% of dogs, regardless of the surgical technique used for shunt closure.4–6 The ability of the liver to adapt to the increased blood flow after shunt attenuation and to grow to normal size may contribute to recovery in individual patients.4,5,7 In dogs with CPSS, diagnostic imaging techniques frequently are used for pre- and postoperative assessment of the anatomy of the shunt and development of the liver and portal vein.8–11 Liver volume has been measured only in a small number of dogs before and after attenuation of the shunt. Although liver volume may not be small in all dogs with CPSS, it was suggested as a prognostic marker of hepatic function and a noninvasive method to evaluate response to therapy.12,13 In humans, liver volume is significantly related to prognosis in hepatic diseases such as cirrhosis and fulminant liver failure.14,15 Therefore, the aims of the present study were to describe 2 noninvasive methods to measure liver volume in vivo and to record liver growth in dogs with extrahepatic CPSS after surgical attenuation of the shunt.
Background: In dogs with congenital portosystemic shunts (CPSS), the ability of the hypoplastic liver to grow is considered important for recovery after surgical shunt attenuation.
Objectives: This study investigated hepatic growth after extrahepatic shunt attenuation in dogs using magnetic resonance imaging (MRI) and computed tomography (CT).
Animals: Ten client-owned dogs with single extrahepatic CPSS.
Methods: Abdominal MRI, CT, or both were performed before and 8 days, 1, and 2 months after shunt attenuation. Liver volumes were calculated from the areas of the MRI or CT images.
Results: Before surgery, median liver volume was 18.2 cm3/kg body weight. Liver volume increased significantly after surgery. Growth was highest between days 0 and 8 and decreased afterward. Median liver volume was 28.8 cm3/kg at 2 months after attenuation. No significant differences in growth were found between dogs with complete or partial shunt closure or between dogs with complete or incomplete metabolic recovery. Volumes measured from consecutively performed MRI and CT images correlated well (r= 0.980), but volumes from MRI images were significantly larger than volumes from CT images (6.8%; P= .008).
Conclusion and Clinical Importance: After shunt attenuation, rapid normalization of liver size was observed. Hepatic growth was not decreased in dogs after partial closure of CPSS or in dogs with subclinical, persistent shunting 2 months after surgery. CT is the preferred imaging method for volumetric estimation because of speed.
congenital portosystemic shunt
magnetic resonance imaging
Material and Methods
Dogs in which a single extrahepatic CPSS was diagnosed and that were consecutively planned for surgical shunt ligation were entered into the study, with informed written consent of the owners. The design of the study was approved by the Ethics Committee on Animal Experimentation. The diagnosis of portosystemic shunting was based on increased plasma concentrations of bile acids and ammonia (NH3) after 12 hours of fasting (reference values: bile acids, 0–10 μM; NH3, 24–45 μM) or an abnormal rectal NH3 tolerance test.16 The presence of a single extrahepatic CPSS was confirmed by ultrasonography. All surgeries were performed by the same surgeon. After exploration of the abdominal cavity via a midline celiotomy, the extrahepatic shunt was ligated over a gauged rod to the smallest diameter that did not induce portal hypertension, using a nonabsorbable 2-0 polyester suture.a All dogs were kept on 1 commercial low protein diet during the study.
Clinical and metabolic recovery was monitored at 8 days, 1, and 2 months after surgery. At each visit, body weight was measured and portosystemic shunting was evaluated by determining 12-hour fasting plasma bile acids and NH3 concentration or performing a rectal NH3 tolerance test.16 Complete recovery was defined by resolution of all clinical signs (clinical recovery) and normal results of the rectal NH3 tolerance test (metabolic recovery).
