Objectives: In this study, we share our approach for care of patients with hepatic venous outlet obstruction after living-donor liver transplant.
Materials and Methods: We retrospectively examined the demographic, clinical, and radiologic data of 35 patients who developed hepatic venous outlet obstruction after living-donor liver transplant. Patients were subgrouped on the basis of onset (8 patients with early onset [< 30 days posttransplant] and 27 patients with late onset [≥ 30 days posttransplant]) and postoperative survival (24 survivors, 11 nonsurvivors).
Results: Patients ranged in age from 1 to 61 years (24 adults and 11 children). All adult patients had undergone right lobe living-donor liver transplant. In the pediatric group, 8 had undergone left lateral segment and 3 had undergone left lobe living-donor liver transplant. Nineteen adult patients and all 11 pediatric patients underwent hepatic venous reconstruction, with all procedures based on common large-opening drainage models using various vascular graft materials. Development of hepatic venous outlet obstruction occurred at mean posttransplant day 233 ± 298.5 in the adult patients and mean post-transplant day 139 ± 97.8 in the pediatric patients. After development of obstruction, the patients underwent 1-6 sessions (1.5 ± 1.1 sessions) of balloon angioplasty. After the first balloon angioplasty procedure, 25% of the adults and 36.3% of the pediatric patients developed recurrence. The early-onset and late-onset subgroups showed statistically significant differences in serum albumin (P = .01), underlying causes (P < .001), time from transplant to obstruction (P = .02), and time from transplant to last visit (P = .02). The survivor and nonsurvivor subgroups showed statistically significant differences in total bilirubin (P = .03) and time from transplant to last visit (P = .03).
Conclusions: Common large-opening reconstruction minimizes hepatic venous outlet obstruction devel-opment after living-donor liver transplant. Balloon angioplasty and/or stenting is almost always the first option in the care of this complication.
Key words : Balloon angioplasty, Venous outflow, Venous reconstruction models
Since the time of the first successful liver transplant in humans (performed by Starzl and colleagues in 1967), liver transplant has emerged as the standard treatment option for a wide array of liver disorders and mainly those related to chronic liver diseases.1-4 Although deceased-donor liver transplant (DDLT) constitutes a large percentage of liver transplant procedures performed in Western countries, most liver transplant procedures performed in many of the Asian countries (including Turkey) use organs obtained from living donors (that is, living-donor liver transplant; LDLT).1 The surgical techniques of DDLT and LDLT differ, as do the posttransplant liver functions resulting from each.
One of the most notable differences in the techniques is the reconstruction approach for venous outflow, which is necessary to ensure proper and adequate hepatic drainage.3 In DDLT, various wide anastomoses are reconstructed between the donor’s inferior vena cava (IVC) and the recipient’s retrohepatic IVC; this approach lowers the rate of postoperative drainage complications to acceptable levels. In contrast, for LDLT, the reconstruction approach is complicated by the presence of multiple venous orifices in the liver grafts, as well as the short hepatic vein and small diameters of veins, all of which can also lead to difficulties with postoperative drainage.3
Various common large opening venous recon-struction models have been developed to minimize the risk of developing hepatic venous outflow obstruction (HVOO) after LDLT.2 The foundation of these common large opening venous reconstruction models was developed by Lee and colleagues (ASAN Medical Center, Seoul, South Korea), based on their vast experience in LDLT. Our hospital has developed, and successfully implemented, various drainage models inspired by the original drainage model of Lee and colleagues.
Materials and Methods
The demographic, clinical, biochemical, and radiologic characteristics of 1011 patients who underwent LDLT at the Inonu University Liver Transplantation Institute (Malatya, Turkey) between November 2007 and April 2014 were retrospectively reviewed. Posttransplant HVOO had developed in 35 patients (3.46%), as discovered during follow-up. The following parameters were evaluated for the 35 patients with HVOO: age, sex, Model for End-Stage Liver Disease score or Pediatric End-Stage Liver Disease score, Child-Pugh score, graft-to-recipient weight ratio, graft type (ie, right lobe, left lobe, or left lobe lateral segment), aspartate aminotransferase level (in U/L), alanine aminotransferase level (in U/L), international normalized ratio, total bilirubin (in mg/dL), albumin (in g/dL), time from liver transplant to development of HVOO, time from liver transplant to last outpatient clinic visit/mortality, time from balloon angioplasty to last outpatient clinic visit/mortality, number of balloon angioplasty procedures, and clinical signs of HVOO (ie, emergence of ascites, pleural effusion, and lower extremity edema). In addition, data regarding the various vascular reconstruction approaches used (ie, common large opening, fence conduit, all-in-one), all of which were performed in the liver graft at the back-table stage prior to implantation, were also examined.
