Changes in body composition after transjugular intrahepatic portosystemic stent in cirrhosis: a critical review of literature
Srinivasan Dasarathy, MD, Department of Gastroenterology and Hepatology, Cleveland Clinic, 9500 Euclid, Avenue, NE4-208, Cleveland, OH 44195, USA
Change in body composition with reduced muscle mass with or without loss of fat mass occurs in 60–90% of patients with cirrhosis. This has an adverse impact on the outcome of these patients and is an understudied area. Transjugular intrahepatic portosystemic stent (TIPS) is now a standard therapy for portal hypertension but its beneficial impact on nutritional indices is not well recognized. We included all publications on TIPS that had any nutritional index as an outcome measure or end point. Given the heterogeneity of the patient population, differences in study design and outcome measures, a meta-analysis was not feasible. Data were summarized and interpreted. A total of eight studies have been published on the changes in body composition after TIPS in cirrhosis in a total of 152 patients followed for 3–12 months. Improvement in fat-free mass and fluid-free or ascites-free body weight was reported in all studies. Plasma leptin, IGF1, insulin sensitivity, rate of glucose disposal and growth hormone did not change after TIPS. One study measured muscle strength that improved. Direct measurement of skeletal muscle mass was not performed in any study. TIPS resulted in an improvement in body composition. Given the clinical significance of skeletal muscle and fat mass in cirrhosis, nutritional indices should be considered to be an important outcome measure in patients with TIPS. The mechanism of these is unclear, but its clinical implication is that this may contribute to the improved survival after TIPS.
Malnutrition is the most common complication in patients with cirrhosis of the liver that adversely affects their outcome before, during and after liver transplantation (1). The term malnutrition and the methods of its assessment in cirrhosis are not clear, and there is no standard definition of these terms. Sarcopenia or loss of muscle mass, reduction in fat mass and loss of both fat and muscle mass or cachexia have all been reported in patients with cirrhosis with prevalence rates varying from 30 to 100%, depending on the assessment parameter used (2, 3). A number of instruments have been evaluated that include (i) subjective global assessment, (ii) anthropometric evaluation that measure triceps skinfold thickness and mid-arm circumference to then calculate the fat and muscle mass, (iii) muscle strength by quantifying grip strength, (iv) determination of body components by bio-electrical impedance analysis (BIA), dual-energy X-ray absorptiometry (DEXA), air displacement plethysmograpy using BodPod®, isotopic tracer distribution, (v) biochemical measures including plasma albumin, pre-albumin, transthyretin, retinol binding protein and (vi) immunological assessment using lymphocyte count and delayed hypersensitivity reactions (2, 3). The prevalence of malnutrition is much higher when biochemical rather than body composition criteria are used (4). Even though malnutrition has not been defined precisely in patients with cirrhosis, most published studies have used a two-compartment model that is an indirect measure of a reduction in fat-free mass and fat mass in patients with cirrhosis (5). Despite the recognized limitations of the two-component model, combinations of anthropometric evaluation and quantification of fat-free mass are the most frequent measures of malnutrition in cirrhosis (6).
Even though reduced muscle mass is not stated explicitly, most descriptive studies on malnutrition in cirrhosis have used reduction in muscle mass (sarcopenia) and fat mass to define malnutrition (2). The potential mechanisms for protein malnutrition in cirrhosis that primarily manifests as reduced skeletal muscle mass include increased whole body and skeletal muscle protein breakdown and reduced rate of protein synthesis in the skeletal muscle (7–13). Measures to reverse protein malnutrition have included increased nutritional intake, amino acid supplements, insulin-like growth factor 1 (IGF1), oxandralone and growth hormone (2). However, these have very limited benefit and their long-term benefits have not been assessed. Failure to increase skeletal muscle protein synthesis in response to different interventions following immobilization has been termed anabolic resistance (14). Impaired protein synthesis response to chronic branched chain amino acid administration has been reported in cirrhosis (15) and may be a consequence of anabolic resistance. Both hepatocellular dysfunction and portosystemic shunting occur in cirrhosis and potentially contribute to this anabolic resistance resulting in sarcopenia and cachexia of cirrhosis. A number of abnormalities contribute to sarcopenia of cirrhosis (Table 1). Many of these have been considered to be secondary to portal hypertension and can potentially be reversed with a reduction in portal pressure. These include improvement in protein losing enteropathy (16, 17), gastroparesis, small bowel dysmotility (18, 19) and control of ascites (20). One potential intervention that lowers portal pressure is the placement of a transjugular intrahepatic portosystemic stent (TIPS) (21). A number of small prospective studies in humans (Table 2) have shown that placing TIPS improves body composition in patients with cirrhosis but this benefit is under recognized. The present review of the published literature was performed to examine nutritional outcomes in patients with cirrhosis who underwent a TIPS procedure. Given the heterogeneity of the data, we have critically reviewed the published literature.