Magnetic resonance imaging (MRI) and computed tomography (CT) were performed before surgery and at 8 days, 1, and 2 months after surgery in anesthetized dogs, positioned in dorsal recumbency. T1-weighted images (time of repetition 560 ms, time of echo 15.0 ms) of 5-mm-thick contiguous slices were made of the entire liver using a 0.2 T open MRI scanner.b CT images were made of the entire abdomen with a single slice helical CT scanner,c using 120 kVp and 280 mA settings, a collimation of 3 mm, a pitch of 1, and 0.7 seconds scanning time per rotation. Images were reconstructed to 2-mm-thick contiguous slices. CT scans were made during a state of apnea, achieved by hyperventilating or disabling intermittent positive pressure ventilation.
MRI images were viewed with a window width of 1,892 and a window center of 891; CT images were viewed with a window width of 150 and a window level of 40 HU. Liver contours were outlined using a mouse-driven cursor on each individual image of the MRI and CT scans, excluding the gallbladder, the caudal vena cava, and the portal vein, where these vessels were not completely surrounded by liver tissue (Fig 1). Liver surface areas were calculated on every image using the standard software programsd,e of the scanners. The sum of the areas of all slices per scan was multiplied with slice thickness to calculate total liver volume. Measurements were performed twice, blinded, and in random order by one of the authors.
All statistical tests were performed using commercial software.f Wilcoxon's signed ranks tests were used to compare 1st and 2nd measurements of liver volumes, volumes calculated from simultaneous MRI and CT scans, hepatic growth, and changes in body weight. Correlations between volume measurements and imaging techniques were analyzed by linear regression. Comparisons of liver volumes with degree of intraoperative CPSS closure and postoperative recovery were performed using Mann-Whitney U-tests. Spearman's ρ correlation coefficients were calculated to determine correlations among age at surgery, body weight, and liver volumes. A P value < .05 was considered significant.
Twelve small breed dogs were entered into the study, but 2 dogs did not survive to the end of the study. One dog suddenly died 10 hours after uncomplicated surgical shunt closure. The cause of death was unknown. Another dog died 2 days after surgery because of persistent abdominal bleeding because of coagulopathy. The group with follow-up measurements consisted of 10 dogs, 7 female and 3 male dogs. Body weight at time of surgery ranged from 1.2 to 6.5 kg, with a median of 5.9 kg. Age ranged from 5.3 to 40 months (median 11 months). The group consisted of 7 Terrier breed dogs, 1 Shih Tzu, and 2 mixed breed dogs. Before surgery, median plasma bile acid concentrations were 175 μM (range, 19–260 μM). During surgery the attenuated CPSS consisted of 6 portocaval shunts and 4 portoazygos shunts. All portocaval shunts and 2 portoazygos shunts were partially ligated, whereas 2 portoazygos shunts were completely ligated (dogs 2 and 6). All 10 dogs did clinically well immediately after surgery and returned to the clinic at 8 days after surgery. At 8 days after surgery, diarrhea and vomiting were observed in dog 1, and in dog 7 mild diarrhea was seen. No clinical signs were reported in the other 8 dogs. Dogs were returned for evaluation approximately 1 and 2 months after surgery: median values were 29 and 64 days after surgery, respectively. Clinical recovery was good in all dogs: signs of hepatic encephalopathy or other clinical signs of portosystemic shunting were not reported. Two dogs had persistent portosystemic shunting (dogs 9 and 10). Despite good clinical response in both dogs, NH3 tolerance tests were still abnormal 2 months after surgery. At 20 and 40 minutes after rectal NH3 administration, plasma NH3 concentrations were >286 and 137 μM in dog 9 and >286 and 150 μM in dog 10 (280 μM being the upper detection limit, and 46 μM the upper fasting reference value). Plasma bile acid concentrations in these dogs were 3.0 μM (dog 9) and 59 μM (dog 10; fasting reference value <10 μM). In the 8 recovered dogs, median plasma NH3 concentration was 7.5 μM (range, <7.0–24 μM) with normal results for NH3 tolerance testing, and median bile acid concentration was 2.5 μM (range, 0–13 μM) 2 months after surgery. A significant increase in body weight occurred between 8 days and 1 month after surgery (P= .005) and between 1 and 2 months after surgery (P= .011).