The primary aim of this study was to gain insight into how our Liver Transplant Institute has managed to attain a low HVOO rate (3.46%) after LDLT and to report the specifications of our hepatic venous reconstruction model so that they may be assessed and applied by our colleagues in clinical practice worldwide. The secondary aim of this study was to determine whether early-onset or late-onset HVOO affects patient survival. We therefore divided the 35 HVOO patients into 2 subgroups based on timing of HVOO: 8 patients with early-onset HVOO (occurrence within the first 30 days posttransplant) and 27 patients with late-onset HVOO (occurrence beyond day 30 posttransplant). A tertiary aim of this study was to determine whether any significant differences existed between the clinical characteristics of the patients with HVOO who survived and those who did not; for this, we subgrouped patients according to survival (24 patients) and death (11 patients).
Hepatic venous reconstruction on the back table
To maximize benefits to patients who receive liver grafts, all venous structures (V5, V8, inferior right hepatic vein) with a diameter of ≥ 5 mm are included in the drainage model. For the reconstruction technique involving the right liver graft, first V5 and V8 and then the inferior right hepatic veins, if present, are extended to the right hepatic vein using allogenic (ie, saphenous vein or iliac artery) or artificial (ie, polytetrafluoroethylene) graft materials (Figure 1B and 1C). Next, the saphenous graft is opened longitudinally and wrapped around the venous orifices, which are pieced together in a fence-like manner (Figure 1D and 1E). The procedure involving the extension of all venous structures on the graft to the right vein and their reconstruction in the form of a single orifice is termed the “all-in-one” venous reconstruction technique (Figure 1F and 1G). The final stage of the procedure (the process of wrapping the saphenous vein like a fence) is called “circumferential fencing” or creating a “fence conduit” or a “common large opening.” In cases when the right inferior hepatic veins and V5 orifices are too large and > 5 cm away from the right hepatic vein orifice, we directly anastomose to the IVC (Figure 2). The wide fence conduit is created by wrapping the saphenous vein graft around the left hepatic vein orifice for left lobe lateral segment grafts (Figure 3). At the back-table stage, we reconstruct the recipient’s IVC before implanting the venous reconstructed graft. A vein portion equal to the longest diameter of the wide orifice reconstructed at the back table is removed from the anterolateral aspect of the IVC (from the hepatic vein orifice, caudally). Then, a wide anastomosis is established between the IVC and the graft’s vein.
Diagnosis of hepatic venous outflow obstruction
Doppler ultrasonography evaluation was ordered for any post-LDLT patient who developed the following signs and symptoms during follow-up: impaired liver function tests, jaundice, abdominal swelling, extremity edema, and respiratory difficulty. Patients who demonstrated slow flow (velocity of < 10 cm/s) or persistent monophasic wave were further evaluated with multidetector computed tomography (CT). Patients who showed a nonenhanced hepatic vein (> 50% stenosis relative to the normal segment in the stenotic segment) or areas of low attenuation in liver on multidetector CT were sent for further examination via hepatic venography to confirm the diagnosis and to apply a therapeutic procedure, if necessary. In particular, patients who showed contrast stasis due to stenosed anastomosis on the venography test were considered as having severe HVOO, and appropriate therapeutic interventions were applied. All patients, following the HVOO diagnosis, were monitored radiologically (via Doppler ultrasonography and/or multidetector CT) and clinically (for evidence of regression of ascites fluid and weight loss). When clinical status improved, the patients were placed on a once-monthly follow-up schedule. Patients with lagging recovery, according to either clinical or radiologic findings, received repeat balloon angioplasty and/or stenting procedures. All patients, regardless of recovery speed or status, received a short-term course (1 mo) of low-molecular-weight heparin during the postoperative period and an indefinite course of acetyl salicylic acid thereafter.
