| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Pediatric Imaging |
1 From the Departments of Pediatric Radiology (S.J., J.C.J., S.H.), Pediatric Surgery (C.L.C.), and Pediatric Gastroenterology (D.C.B.), Children’s Hospital, University Hospital of Geneva, 6 rue Willy Donzé, 1112 Geneva, Switzerland. Received July 23, 2001; revision requested September 24; final revision received May 1, 2002; accepted May 14. Address correspondence to S.J. (e-mail: sigrid.jequier@hcuge.ch).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Thirty-nine children received 42 liver allografts during a 10-year span. Doppler ultrasonographic (US) recordings of hepatic venous wave patterns were reviewed. Nine children (ten grafts) with continuously monophasic flow were not included in the study. Changes from triphasic to monophasic flow were recorded and correlated with clinical findings in all 30 patients and biopsy findings in 25. Biopsy results were compared with US data recorded within 24 hours of biopsy. Standard statistical tests were conducted to assess value of Doppler US in diagnosis of graft rejection. Significance was assessed with 
2 statistics.
RESULTS: Of 113 Doppler US recordings in 30 children, 74 showed an episode of change in flow from triphasic to monophasic in 27 patients; biopsy correlation existed for 39 episodes. Thirty-five episodes were due to acute graft rejection (true-positive results). Thirty-nine episodes were due to a variety of pathologic causes (determined with biopsy results for 12 and by clinical means for 27) (false-positive results). Thirty-six assessments were true-negative (US and biopsy results negative for rejection); three were false-negative. When US results were evaluated against clinical and biopsy data, analysis revealed that change to monophasic flow predicted rejection with sensitivity of 92% (35 of 38) and specificity of 48% (36 of 75). Negative predictive value of evidence of persistent triphasic flow was 92% (36 of 39). In the subgroup of US findings with biopsy correlation, specificity increased from 48% (36 of 75) to 75% (36 of 48). It was zero (0 of 27) for the group with clinical correlation only.
CONCLUSION: Change of hepatic venous flow pattern from triphasic to monophasic is sensitive but nonspecific for detection of graft rejection. Evidence of persistent triphasic flow helps eliminate the possibility of graft rejection with a high negative predictive value.
© RSNA, 2002
Index terms: Blood, flow dynamics, 57.12984 • Hepatic veins, US, 958.12984 • Liver, transplantation • Ultrasound (US), in infants and children
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Acute (cellular) rejection occurs in 50%–70% of cases (3,4), usually in the first 3 weeks after transplantation. It manifests clinically as fever and elevation of liver enzymes and total serum bilirubin. Histopathologically, it is defined by the Snover triad of portal hepatitis with mixed inflammatory infiltrates, portal and central venous endothelialitis, and lymphocytic cholangitis (bile duct injury) (5). This form of rejection is reversible with appropriate treatment (4). Early diagnosis is therefore important.
Clinical symptoms and laboratory tests are nonspecific and cannot enable the distinction of hepatitis from acute allograft rejection. Liver biopsy is often necessary to establish the diagnosis of acute rejection so that treatment with increased doses of cyclosporine and steroids can be initiated. A noninvasive test for acute graft rejection would be extremely helpful to clinicians. Doppler ultrasonography (US) of liver transplant hemodynamics could be useful in this context.
Normal hepatic vein flow usually has a triphasic pattern at Doppler US (Fig 1): The flow proceeds toward the inferior vena cava (IVC) and right side of the heart during cardiac diastole, proceeds back toward the liver during right atrial systole, and comes to a short stop or shows slightly reversed flow during right ventricular systole (6).
|
In a previous study of hepatic venous flow in healthy children, we observed that only 42% had a triphasic flow pattern in all hepatic veins (8). Different flow patterns from one vein to another could be seen within the same liver. A monophasic flow pattern was more often seen in the right hepatic vein and in patients of younger age. No change in the flow pattern of a given vein was seen at examinations conducted with various physiologic conditions in healthy children (9). We concluded that only a change of triphasic flow to monophasic flow in a given hepatic vein should be considered an indicator of possible hepatic parenchymal disease.