In the 1st 2 dogs, only MRI scans were performed. In dogs 3–5, both MRI and CT scans were performed and in dogs 6–10, only CT scans were performed to estimate liver volume. Two consecutively planned scans in 1 dog were not performed (1 MRI, 1 CT) and 4 CT examinations were lost because of technical problems. A total of 46 volumetric scans were made and analyzed. The results of the volumetric measurements were normally distributed. There were no significant differences between duplicate measurements made on CT images or between duplicate measurements made on MRI images. Measurements made on MRI images were compared with consecutive measurements made on CT images in dogs 3–5 (11 CT scans and 11 MRI scans). These values correlated very well (r= 0.980, Fig 2), but a significant difference was found (P= .008). On average, volumes estimated from CT images were 6.8% less than volumes estimated from subsequent MRI images.
For evaluation of liver growth, only values derived from CT measurements were used. In both dogs in which no CT scans were made (dogs 1 and 2), the MRI-derived measurements were corrected to fit CT data using a factor of 0.932. Box-and-whisker plots of liver volume and body weight after surgical attenuation of the shunt are shown in Figure 3. Liver volumes and growth were also expressed in cm3/kg body weight. The median liver volume before surgery was 18.2 cm3/kg (range, 16.4–32.7 cm3/kg). The largest hepatic growth occurred in the period between days 0 and 8. Median hepatic volumes significantly increased in this period by 47.9 cm3 (P= .018) or 8.6 cm3/kg (P= .018) to 26.8 cm3/kg. Between 8 days and 1 month postoperatively, median hepatic growth was of 12.4 cm3 (P= .043). During this period, there was no significant increase in liver volume per kilogram body weight. The gain in liver volume per kilogram body weight was maintained at 1 and 2 months after surgery: mean volumes at 1 and 2 months after surgery were 28.5 and 28.8 cm3/kg, respectively, which were significantly larger than before surgery (both P values were .008). Between 1 and 2 months after surgery, there was no significant liver growth. After 2 months, liver volume increased to 152% compared with preoperative values (146% relative to body weight).
Hepatic growth in the 2 dogs that had residual portosystemic shunting (dogs 9 and 10) was not significantly different from the 8 dogs with complete metabolic recovery from portosystemic shunting at 8 days, 1, or 2 months after surgery. Individual measurements actually showed an increase of liver volume beyond the median value in these dogs (177% and 156% relative to body weight, respectively).
We found no significant differences between 2 dogs that had their shunts completely closed (dogs 2 and 6) and 8 dogs whose shunts were only partially closed at 8 days, 1, or 2 months after surgery.
Although before surgery no significant correlation existed between age and liver volume per kilogram body weight, this correlation became significant 1 month after surgery. Spearman's ρ correlation coefficients were 0.95 at 1 month (P < .001) and 0.85 at 2 months postoperatively (P= .004). The youngest dogs had the largest relative liver volumes after surgery. No significant correlation was found between hepatic growth and increase in body weight.
Performing MRI and CT in dogs with CPSS, allowed clinically relevant results to be obtained with respect to liver volume before and after shunt attenuation. Hepatic volumes estimated from consecutively performed MRI and CT images correlated well, but volumes from MRI scans were significantly larger than volumes from CT scans. Before surgery, median liver volume was 18.2 cm3/kg body weight. Liver growth was highest between days 0 and 8 and decreased afterward. Liver volume per kilogram body weight increased by 46% in 2 months after shunt attenuation. No significant differences in growth were found between dogs with or without complete shunt closure or between dogs with and without complete metabolic recovery.