Table 1. Potential causes of sarcopenia of cirrhosis
|Reduced caloric and protein intake|
|Anorexia, nausea, emesis, gastroparesis,* small intestinal dysmotility*|
|Portal hypertension with impaired mucosal absorption*|
|Recurrent gastro-intestinal bleeding*|
|Early satiety because of ascites and portal hypertensive gut mucosal congestion*|
|Hospitalization and starvation related to procedures|
|Bile salt deficiency-induced malabsorption in cholestasis|
|Unpalatable diets – sodium restriction*|
|Impaired protein synthesis response|
|Growth hormone deficiency, acquired resistance to growth hormone|
|Impaired signalling pathways in skeletal muscle|
|Altered energy response with reduced plasma branched chain amino acids|
Table 2. Nutritional status and body composition changes after transjugular intrahepatic portosystemic stent in cirrhotic patients
|Trotter et al. (31)||35||Refractory ascites||Body weight||Body weight, Child's score, ascites at 2 and 8.8 months after TIPS||At 2 months weight decreased|
At 8.8 months weight increased by 5.5 kg
|After the initial loss of weight because of disappearance of ascites, there was weight gain without fluid accumulation|
|Allard et al. (24)||14 (10 completed the study)||Ascites||REE, TBN, TBP, TBF, muscle function test||Patients were assessed at baseline before, and at 3 and 12 months after TIPS||Improved REE, TBN, and TBF|
No change in TBP
No change in muscle function test
|First study to assess TIPS effect on nutritional status|
Small number of patients and no controls
|Nolte et al. (29)||31 patients. Controls: 19 cirrhotic patients without TIPS||–||Serum leptin, BMI, CP score||Patients were assessed at baseline and 3 months after TIPS||Leptin levels increased after TIPS|
Ascitic-free BMI increased in women but not men
No change in CP score
|Investigated the influence of serum leptin on nutritional status after TIPS|
Found that elevated leptin levels were not a major reason for poorer body composition in cirrhotics
|Plauth et al. (30)||21 patients||14 with EV|
–Seven with ascites
Anthropometry, REE, nutritional intake (NI)
|Patients were assessed at baseline and at 6 and 12 months||Improved BCMTBP and BCMBIA|
Improved REE and NI (because of increase in protein and carbohydrate intake)
|Confirmed the improvement in nutritional status after TIPS|
The improvement was more readily observed than the study by Allard and colleagues
|Holland-Fischer et al. (26)||26 patients of which 17 completed study|
Ascites+GI bleed prevention 2; variceal bleed prevention-5
|BMI, BCMBIA, IGF, IGFBP, HOMA, albumin||Baseline, 1, 4, 12, 52 weeks||BCMBIA increased by 3.2 kg (13.1%) at 52 weeks|
Weight, IGF1, IGFBP 1, 2 and 3, fasting glucose unchanged. Fasting insulin and glucagon increased by 58%
|TIPS results in increased BCM by 1 year and was not because of changes in IGF1|
|Camci et al. (25)||12 (six completed study)|| ||BMI, estimated muscle mass, CLDQL score, serum albumin||Baseline, 3 months, 6 months||BMI increased from 21.4 to 25.5 kg/m2|
Estimated muscle mass increased from 16.6 to 20.5. Serum albumin increased from 2.46 to 2.76 and CLDQL score improved from 130 to 150
|Small study but showed that improvement in ascites and nutritional indices were accompanied by improved QOL|
|Montomoli et al. (28)||21 patients. Patients were divided into normal weight (NW) and overweight (OW) with BMI >25||12 with ascites|
–Seven with EV
–Two with both
|Body composition by BIA (dry lean mass, fat mass and TBW), FFA, insulin and HOMA-IR and GEC||Patients were assessed at baseline and 1, 4, 12, 52 weeks after TIPS||Dry lean mass increased in NW but not OW patients|
FFA increased in OW
Insulin and HOMA-IR increased in NW
Liver function as assessed by GEC remained stable in NW while in OW there was a trend towards deterioration
|First study to address the impact of pre-TIPS overweight and obesity on the post-TIPS nutritional status and insulin levels|
|Holland-Fischer et al. (27)||12 patients (11 data available)||Intractable ascites-7|
Recurrent variceal bleeding and ascites-4
|BIA measured BCM, BMI, fat-free mass, REE, GTT, hyperinsulinaemic clamp||Before and 6 months after TIPS||BCMBIA increased 4.8 kg (19.3%). FFM increased by 5.7 kg (10.9%)|
REE was unchanged
Insulin sensitivity, rate of disposal of glucose, insulin clearance, growth hormone, cortisol unchanged unchanged after TIPS
|Authors have shown previously that body composition improves after TIPS. This report showed that insulin clearance, growth hormone, cortisol and glucagon levels were unchanged after TIPS and were probably not responsible for the changes in body composition observed|