In the present study, we applied this criterion in the examination of all children with hepatic orthotopic liver allografts in our institution to evaluate whether a change in liver vein flow pattern can be seen during hepatic graft rejection and if it is reversible with treatment.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
In the study group of the remaining 30 patients (who received a total of 32 allografts), there were 17 girls and 13 boys, ranging in age from 6 months to 14 years and 4 months (mean, 3 years and 10 months). The 32 allografts consisted of 22 split grafts—usually all or part of the left lobe of the liver—and 10 full livers.
The initial preoperative diagnosis was congenital biliary atresia and unsuccessful Kasai procedure in 22 patients, sclerosing cholangitis in two patients, and cryptogenic cirrhosis, Alagille syndrome, tyrosinemia, Crigler-Najjar syndrome, abetalipoproteinemia, and mushroom intoxication in one patient each. A second liver graft was implanted because of primary graft malfunction in one patient and chronic rejection in another.
The study was approved as both a retrospective and prospective study by the hospital ethics committee in November 1998. Informed consent was not deemed necessary by our ethics committee because all studies we evaluated were requested by clinicians for necessary patient care.
Doppler US Studies
Timing.—Postoperative US studies of the abdomen with a hepatic Doppler examination were performed daily and sometimes twice a day or more often if the vascular anastomoses had been particularly difficult for the surgeons to perform and flow was difficult to find during the Doppler US study. If no vascular or other complications were detected during the first 10–14 days after surgery, subsequent Doppler US studies were spaced out progressively, according to the clinical progress of the patient, over days, weeks, and then months, until only a yearly US examination was performed.
Equipment.—A Diasonics machine (Sunnyvale, Calif) with a 5-MHz transducer was used initially, an Acuson XP128 machine (Mountainview, Calif) with transducers of 7.5-, 5.0- and 4.0-MHz was used after 1992, and an Acuson Sequoia machine with transducers of 8.0- to 3.5-MHz was used after 1999. Many patients were scanned with different equipment over time—that is, with an Acuson XP unit in the intensive care unit and with a Sequoia machine in the radiology department.
Technique.—The highest-frequency US probe was used for maximum resolution of gray-scale imaging according to the age and weight of the patient; a 5.0-MHz or 3.5-MHz Doppler US probe was usually used. An anterior approach was used if possible, and a lateral approach was used when necessary because of the presence of surgical dressings or air in the intestine. Angle correction for an angle of 60° was attempted but was not always possible in crying or restless young children; flow velocity was therefore not routinely measured. The velocity scale was set as low as possible, and no filter was added.
Hard copies of the Doppler US images acquired in all patients were obtained, and a retrospective analysis of the Doppler spectral curve in the hepatic veins was performed in the first 26 patients. Doppler US images of the same vein with an insonation angle as similar as possible from one examination to the next were chosen. A prospective study was carried out in the last four patients.
The US examinations were performed by four different pediatric radiologists (S.J., S.H.) and over 20 different residents supervised by one of the four pediatric radiologists.
Interpretation.—The hepatic venous flow patterns in all retrospective and prospective Doppler US studies were analyzed by the same pediatric radiologist (S.J.). Flow was considered to be triphasic when the wave of reversed flow during atrial systole crossed the baseline of the spectral analysis curve. It was considered to be modulated (biphasic) when a wave reversed during atrial systole came close to the baseline without crossing it. A flow pattern was labeled monophasic when it was flat, with no or minimal (<25%) oscillations of the distance between the baseline and peak diastolic flow. The return of monophasic flow to well-modulated biphasic or triphasic flow was considered a return to normal flow.