Liver volume measurement has been reported in dogs with CPSS using radiography, ultrasonography, and CT.12,13,17 On abdominal radiographs, superposition of abdominal organs can make identification of the caudal border of the liver impossible.18 However, radiographs were used to compare liver size in small dogs with CPSS and healthy dogs of the same breeds. The hepatic area in dogs with CPSS was reported to be 48% smaller than in healthy dogs.12 Despite differences in imaging techniques and dogs, we found a similar difference in liver volume between small-sized dogs with a CPSS before surgery and after their recovery 2 months postoperatively (46%).
Liver volume estimation with abdominal helical CT has been reported to be very reliable. It has been regarded as the most accurate imaging technique for in vivo evaluation of organ volume and is considered the gold standard.19,20 Volume measurements with CT have been used in large numbers of human patients with liver disease, because CT is easy, safe, and rapid.15,21 CT also is capable of rapid and accurate measurements of liver volume and growth in small animals such as dogs.13,22 CT volume measurements are reported to be accurate for organs larger than 10 mL.20 The smallest liver volume in our patients was 23.8 cm3 in a dog that weighed 1.2 kg (dog 9, day 0). Thus, liver volumes <10 mL will not often be encountered in dogs.
MRI also offers an accurate means of determining liver volume in vivo.23 MRI has been used in dogs with CPSS for diagnostic purposes or to evaluate hepatic encephalopathy-related brain changes before and after surgical attenuation, but liver volume measurements in dogs using this technique have not been reported before.24–26 Compared with CT, measured MRI volumes were 6.8% larger. Differences in slice thickness (MRI 5 mm, CT 3 mm) could have contributed to less accurate measurements on MRI scans, but the most likely cause of differences between MRI and CT-derived liver measurements is respiratory movement of the liver during the MRI scans.27 During CT scanning, dogs were held in apnea, which is not possible during MRI scanning because of acquisition time. Because respiratory movement during MRI can be responsible for overestimation of liver volumes and because CT is fast, CT is preferred for volumetric measurements of the liver. CT and MRI were not consistently used in all dogs. In the 1st 2 dogs, the CT scanner was temporarily not available. In the next 5 dogs, CT and MRI were used to compare both techniques. This was abandoned because performing both imaging techniques before surgery resulted in prolonged anesthesia time and severe hypothermia. The resulting increase of surgical risk was considered unacceptable. Therefore, volumetric studies in the last 5 dogs only used CT scanning.
In this study, larger liver volumes were found in dogs with CPSS before and after surgery than reported previously. Stieger et al13 measured a mean preoperative liver volume of 15.5 cm3/kg in 21 CPSS dogs and postoperative volumes of 11.7 cm3/kg in 2 intrahepatic CPSS dogs and 22.8 cm3/kg in an extrahepatic CPSS dog, using quantitative CT. The median liver volume 2 months after surgery (28.8 cm3/kg) in this study also is higher than the reported volumes in normal adult dogs or humans. The average liver volume in normal dogs was 24.4 cm3/kg and normal human adult liver size is 23.5 cm3/kg.13,20 In young children, up to around 8 years old, relative liver volume is much larger than in adults. Mean liver volumes in children range from 34.1 cm3/kg in infants to 23.8 cm3/kg at an age of 7 years.28 Although postoperative liver size in this study seems more comparable with normal liver size in children, the median age in our dogs was 11 months, which is considered young adulthood in small breed dogs. However, the median age in our dogs was less than the age of the other reported extrahepatic CPSS dog that was measured after shunt ligation and the dogs that were used to determine normal liver volumes.13 Furthermore, the significant correlation that was found in this study between age and proportional liver volumes after recovery is compatible with the age-related difference in liver volume reported in humans.28 Besides age, different estimations of liver volumes may be a result of differences in measurement technique or breeds that were used. Although no differences were reported between dogs with intrahepatic and extrahepatic shunts before surgery, additional studies are necessary to determine shunt specific differences with respect to liver volume and growth after surgery.13