Statistical analyses were performed with the SPSS Statistics software package (SPSS: An IBM Company, version 22, IBM Corporation, Armonk, NY, USA). Data are presented as means ± standard deviation (SD). The independent sample t test was applied for comparison of continuous variables, and the Pearson chi-square test was applied for comparison of categorical variables. The Kaplan-Meier estimate test was utilized to determine patient cumulative survival rates after angioplasty. P < .05 was considered statistically significant for all statistical analyses.
This study retrospectively reviewed the demo-graphic and clinical information of 35 patients with HVOO, including 24 adult patients and 11 pediatric patients, ranging in age from 1 to 61 years. All adult patients received a right lobe graft, whereas the pediatric patients received either a left lobe lateral segment graft (8 patients) or a left lobe graft (3 patients). Mean time to development of HVOO (days posttransplant) was 233 ± 298.5 (range, 14-1440) for the adult patients and 139 ± 97.8 (range, 15-300) for the pediatric patients. For both adult and pediatric patients, the most common signs and symptoms of HVOO were abdominal pain, abdominal swelling, respiratory difficulty, dermal and scleral icterus, and peripheral edema. All adult patients presented with moderate ascites and pleural effusion upon HVOO development; these symptoms were present in only 82% of the pediatric patients. In total, only 5 patients developed marked lower extremity edema (4 adults and 1 child). The general clinical and HVOO-related characteristics of all patients (adult and pediatric) are summarized in Table 1.
Two of the 3 pediatric patients who received the left lobe graft had the vein draining segment 1 included in the reconstruction model. Circumferential fencing was performed in all 8 of the pediatric patients who received a left lobe lateral segment graft. For the adult patients who received the right lobe graft, the all-in-one model was used for venous drainage. In 7 of the adult patients, V5 and/or V8 were extended to the right hepatic vein with the help of vascular grafts, after which a common large opening venous drainage model was reconstructed using the circumferential fencing technique. In these patients, the orifices of the right inferior hepatic veins that were > 5 mm were wrapped with saphenous vein grafts and anastomosed to the IVC. For the remaining 5 adult patients, no graft material was used for the venous reconstruction.
Hepatic venous outflow obstruction treatment by balloon angioplasty was carried out 1 to 6 times (means ± standard deviation of 1.5 ± 1.1) for each patient. Monitoring after the first balloon angioplasty session was carried out for an average of 1193 days (range, 51-2310 days) in the adult patients. Eighteen of the adult patients did not develop any restenosis after the first angioplasty procedure; however, 6 patients (25%) developed restenosis within 1 to 3 months (Figure 4). For 2 of the 6 patients who developed restenosis, a second balloon angioplasty procedure was performed and an expandable metallic stent was implanted. The other 4 patients underwent balloon angioplasty without stenting. Twelve of the 18 adult patients who did not develop restenosis at any time during the long-term monitoring period experienced complete recovery. Unfortunately, the remaining 6 patients stopped attending the monitoring appointments, and the outcome of the angioplasty procedure cannot be definitely ascertained. However, each of these 6 patients later presented to our clinic with impaired general status. Two of these patients were offered the option of retransplant, but 4 did not fit the criteria (Figure 5). The pediatric patients were monitored for a mean of 370 days (range, 240-539 days) after the first angioplasty procedure. Although no clinical and/or radiologic signs of restenosis developed in 7 of the pediatric patients, 4 patients (36.3%) developed restenosis early (Figure 6). Although 1 patient with early-term restenosis underwent a repeat balloon angioplasty procedure, that patient died due to graft failure. In 2 of the remaining 3 patients, repeat balloon angioplasty plus expandable metallic stent implantation was performed; the remaining 1 patient underwent repeat balloon angioplasty sessions without stenting (Figure 7).
Patients presented to our clinic with signs of HVOO at posttransplant days 14 to 1440 (mean ± SD of 207.5± 263 days), representing the 2 groups based on the timing of the emergence of the clinical and radiologic signs: early-onset (8 patients) and late-onset (27 patients). Significant differences were shown between these groups with respect to underlying factors (P < .001), concentration of albumin (P = .01), time from LDLT to the emergence of HVOO (P = .02), and time from LDLT to the last visit (P = .02). The mortality rate was 62.5% in the early-onset group and 22.2% in the late-onset group but did not meet the statistical threshold for significance. Characteristics of these various subgroups are detailed in Table 2.