Chart review.—All charts and all biopsy results were reviewed (S.J.). All episodes of change of hepatic venous flow from triphasic to monophasic were compared with the clinical and biopsy findings recorded at that time. Results from only those 78 biopsies for which correlative Doppler US findings existed (ie, the biopsy was performed within 24 hours before or after a Doppler US examination) were included. Biopsy results were interpreted by a pediatric pathologist.
Additional postoperative complications such as vascular thromboses, bilomas, hematomas, ascites (ie, intraabdominal fluid collections); pleural effusion and other intrathoracic diseases; diaphragmatic paralysis; sepsis; dehydration; peritonitis; pancreatitis; cardiovascular problems; and renal impairment were recorded and compared with Doppler US findings if Doppler US had been performed at the time the complication was observed.
Statistical Analysis
The statistical analyses were performed (J.C.J.) with Statistica software (Statsoft, Tulsa, OK) for Windows (Microsoft, Redmond, Wash). In 113 separate instances, graft rejection was suspected. Consequently, a range of diagnostic procedures was undertaken to confirm or exclude a rejection. These diagnostic procedures included the recording of signs and symptoms, observation of the patient’s clinical course and response to antirejection drug treatments, biochemical investigations, and, in a number of instances, liver biopsy. This sequence of procedures had the purpose of reaching a dichotomous (ie, either positive or negative) end result as to rejection. It represented the independent variable, or standard of diagnosis.
The hepatic US flow pattern was recorded within 24 hours of the diagnostic procedures. Interpretation of the flow patterns was also designed to yield a dichotomous result—that is, the presence or absence of a triphasic flow pattern. This recorded interpretation represented the dependent variable. All sonographers were blinded to the ongoing diagnostic results.
All 113 instances of suspected graft rejection, referred to herein as "episodes," were sufficiently distinct and separated in time to be considered independent from one another, even if they occurred in the same patient. No adjustment for covariance within a patient was provided. The inference of the analytic results was thus based on the episode, rather than the patient, as a statistical unit.
The clinical expression of rejection at its onset can vary widely and remain nonspecific. For that reason, the 113 episodes were divided into (a) those in which a liver biopsy (an invasive procedure) was required for diagnosis and (b) all others. The 78 episodes requiring a liver biopsy, referred to as "biopsy episodes," were generally harder to confirm than the 35 episodes not requiring a biopsy, which were referred to as "clinical episodes." These latter episodes were easier to confirm, either because they were of short duration or because they were acute. Therefore, the subdivision into biopsy episodes and clinical episodes was retained for comparison.
The usual computations of diagnostic predictive values were performed for Doppler US assessment of hepatic venous flow patterns: percent sensitivity as the proportion of true-positive episodes divided by the total number of rejection episodes, specificity as the proportion of true-negative episodes divided by the total number of episodes without rejection, positive predictive value as the proportion of true-positive findings divided by the total number of positive Doppler US findings (of change from triphasic flow to monophasic flow), and negative predictive value as the proportion of true-negative findings divided by the total number of negative Doppler US findings. Positive likelihood was calculated as the ratio of sensitivity to 1 minus specificity, and negative likelihood was calculated as the ratio of 1 minus sensitivity to specificity (10).
Accuracy (sometimes called percentage of agreement) is the percentage of the sum of true-positive and true-negative findings versus the total number of observations. The 
statistic expresses the proportion of agreement observed, taking into account the proportion of agreement that can be expected by chance. results vary between -1, meaning total disagreement, and +1, meaning perfect agreement; zero represents a result obtained by chance only.