All published literature as full reports in the English language on TIPS placement were included in this review, if any of the nutritional end points was considered to be a primary outcome measure. Additionally, studies on TIPS that did not use nutritional indices as a primary outcome measure were also reviewed to determine whether they were used as a secondary end point. Medline, Ovid, EMBASE and Google scholar databases were searched using the key words: TIPS, Body composition, skeletal muscle, nutritional indices, sarcopenia and malnutrition. However, none of the studies on TIPS included nutritional parameters as a prespecified outcome measure if it was not a primary outcome measure. Because the precise definition of malnutrition and sarcopenia in cirrhosis is not available, nutritional indices that were evaluated included body mass index, body composition and muscle strength. None of the studies examined anthropometric measures or quantified body composition using methods other than bioelectrical impedance. Despite the known limitations of bioelectrical impedance (22), we accepted that this as a valid measure of body composition because it has been shown to be useful even in the presence of ascites (23). Even though grip strength was not quantified as a measure of muscle strength, one study used electrical stimulation and muscle relaxation of the adductor pollicis (24). For purposes of this review, follow-up was considered from the date of the first assessment to the last assessment reported for nutritional indices and outcomes. We did not restrict the number of patients in the studies included because all of them have few patients. All participants who had TIPS were included. No quality measures were used to select the studies because we wanted to review the entire published literature. A systematic meta-analysis was not performed given differences in the outcome measures, heterogeneous patient population, aetiology of liver disease, severity of liver disease, indication for TIPS, difference in measures of body composition before TIPS and lack of the raw data in the patients studied. Statistical calculations were not performed because a systematic meta-analysis was not feasible and the levels of significance reported by the authors were used where necessary.
After the initial report of a prospective study that showed that cirrhotic patients had improved whole-body weight after TIPS, a total of eight prospective studies were identified whose prestudy endpoint included one or more parameters to measure nutritional outcome (Table 2) (24–31). The nutritional outcomes that were reported included weight, body mass index, body cell mass measured using BIA, fat-free mass, lean body mass, serum albumin and estimated muscle mass. One of the studies published to date reported muscle force measured by the response to electrical stimulation (24). No studies reported mortality of these patients at the end point time. Of the total 152 patients evaluated in these studies, the reported duration of follow-up varied from 3 to 12 months. Body weight and body mass index in all studies increased significantly with the loss of ascites. Four studies measured body cell mass and all of them showed significant improvement (26–28, 30).
The mechanism of improvement in body composition has been studied by three authors (26, 27, 29). Specifically, plasma leptin, IGF1 and IGF-binding proteins, insulin, glucagon, growth hormone and cortisol levels were quantified before and at different times after TIPS. Plasma leptin, IGF1 or IGF binding proteins, growth hormone, plasma cortisol and insulin sensitivity did not change in response to TIPS. Resting energy expenditure (REE) was measured by three authors in a total of 44 patients and showed that mean REE measured using calorimetry increased in 33 patients in two studies (24, 30), while in one study involving 11 patients, there was no change in REE 3 months after TIPS (27). In two studies, the predicted REE using the Harris Benedict equation was reported with divergent findings (24, 30). One study reported that patients were hypermetabolic (30), while in the other study patients were hypometabolic (24). Because raw data were not available, the number of patients in each study was small and the change in all of them was probably not significant, it was concluded that TIPS did not result in a significant alteration in REE. The mean caloric intake was reported and increased following TIPS in 21 patients in one study (30) and did not change in two studies (24, 26).
The indication for TIPS included refractory ascites in the majority and oesophageal varices with bleeding or a combination of the two (Table 2). In the studies that included only patients with refractory ascites (24, 31), improvement in body composition was similar to that in patients with ascites as well as gastrointestinal bleeding (26–28, 30).