Eleven patients died during the follow-up period (mean ± SD of 854 ± 804.8 days; range, 66-2465 days). For analysis, the patients were divided into 2 subgroups according to mortality: survivors (24 patients) and nonsurvivors (11 patients). Although the 2 subgroups differed significantly with regard to mean total bilirubin level (P = .03) and time from LDLT to the last visit (P = .03), all other demographic parameters examined were comparable (Table 3). The most notable difference between the 2 groups was in the level of bilirubin, with the latter being significantly greater for the nonsurvivors.
Various alternative techniques have been developed in the past quarter century to expand the donor liver pool and to overcome sizing problems between the recipient and the donor, particularly in pediatric patients. These techniques include reduced-size DDLT, split-size DDLT, segmental LDLT (right, left, and left lateral segment), and monosegment LDLT.1,2,5 In parallel to the introduction of these alternative techniques aimed at expanding the donor organ pool for LT, the incidence rates of formerly relatively rare vascular and biliary complications have increased in cases of segmental liver transplant. According to previous reports, most post-LDLT complications are caused by limited graft volume (graft-to-recipient weight ratio of < 0.8), bile duct variations (eg, multiple bile ducts and narrow bile ducts), complex venous drainage pattern (eg, V5, V8, right inferior hepatic veins and middle hepatic vein), short and narrow venous structures, abnormal portal vein patterns (duplex or triplex portal vein), and hepatic artery variations (eg, multiple arteries, short artery, and narrow artery).4
Hepatic venous outflow obstruction is a rare, albeit serious, vascular complication of liver transplant, which may lead to graft loss or death in the absence of appropriate treatment.2,4,6 Promptly recognizing and addressing this condition are vital steps for graft function.3 Early-onset HVOO can result from a number of factors, such as a tense suture line, bending of a redundant hepatic vein, mismatch between the size of recipient and graft, low recipient-to-donor body mass index ratio, compression of the IVC by a large graft, and thrombosis.2,4,7 In contrast, the cause of late-onset HVOO includes intimal hyperplasia, fibrosis development secondary to inflammation (ie, abscesses or bile leaks), twisting of an anastomosis, external compression due to graft growth, and anastomosis strictures secondary to fibrotic changes in the dissection field.2,4
After a first description of the standard end-to-end cavocaval anastomosis technique for DDLT, researchers were stimulated to develop improved anastomosis techniques, such as the piggyback and the modified piggyback to minimize the rate of HVOO. As a result, rates of venovenous bypass-related complications dropped remarkably and the incidence of posttransplant HVOO has been reduced to acceptable levels. The greatest advantages of DDLT lie in its feature of graft retrieval together with retrohepatic IVC and its ability to facilitate reconstruction of a cavocaval anastomosis of sufficient size. On the other hand, a short hepatic vein stump, a narrow hepatic vein orifice, and multiple venous orifices (middle hepatic vein, V5, V8, right inferior hepatic vein, etc) of the living right lobe can lead to serious problems during the venous reconstruction procedure.3 Thus, a common large opening venous drainage model is crucial for avoiding post-LDLT HVOO.
Despite being affected by several factors, such as donor origin (DDLT or LDLT), graft type (right, left, left lateral, or reduced size), and patient age, the rate of overall posttransplant HVOO is between 0.8% and 16.6%.3,4 Whereas the reported incidences of post-LDLT HVOO range between 2% and 16.6%, the reported incidence of post-DDLT is substantially lower, ranging from 0.8% to 2.5%.3,8 When considering graft type, the post-LDLT HVOO incidence ranges between 0.8% and 8.8% for right lobe and between 4.5% and 12.5% for left lobe/left lobe lateral segment.6,9,10 Our Liver Transplant Institute had an overall incidence of post-LDLT HVOO of 3.46% (2.8% for adults and 7.4% for children). Our experience in right lobe implantation and the relative technical simplicity of right lobe venous reconstruction may represent the major factors supporting our overall success with application of this procedure and its outcomes. Another reason may be the more rapid left lobe regeneration, during which the left hepatic vein is at higher risk of distortion and tension. Moreover, these reasons may explain why the incidence of HVOO was relatively higher in our hospital for the pediatric cases, in which left lobe/left lateral segment grafts were used, as well as for adult cases undergoing left lobe implantation. In agreement with this idea, Kitajima and associates7 recently reported that left lobe implantation was an independent risk factor for post-LDLT HVOO.