The Youden J statistic represents the difference between sensitivity plus specificity minus 1, does not adjust for the total number of measurements, and is interpreted like when it is between -1 and +1. 2 is the classic test for computing the probability of nonhomogeneous partition of nominal data. The Yates correction was performed separately for the three groups (US results with correlation with all information, US results with biopsy correlation only, and US results with clinical correlation only). US results with correlation with all sources of information were evaluated against US results with biopsy correlation only by using the 2 test of conformity. The comparison of results is presented as an overall effect and is decomposed into all four categories of predictive results.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Doppler US Study Results
The use of different equipment by different sonographers had no apparent influence on the observed hepatic venous flow patterns. One hundred thirteen Doppler US studies were performed in 30 patients. Biopsy correlation was available for 78 of the US studies, and clinical correlation was available for 35 of the studies (Table 1). Twenty-seven children had 74 episodes of documented change from triphasic hepatic venous flow to monophasic flow.
|
Three children had triphasic venous hepatic flow at all times and no clinical or biopsy evidence of acute rejection. Results of three liver biopsies performed in two of these three patients were negative for rejection. The hepatic venous flow also remained triphasic in the patient with primary graft malfunction. This patient underwent an emergency second liver transplant procedure 4 days after the first one. These patients are included in the group of true-negative cases.
Four patients had different flow patterns in different hepatic veins within the same liver (in two whole-liver grafts and two split grafts): monophasic in one vein (usually a right branch vein) and triphasic in the others. Only a change in the vein with an initially triphasic flow pattern was taken into account. The use of different equipment did not result in changes in the Doppler wave pattern of any hepatic vein.
There were 35 true-positive Doppler US results. That is, in 35 episodes, a change in the hepatic venous flow pattern corresponded to an incidence of acute rejection (Fig 2). Twenty-seven of these 35 episodes proved to be episodes of acute rejection at liver biopsy; eight proved to be episodes of acute rejection on the basis of the patient’s favorable clinical response to treatment after a clinical and laboratory diagnosis of acute graft rejection had been rendered. Clinical evolution was considered favorable when laboratory values normalized with antirejection treatment. Some patients experienced up to three episodes of acute rejection.
|
|
|
The length of time between the occurrence of monophasic flow and a return to triphasic flow was as short as 1 day (in one instance) and was usually around 3–5 days, but this length could not be determined precisely in many of our patients because most did not undergo daily US examinations beyond the first week after surgery. The one late instance of acute rejection (4 years after transplantation) was first suspected during the patient’s yearly check-up, was confirmed with biopsy, and was treated successfully.
In only two patients with triphasic hepatic vein flow that changed to monophasic flow, no return to a pulsatile wave pattern was observed with treatment. Rejection in these patients became chronic, and a second liver transplant procedure was performed in one (the other was waiting to receive a new liver when this study concluded). Results in these patients are included in the group of true-positive results.
There were 36 true-negative Doppler US results. That is, there were 36 clinical incidences of deterioration of liver function in which the hepatic venous flow pattern remained triphasic and liver biopsy revealed findings other than those indicative of graft rejection. Biopsy in these cases revealed hepatitis (cytomegalovirus hepatitis, Epstein-Barr hepatitis, adenovirus hepatitis, or hepatitis C), nonspecific inflammation, bile duct inflammation, ischemic necrosis, centrilobular necrosis, drug-related cholestasis and steatosis, primary graft malfunction (observed in one instance), improvement of previously diagnosed graft rejection, or minimal changes (revealed at biopsy performed during plication of the right side of the diaphragm in one patient). Normal findings were revealed at biopsy during surgical correction of gastrointestinal bleeding in one patient and at routine biopsies performed in the first few patients. The length of time between transplantation and the biopsy that yielded these "negative" results varied from 4 days to 3 years.
There were 39 false-positive Doppler US results. Of 39 episodes of change in flow pattern from triphasic to monophasic, 12 proved not to represent acute rejection at biopsy, and 27 correlated with no clinical findings of acute rejection. Three of the 12 biopsies revealed ischemic or infectious cholangitis; three, fibrosis; two, centrilobular congestion and/or necrosis; and one each, lymphoproliferative disease (Fig 3), cholestasis, hepatitis, and normal findings.
|
|
|
There were three false-negative Doppler US results. That is, three instances of biopsy-proved graft rejection were missed at Doppler US. Hepatic venous flow pattern was triphasic in one child with findings of mild graft rejection at biopsy performed 2 years and 4 months after transplantation, in one who still showed signs of residual rejection at follow-up biopsy for acute rejection but was improving, and in one who was found to have mild graft rejection at biopsy on day 35 after two episodes of treated graft rejection in which the hepatic venous flow had transiently become monophasic. The acute rejection in these last two patients was improving, but the Doppler US results in these patients were still considered to be false-positive because rejection with endothelialitis was described in the pathology report.