The present critical review examined the published data and showed that one of the understudied outcomes after TIPS is the beneficial effect on nutritional indices including body cell mass measured by BIA. This suggests that the anabolic resistance of cirrhosis can be overcome by specific interventions. Despite the extensive data that malnutrition adversely affects the outcome of patients with cirrhosis, few studies exist that have shown that improved nutritional parameters result in better outcome in cirrhotic patients (2). A major limitation of the published data is that reduction in muscle mass or sarcopenia has not been distinctly separated from reduction in fat mass (32). However, mid-arm muscle area and grip strength have been shown to correlate significantly with the outcome in patients with cirrhosis before and after liver transplantation (3). Imaging using computed tomography (CT) reconstruction and magnetic resonance imaging (MRI) have been used as measures of skeletal muscle mass in vivo in humans (33–35). Recently, CT-measured psoas muscle area at the L4 vertebra was shown to correlate with survival after liver transplantation (32). These data suggest that reduced muscle mass has an adverse impact on the survival of patients with cirrhosis. Even though the survival advantage of improvement in fat-free mass, an indirect measure of whole-body skeletal muscle mass has not been specifically evaluated, we speculate that improvement in body composition may contribute to the survival advantage reported after TIPS (21, 36).
Four studies reported an ascites-free increase in BMI (24, 25, 29, 31). This may be because of either an increase in fat mass or fat-free mass. Estimated muscle mass that was a clinical measure was also reported by one of the authors to have increased significantly after TIPS (25). These data in small numbers of patients in each individual study had a follow-up duration ranging from 3 to 12 months. No study reported nutritional alterations or changes in body composition after 12 months. Even though precise measures of skeletal muscle mass were not measured after TIPS, bioelectrical impedance analysisbody cell mass (BIABCM) was used as a measure of fat-free mass in four studies (26–28, 30) and estimated muscle mass in one study. These studies showed an increase in fat-free mass after TIPS (25). The fat-free mass in humans is comprised predominantly of skeletal muscle (37–40). Therefore, an increase in BIABCM is an indirect evidence of improvement in muscle mass (41, 42). Even though the terms fat-free mass and lean body mass have been used interchangeably, it must be emphasized that there is a difference between the two. Lean body mass is an in vivo entity that contains the small percentage of non-sex-specific essential fat equivalent to approximately 3% of body mass (located chiefly within the central nervous system, bone marrow and internal organs). In contrast, fat-free mass represents the body mass devoid of all extractable fat and it is an in vitro entity appropriate to carcass analysis. In normally hydrated healthy adults, the fat-free mass and lean body mass differ only in the essential fat compartment (43, 44). Most instruments that are currently used to assess body composition including BIA, DEXA and Bodpod® measure the whole-body fat mass or fat-free mass (6, 22, 45). In most studies in cirrhotics, body composition has been quantified using BIA that measures fat-free mass (46, 47). Based on limited human carcass studies, and studies in animals, it has been estimated that about 60% of fat-free mass is comprised of skeletal muscle (40, 48, 49).
Reproducible in vivo measures of skeletal muscle mass that are believed to be relatively precise include mid-arm muscle area, measures of skeletal muscle mass in the extremities using MRI, DEXA or CT scan (42, 50, 51). Cost, availability, need for training and technical expertise to interpret the observations using these measures make bioelectrical impedance a frequently used measure of fat-free mass (52). However, BIA is based on the principle of electrical conductivity and depends on the whole-body water content and is believed to measure total body water and extracellular water (ECW). However, strictly speaking, BIA measures a weighted sum of ECW and intracellular water resistivities (∼25%) (53). Therefore, the single-frequency BIA that has been used in all the studies reported to date in patients before and after TIPS cannot determine differences in intracellular water. The current BIA methods are not considered to be sufficiently accurate to assess total body water under conditions of hydration change (54). Because cirrhosis is accompanied by the increased whole-body water, altered hydration of individual organs especially in the skeletal muscle compartments may violate the assumptions relating to the consistency of density and hydration factor, both of which form the basis of the calculations of fat-free mass using BIA (55). Therefore, data using BIA in cirrhotics cannot be used as a precise measure of fat-free mass or skeletal muscle mass. The recently described CT or MR-imaging-derived muscle volume is considered to be better measure, but have not been assessed following TIPS (32). Of the various measures of skeletal muscle function, the most reproducible and robust measure that predicts outcome not only in patients with cirrhosis but also in other conditions characterized by sarcopenia is grip strength, but this has not been reported in patients following TIPS (56–58). Even though there are no data on long-term nutritional changes after TIPS, a review of the published data suggests an improvement in fat-free mass and liver disease-related quality of life improved after TIPS (59).