Another important finding of our study was a significantly higher total bilirubin level in the nonsurviving patient group than in the surviving patient group (P = .03). No study in the current literature has addressed the role of serum bilirubin levels in deceased patients after HVOO, which precludes deeper reasoned consideration of our results. However, it is well known that hyper-bilirubinemia is toxic to many organ systems, particularly the kidneys; therefore, it can directly or indirectly affect mortality in this patient population. Serum bilirubin level is the most important parameter of the Model for End-Stage Liver Disease score (the major score used to determine ranking of liver transplant recipients on transplant wait lists). Thus, serum bilirubin level can be an important parameter reflecting survival probability of patients diagnosed with HVOO.
Posttransplant HVOO is diagnosed on the basis of one or a combination of clinical signs and symptoms (eg, ascites, pleural effusion, or hepatomegaly), findings from a radiologic evaluation (ie, ultraso-nography, multidetector CT, or conventional venography), and findings from histopathologic analysis (eg, congestion, hemorrhage, or necrosis around the central vein).8,11 The foremost and most commonly used noninvasive diagnostic tools for radiologic diagnosis are Doppler ultrasonography and multidetector CT, with the latter having both a greater sensitivity (87% vs 97%) and greater specificity (68% vs 86%), as well as offering excellent spatial resolution for parenchymal changes suggestive of venous congestion on portal and delayed phase images.4 In cases with equivocal diagnosis or those that require intervention, conventional angiography is the procedure of choice. Hepatic venography is quite useful for determining both the pressure gradient around the anastomosis and the decision of whether to proceed with therapeutic procedures, such as balloon angioplasty/stenting. Although the pressure gradient plays an important role in diagnosis of HVOO, no consensus has been reached yet as to the correlation of that gradient with stenosis severity and its cut-off level. Some authors have advocated a pressure gradient > 3 mm Hg as criterion for treatment planning, whereas others have indicated that that levels should be > 5 to 6 mm Hg.3,4,6,8 The most important limitation of the present study is the lack of any gradient measurement for making the diagnosis of HVOO; unfortunately, our radiology unit lacks a manometry device for measuring that pressure gradient. Thus, we diagnosed HVOO and rated the success of the balloon angioplasty or stenting procedure on the basis of the view of a radiologist with 10 years of experience in interventional radiology, along with analyses of postprocedural clinical/radiologic data.
Timing of symptom onset and the experience in interventional radiology of the treating center are the 2 most important factors dictating the approach to HVOO cases. There is no consensus in the literature as to which of the options of balloon angioplasty, balloon angioplasty plus stenting, or primary stenting should be selected for treatment of early-onset HVOO cases.6,10 Although authors have reported favorable outcomes with primary stenting, they have also advocated that balloon angioplasty is still the best option for early-onset HVOO, suggesting that stents both increase the tendency for thrombosis and cause technical difficulties in the case of repeated liver transplant.3 In contrast, the rationale of authors advocating avoidance of balloon angioplasty are as follows. First, rupture risk increases with balloon angioplasty in fresh anastomoses (≤ 4 weeks). Second, angioplasty is incapable of resolving complications such as IVC compression by the graft or bending of a redundant hepatic vein, which are the major causes of early-onset HVOO. Third, there exists a risk of elastic restenosis after balloon angioplasty.3 Accordingly, those authors advocate that primary stenting should be preferred over balloon angioplasty.3
The technical success rates after primary stenting and balloon angioplasty have been reported as ranging from 96% to 97.6% and from 91% to 100%, respectively.3,4,6,12 The clinical success rates reported range between 82.2% and 90%.3,6,8 The patency rates at postprocedure months are reported as 78% for 1 month, 67% to 80% for 3 months, 65% to 67% for 6 months, 56% to 82.3% for 12 months, 57% to 75% for 36 months, 56% to 72.4% for 60 months, and 52% for 120 months.3,6,8,12 Among the patients presented here, 48.6% completely recovered after the first angioplasty procedure and did not require a repeat angioplasty; however, 70% of patients with early-onset HVOO completely recovered with repeat angioplasty/stenting. The long-term success rate of interventional radiology was 88.9%, although only 72.7% for our pediatric patients.