Results of evaluation of the possible effect of other clinical conditions on the hepatic venous flow pattern are as follows: Ascites of variable amounts developed in all children within days of surgery after the surgical drains were withdrawn but had no apparent influence on the hepatic venous flow pattern. Portal vein thrombosis was observed in one child whose hepatic venous flow remained triphasic. Intrathoracic disease—partial atelectasis (usually in the right upper or right lower lobe) and/or pleural effusion (right-sided in 10 children, bilateral in two, and left-sided in one)—occurred in all children except three but did not influence hepatic vein flow patterns.
Paresis of the right side of the diaphragm in 10 children and paralysis of the diaphragm in five children, which necessitated diaphragmatic plication in five cases, did not influence the hepatic venous flow pattern, except in one child whose hepatic venous flow became monophasic for 3 days after diaphragmatic plication but who did not have any other apparent problems.
Peritonitis, septic shock, pancreatitis with a large pseudocyst, dehydration, cardiac disease with an atrial thrombus, and renal insufficiency were, however, observed in association with a change in the hepatic venous flow pattern from triphasic to monophasic (Table 2).
Statistical Results
Statistical analysis of the results of the Doppler US examinations to establish their value in the diagnosis of acute graft rejection was performed on three different levels: Results of Doppler US examinations of the hepatic veins were compared with biopsy findings alone, with clinical findings alone, and with the combination of biopsy and clinical findings in the diagnosis of acute graft rejection. The results are summarized in Table 1 and are illustrated in Figure 4.
|
Doppler US-observed episodes for which biopsy information was the reference standard (78 of 113) provided the best information, because the number of false-positive diagnoses was low and a complete set of data was available for estimation of the diagnostic value of changes in US flow patterns. Specificity, positive predictive value, positive likelihood ratio, percentage of agreement, and and Youden J values were higher for the group of Doppler US-observed episodes with biopsy correlation than for the group of episodes evaluated against all information, including clinical status. The sensitivity of 90% (27 of 30) and specificity of 75% (36 of 48) for the group of findings with biopsy correlation were different from those values yielded by chance alone. The positive predictive value of 69% (27 of 39) is of borderline value for detecting rejection in clinical practice, whereas the negative predictive value of 92% (36 of 39) is useful for excluding acute graft rejection in about nine of 10 instances in which the US flow pattern remains triphasic.
The remaining computations for the group of Doppler US findings for which biopsy information was used as the reference standard confirmed the strength of these statistical findings: Accuracy was 81% (63 of 78), positive likelihood ratio was 3.60, and negative likelihood ratio was 0.133; both ratios were different from 1 (the ratio yielded by pure chance). was 0.615 and Youden J was 0.650 (both different from zero). The 2 value with the Yates correction for one degree of freedom was highly significant at 28.65, with a P value of less than .001, and showed that the distribution of the proportions in the diagnostic categories was not homogeneous.
The third level of information against which Doppler US findings were compared consisted of all findings observed in all 113 episodes (ie, the sum of data gathered with and without biopsies). It represents an overall view of the value of Doppler US as practiced at our hospital independently of the decision to perform a biopsy. When Doppler US findings were compared with this level of information, the diagnostic predictive value of Doppler US was found to be different from that of chance alone.