Regeneration of normal liver is very fast; liver volume is restored in 7 days after 2/3 hepatectomy in rats and mice.29 The increase in liver volume in this study also was remarkably rapid during the 1st week after surgery. Body weight directly influences the results of liver volumes expressed per kilogram. In dogs with CPSS, retarded growth or leanness is a common finding, so after successful treatment both young and adult patients are expected to gain body weight. Although average body weight did not significantly change in the 1st week after surgery, in several dogs body weight had decreased at 8 days after surgery. In 2 dogs, this weight loss may have been caused by vomiting and diarrhea. Weight loss also can be related to the high plasma bile acid concentrations before surgery and the substantial decrease in bile acid concentrations 8 days after surgery in most dogs. Bile acids are known to inhibit 11β-hydroxysteroid dehydrogenase type 2, an enzyme that is responsible for protecting the mineralocorticoid receptors from activation. Activation of these receptors in patients with high plasma bile acid concentrations leads to renal sodium retention and therefore retention of fluid.30,31 After normalization of bile acid concentrations, fluid retention is reversed and body weight may decrease. Decreased body weight increases relative liver volume. However, perfusion state has a major impact on liver volume. So, if body weight is decreased because of reversed fluid retention, the absolute liver volume (in cm3) is expected to be lower too. Body weight did increase significantly between 8 days and 1 month after surgery (P= .005), which is not surprising in dogs that are still young and growing, but also may be expected as a compensating gain of weight in older dogs that recover from secondary retarded growth and leanness. Although absolute liver volumes further increased during this period, the gain of weight camouflaged hepatic growth expressed in cm3/kg.
Stieger et al13 suggested that dogs with smaller hepatic liver volume before surgery have poorer outcome. He proposed repeated CT postoperatively as a useful measure of response to therapy. We found no significant differences in preoperative liver volumes or liver growth between dogs with complete recovery and dogs with persistent shunting. Recently, Doran et al17 reported that dogs that were tolerant to full closure of the shunt during surgery had significantly greater preoperative liver volume/body weight ratios than dogs intolerant of occlusion. We did not find any significant differences between preoperative or postoperative volumetric results of dogs with completely closed shunts and dogs with shunts closed to the smallest diameter possible. The number of cases may be too small to find slight differences between dogs with complete and semicomplete recovery or between dogs with complete and those with partial closure. Measurements in a large number of dogs, including those with incomplete resolution of clinical signs, may help to identify such differences.
Recovery rates, and possibly also hepatic size or growth, vary among dogs after attenuation of a CPSS. This may be caused by differences in shunt localization, breed, and population.32,33 Also the surgical technique that is used for CPSS closure may affect recovery and hepatic growth. This study was restricted to dogs with extrahepatic CPSS referred to 1 university clinic, using an identical ligation technique in all dogs. Although the concepts of this study may be equally valid to other CPSS dogs, results of hepatic volume and growth measurements may yield different results in different centers because of local variation of the above factors.
In conclusion, CT and MRI are both suitable noninvasive methods to measure liver volumes in dogs, but CT is preferred because of its speed. Liver volumes in dogs with extrahepatic CPSS are smaller than reported volumes in normal dogs. After attenuation of the shunt, a rapid increase of liver volume occurred within 8 days and complete normalization was measured in dogs with complete recovery of portosystemic shunting, and also in dogs with subclinical postoperative shunting. Partial closure of CPSS was sufficient to obtain complete recovery and normalization of liver volume in 6/8 dogs.
aEthibond excel 2-0, Ethicon Inc, Somerville, NJ
bMagnetom Open Viva scanner (0.2 T), Siemens Nederland N.V. Medical Solutions, The Hague, the Netherlands
cHelical CT Secura scanner, Philips Medical Systems Nederland BV, Best, the Netherlands
dNumaris version VB33G, Siemens Nederland N.V. Medical Solutions
eEasyVision release 5.1, Philips Medical Systems Nederland BV
fSPSS for Windows, release 16.0.1, SPSS Inc, Chicago, IL
The authors thank all collaborators of the Divisions of Diagnostic Imaging and Anesthesiology for their support.