The major finding of this review is that body composition improves after TIPS with some studies reporting an increase in fat-free mass. These results are unexpected because animal data in the portacaval anastamosis (PCA) rat, a model of the nutritional consequences of portosystemic shunting have shown significant reduction in muscle mass (60, 61). The lower muscle mass in the animal model has been recently shown to be because of an upregulation of transforming growth factor-β superfamily member, myostatin that inhibits protein synthesis and impairs satellite cell function. Satellite cells are myogenically committed stem cells that are responsible for the maintenance of skeletal muscle mass as well as reversal of muscle atrophy. In this model, reversal of these changes have been reported by the use of follistatin, a myostatin antagonist without reversing the shunt or alteration in the biochemical and hormonal consequences of portosystemic shunting (62).
A number of potential reasons for this discrepancy between the reported reduced muscle mass in the PCA rat and the improved fat-free mass in humans after TIPS exist. In patients with TIPS, complete portal diversion does not occur and there is continued hepatopetal blood flow in the portal vein in contrast to the end to side PCA rat in which there is complete portal venous deprivation (63–65). Additionally, the PCA rat does not have any intrinsic liver disease and is a model of portosystemic shunting without portal hypertension, in contrast to humans who have significant portal hypertension that is reversed following TIPS (21). Despite the deleterious effects of TIPS that include aggravation of hyperammonaemia, potential worsening liver and renal function (66), a number of beneficial effects follow TIPS and include re-absorption of ascites, improvement in splanchnic venous return from the intestine, reversal of protein losing enteropathy, prevention of further episodes of rebleeding or paracentesis, all of which potentially prevent further reduction in muscle mass. Improvement in body composition after TIPS can also be because of a reversal in the hypermetabolism of cirrhosis. In two studies, however, hypermetabolism did not decrease (27, 30). In another study, there was an increase in measured REE compared with the calculated REE (24). Other potential mechanisms for improved body composition include alterations in regulatory signalling proteins that increase skeletal muscle protein synthesis to overcome the anabolic resistance in cirrhotic patients. These putative mechanisms include a reduction in plasma leptin concentration or an increase in circulating IGF1. Reduction in circulating leptin would improve appetite, decrease skeletal muscle AMPK phosphorylation with subsequent improvement in muscle protein synthesis (67). IGF1 directly stimulates protein synthesis and inhibits proteolysis in skeletal muscle (68). These concepts are, however, underexplored and one study each has reported that plasma concentration of leptin or IGF1 did not change after TIPS, suggesting that these alterations are unlikely contribute significantly to the reversal of anabolic resistance and improved body composition following TIPS (26, 29).
Several limitations of the reported studies on nutritional outcomes after TIPS exist. Sarcopenia was not documented by precise anthropometric measures or imaging criteria and indirect measures of skeletal muscle mass and fat mass were quantified using BIABCM in the few studies that reported these measures. The impact of improvement in nutritional measures was not related to survival. A systematic meta-analysis could not be performed as stated earlier because of the heterogeneity of patients, differences in study design and variable outcome measures. Alterations in nutritional indices were reported over a relatively short duration of follow-up. Despite these limitations, this is the first critical review of the published literature on post-TIPS nutritional indices in humans. This report also provides the rationale for using specific and predefined nutritional outcomes to evaluate potential long-term beneficial effects of these improvements after the placement of TIPS.
These studies in combination with the recent report that early TIPS for variceal bleeding improves survival (21) suggest that one mechanism for this could be a reversal of sarcopenia, even though this was not evaluated in this study. Our report suggests that future studies on TIPS should include quantifying nutritional indices including CT or other precise measures of muscle mass, anthropometric indices and muscle strength (42) and the impact of these changes on survival and quality of life. Given the interest in measures to reverse cachexia and sarcopenia that are becoming increasingly available for potential clinical trials (69), the impact on the alteration in signalling pathways and the role of changes after TIPS that are responsible for improvement in body composition and the specific compartment in which this improvement occurs will also be of enormous clinical interest. Despite the reported improvement in body composition after TIPS, two questions remain that include whether the improvement was a result of reversal of sarcopenia and whether this contributed to the improved survival after TIPS (21).
We conclude that improvement in body composition with an increase in fat-free mass that reflects skeletal muscle mass occurs following TIPS in patients with cirrhosis. These data are especially exciting because they suggest that the anabolic resistance in cirrhosis is not absolute and can be overcome. The mechanism(s) responsible for this are not known and may be related to reversal of portal hypertension and its consequences. Limited published data suggest that future prospective studies on TIPS should include precise measures of skeletal muscle mass and nutritional indices as outcome measures.