Sensitivity was 92% (35 of 38), specificity was 48% (36 of 75), positive predictive value was 47% (35 of 74), and negative predictive value was 92% (36 of 39); these values point to the importance of false-positive results. Positive likelihood ratio was 1.77, negative likelihood ratio was 0.164, accuracy was 62.8% (71 of 113), was 0.325, Youden J was 0.401, and 2 with Yates correction was 16.22, with P less than .001 (highly significant). These statistical results show that Doppler US is useful and its results are different from those obtained by pure chance. Doppler US findings of persistent triphasic flow excluded graft rejection with a negative predictive value of 92% (36 of 39).
Finally, a comparison between the four categories of results obtained with all information versus those obtained with biopsy (Fig 4) yielded a 2 test of conformity value of 63.12, with P less than .001 (highly significant), even when the unequal sums of measurements between both groups were taken into account. The result obtained when the unequal sums of measurements were taken into account (2 = 60.75, P < .001) was due almost entirely to the difference in frequencies between the two groups in the category of false-positive findings.
Results of all other subgroups of findings were either identical (ie, between the true-negative and false-negative groups) or not significantly different (2 = 2.37 for the true-positive groups).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The opinions of various authors are divided as to the importance of the various hemodynamic parameters of the hepatic veins in acute rejection. Different researchers have measured different flow parameters: the portal vein and hepatic vein damping indexes, the hepatic artery resistive index, and change in flow velocity (11,12). Portal vein hemodynamics have generally been well studied and are known to vary physiologically with food intake and exercise. Kok et al (11) have shown that there are no alterations in portal venous flow that specifically indicate graft rejection. Hepatic artery perfusion also varies physiologically. Changes in hepatic artery flow and resistive index have proved unhelpful in predicting graft rejection (12,13).
Hepatic vein flow, however, is relatively constant in its patterns; it usually shows a triphasic spectral Doppler analysis wave pattern, which mostly reflects changes in flow related to cardiac activity: forward flow toward the heart during diastole and a reversed wave during right atrial contraction (increased by tricuspid valve insufficiency) and, to a lesser degree, during right ventricular contraction. The flow pattern becomes flat and monophasic in patients with liver cirrhosis and fibrosis (7). Hepatocellular diseases such as hepatitis without fibrosis, intralobular degeneration, and necrosis were not associated with changes in the hepatic vein flow pattern in a study by Colli et al (14).
Previous studies of Doppler US in the assessment of liver graft hemodynamics have been controversial. Harms et al (15) studied adult liver transplant recipients. They found that the resistive index of the hepatic artery was of no help in the diagnosis of graft rejection. However, they did observe a decrease in the damping index (calculated as minimum velocity shift divided by maximum velocity shift) of the portal vein and the hepatic veins during acute rejection; this finding had a sensitivity of 75% and a specificity of 91%.
Zalasin et al (16) found the hepatic vein Doppler wave form pattern to be useless in the diagnosis of graft rejection in adults; they did not, however, perform serial US examinations. To our knowledge, only one study, performed by Britton et al (17), in which hepatic vein wave patterns were evaluated in 41 children with 50 liver transplants followed up for an average of 36 days after surgery, recorded positive results. Britton et al (17) recommend that serial Doppler US examinations be performed in such patients and evidence of changes in flow pattern be sought. Their study showed that de novo reduced pulsatility of the hepatic veins indicated acute rejection with a sensitivity of 92% and a specificity of 100%. They observed no patient who had a triphasic flow pattern and acute graft rejection. The change in flow pattern was thought to occur as the result of a decrease in liver compliance because of lymphocytic infiltration during rejection.
We were unaware of the results of the study of Britton et al (17) when we began retrospectively and prospectively to evaluate a change in the hepatic vein flow pattern as an indicator of graft rejection in a similar population of children with liver transplantation. Our study was conducted not for a short postoperative period only, but rather over a prolonged time of up to 10 years; to our knowledge, a study of this duration had not previously been conducted. In a search of the literature, we did not find any study of liver graft hemodynamics conducted years after transplantation. In agreement with published data of acute rejection, a majority of changes in our patients were observed during the first 2 months after transplantation, but some changes occurred much later and some were recurrent. No change in flow pattern due to rejection and no instance of rejection occurred later than 4 years after surgery.