Because irreversible dense fibrotic changes emerge around the anastomosis in most late-onset HVOO cases, the first option is almost always balloon angioplasty.2,6 Recurrence may be shown in 20% to 45% of cases after the first balloon angioplasty, and either repeat angioplasty or metallic expandable stent placement may be appropriate approaches for treating these cases.2 In our study, 28.6% of patients with HVOO developed recurrence after the first angioplasty. Because we have no experience with early-term stenting, we cannot make solid interpretations about this finding; however, we can note that patients who received stents at a later period developed no recurrence. Another major limitation of our study is the lack of preprocedural and postprocedural biopsy reports to verify whether histopathologic alterations (fibrosis) occurred in the liver parenchyma. These results are important because histopathologic outcomes may not necessarily be favorable, even though long-term patency is achieved in a significant proportion of patients after radiologic intervention.2
The debate continues regarding which approach, and its related considerations, to take with patients who have HVOO recurrences after radiologic intervention. Some authors have suggested that a second angioplasty procedure is required when recurrence ensues after a first balloon angioplasty procedure, with stenting being considered appropriate for cases that recur despite a second angioplasty procedure.8,10,13 In one study, satisfactory results were attained with repeat angioplasty, with no cases of subsequent rupture, thrombosis, or mortality.13 Thus, repeat angioplasty could be a safe treatment option. Several other authors have advocated that stents may adversely affect vascular growth in pediatric patients, suggesting that stenting is not appropriate for this patient population. Still, several surgeons are reluctant to put forth an option on advocation for stenting as it may create difficulty for transplant procedures.
DOI : 10.6002/ect.2017.0045
From the 1Department of Surgery, the 2Liver Transplant Institute and the
3Department of Radiology, Inonu University Faculty of Medicine, 44280 Malatya,
Acknowledgements: The authors have no sources of funding for this study and have no conflicts of interest to declare. S. Yilmaz, M. Yilmaz, and B. Barut performed the surgical procedure and patient care; S. Akbulut, S. Koc, and V. Soyer designed the literature review and organized the report; S. Koc and M. Yilmaz M collected the data; R. Kutlu performed invasive radiologic procedures and provided radiologic information; S. Akbulut wrote the paper.
Corresponding author: Sami Akbulut, Department of Surgery and Liver Transplant Institute, Inonu University Faculty of Medicine, Turgut Ozal Medical Center, Elazig Yolu 15. Km, 44280 Malatya, Turkey
Phone: +90 422 3410660
Figure 1. Back-Table Venous Reconstruction of the Right Liver Lobe Via the All-in-One Technique
Figure 2. When Distance Between the Right Inferior Hepatic Vein and Right Hepatic Vein Was > 5 cm, a Saphenous Vein Graft Was Wrapped Around the Right Inferior Hepatic Vein to Form a Second Fence Conduit
Figure 3. Reconstruction of a Fence Conduit to Achieve Optimum Drainage in the Left Lobe Liver Graft
Figure 4. Inferior Vena Cavography and Hepatic Venography Procedures on a Right Hepatic Venous Obstruction That Took Place in a Patient Who Underwent a Right Liver Graft Transplant
Figure 5. Treatment Algorithm for Adult Patients With Hepatic Venous Outflow ObstructionFigure 6. Hepatic Venography Performed for Severe Left Hepatic Venous Obstruction That Occurred in a Pediatric Patient Who Underwent a Left Lobe Lateral Segment Implantation
Figure 6. Hepatic Venography Performed for Severe Left Hepatic Venous Obstruction That Occurred in a Pediatric Patient Who Underwent a Left Lobe Lateral Segment Implantation
Figure 7. Treatment Algorithm for Pediatric Patients With Hepatic Venous Outflow Obstruction
Table 1. Demographic and Clinical Characteristics of the Adult and Pediatric Patients With Hepatic Venous Outflow Obstruction
Table 2. Comparison of the Early-Onset and Late-Onset Hepatic Venous Outflow Obstruction Subgroups in Terms of Demographic, Clinical, and Biochemical Characteristics
Table 2 (cont). Comparison of the Early-Onset and Late-Onset Hepatic Venous Outflow Obstruction Subgroups in Terms of Demographic, Clinical, and Biochemical Characteristics
Table 3. Comparison of the Surviving and Nonsurviving Subgroups in Terms of Demographic, Clinical, and Biochemical Characteristics
Table 3 (cont). Comparison of the Surviving and Nonsurviving Subgroups in Terms of Demographic, Clinical, and Biochemical Characteristics