In addition to acute graft rejection, there are other postoperative hazards that can change the hemodynamics of a grafted liver. They occur mostly during the first few days after surgery: portal vein thrombosis in 7.6% of patients and hepatic artery thrombosis in 16% of patients (18). IVC thrombosis has been observed to occur after transplantation in 7% of patients in our hospital. Fluid collections in and around the transplanted liver are seen in up to 42% of patients (19). These may represent hematomas or bilomas and could lead to compression of the liver graft and IVC and changes in the hemodynamics of the hepatic veins. Because the triphasic hepatic vein flow pattern reflects cardiac activity, it seems reasonable to postulate that any generalized cardiovascular event such as septic shock, dehydration, or renal insufficiency would also influence the hepatic venous flow pattern. Experimental studies would be needed for an evaluation of the precise effect of IVC compression and other factors on the hepatic venous flow pattern.
Our study documents a change in the hepatic venous flow pattern in a large variety of clinical conditions such as IVC thrombosis, IVC compression by large intrahepatic or perihepatic fluid collections, Budd-Chiari syndrome, systemic vascular collapse, and renal insufficiency. A shortcoming of this study was that we had no formal proof of the influence of these pathologic clinical conditions on the hepatic venous flow pattern. The only reason to believe that these factors were responsible for the change in the flow pattern was the return to a triphasic flow pattern after correction of these factors.
The large number of false-positive findings (n = 39) in the "all information" statistical group in our study resulted from the inclusion of the large group of clinical data (27) with the much smaller group of biopsy data (12). The comparative lack of correlative biopsy findings was due to the fact that the pathologic nature of many of the clinical conditions we observed to be associated with a change in the hepatic venous flow pattern did not require a liver biopsy for diagnosis (eg, renal insufficiency, dehydration).
Conversely, most true-negative Doppler US results were associated with an uncertain clinical status in which biopsy was called for so that possible deterioration of the liver could be evaluated.
Other limitations of this study were its purely clinical and largely retrospective nature, the small number of patients, and a lack of biopsy confirmation of eight cases of acute rejection.
The large percentage of false-positive episodes with change from triphasic hepatic venous flow to monophasic flow in the absence of graft rejection is sufficient for us to conclude that acute graft rejection cannot be reliably diagnosed by means of observation of a change in the hepatic venous flow pattern alone.
However, we believe that our study has relevance for clinical practice. Our results show that the value of Doppler US in the diagnosis of acute graft rejection is different from what could be expected with chance alone: The sum of the sensitivity and specificity values was higher than 100%. The positive likelihood ratio (1.77) was higher than 1, and the negative likelihood ratio (0.164) was smaller than 1; the and the Youden J values were different from zero.
A change in hepatic venous flow pattern from triphasic to monophasic is sensitive but nonspecific for the diagnosis of acute graft rejection.
When graft function deteriorates and the hepatic venous flow pattern remains triphasic, evidence of liver disease other than graft rejection should be sought, and a biopsy may be needed. Evidence of persistence of a triphasic hepatic venous flow helps eliminate the possibility of graft rejection with a high negative predictive value.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, S.J.; study concepts and design, S.J.; literature research, S.J.; clinical studies, S.H., C.L.C., D.C.B.; data acquisition, S.H., S.J., C.L.C.; data analysis/interpretation, J.C.J.; statistical analysis, J.C.J.; manuscript preparation, definition of intellectual content, and editing, S.J.; manuscript revision/review and final version approval, D.C.B.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Bolognesi, D. Sacerdoti, C. Mescoli, V. Nava, G. Bombonato, C. Merkel, R. Merenda, P. Angeli, M. Rugge, and A. Gatta Acute Liver Rejection: Accuracy and Predictive Values of Doppler US Measurements--Initial Experience Radiology, May 1, 2005; 235(2): 651 - 658. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
