html xmlns=”http://www.w3.org/1999/xhtml”>
28
Alcoholic liver disease
Screening
In heavy drinkers, liver-related mortality is mainly attributed to cirrhosis and hepatocellular carcinoma (HCC). Therefore, the main objectives of the screening are: to identify patients with significant liver injury; to characterize the main risk factors for HCC; and to diagnose HCC early.
We focus on the non-invasive screening of cirrhosis and on screening of HCC.
Non-invasive screening of cirrhosis
Assessment of the stage and severity of liver injury requires liver biopsy, an invasive procedure associated with complications leading to death in 0.02% of patients [1]. Less than 30% of heavy drinkers have significant liver injury such as extensive fibrosis, alcoholic hepatitis or cirrhosis [2]. Routine liver biopsy is not essential in 70% of heavy drinkers. The recent development of non-invasive methods for the screening of extensive fibrosis and/or cirrhosis is an important progress to identify the subgroup with significant alcoholic liver disease (ALD) and will avoid screening with liver biopsy.
Initial development of biomarkers of fibrosis
Serum hyaluronate was identified to be useful for the diagnosis of cirrhosis [3–5], with a good AUROC for the diagnosis of significant fibrosis (Metavir class ≥ F2 versus F0–F1).
Prothrombin time (PT) predicts liver fibrosis and is inversely correlated with the area of fibrosis [4, 6]. A study showed that PT has a high diagnostic accuracy for severe fibrosis or cirrhosis, particularly in those with ALD [7]. Results of the studies evaluating serum procollagen III propeptide (PIIIP) in assessing fibrosis are controversial [4, 8–11]. A major problem, beside the heterogeneity of the assays, is that even with the same assay the PIIIP screening cutoff is unknown.
Apolipoprotein A-I (ApoA-I) has an independent and discriminative value for the diagnosis of fibrosis [12]. A simple index called PGA, developed by combining PT, γ-glutamyl transpeptidase and ApoA-I was efficient for screening of cirrhosis [6].
Serum α-2 macroglobulin has been evaluated as a marker of cirrhosis. Levels were higher in patients with cirrhosis than in patients without [13,14]. Based on these results, the hypothesis that serum α-2 macroglobulin could improve the accuracy of PGA index for the diagnostic of cirrhosis was confirmed [15].
These studies constitute the background for the development of the new blood tests for screening of fibrosis. As an example, PT, Apo A-1 and serum α-2 macroglobulin have been incorporated in the formula of non-invasive methods such as Fibrotest® and Fibrometer®.
Non-invasive tests to screen for fibrosis
Several new blood tests combining different biomarkers of fibrosis are now available. These tests have been initially designed for patients with hepatitis C but seem to be as efficient in patients with ALD.
Aspartate aminotransferase (AST) to platelet ratio index (APRI) is an accurate non-invasive marker of fibrosis in chronic hepatitis C, and has been evaluated in heavy drinkers. A total of 1308 subjects from two studies of ALD were evaluated, with a liver biopsy available from 781 non-cir-rhotic patients and a history of decompensation in 527 [16]. In the 507 patients with biopsy-confirmed fibrosis, the sensitivity of APRI for significant fibrosis was 13.2% and the specificity 77.6%. Twenty percent were misclassified. Thus, APRI has low sensitivity and specificity for the diagnosis of significant fibrosis in patients with ALD. The use of this test to screen fibrosis is not recommended.
Fibrotest® combines alpha-2-macroglobulin, hap-toglobin, GGT, ApoA1 and bilirubin, corrected for age and sex. It was designed to predict advanced fibrosis in chronic hepatitis C, and is also known to have high predictive values for the detection of significant fibrosis in ALD [5, 17]. In a study of 221 consecutive patients with biopsy-proven ALT, the mean Fibrotest® value ranged from 0.29 in those without fibrosis to 0.88 in those with cirrhosis. For the diagnosis of cirrhosis, the AUROC was very high at 0.95 [5]. FibrometerA®, combining PT, alpha-2-macroglob-ulin, hyaluronic acid and age has similar accuracy in ALD [18]. In the validating step, the Fibrometer® AUROC curve was 0.892 in overall patients and 0.962 in patients with ALD. Hepascore® combines bilirubin, GGT, hyaluronic acid, alpha-2-macroglobulin, age and sex. The diagnostic accuracies of Fibrotest®, Fibrometer® and Hepascore® were compared in patients with ALD [19]. The diagnostic values of FibrometerA® and Hepascore® did not differ from that of FibroTest® for advanced fibrosis (AUROCs around 0.80) and cirrhosis (AUROCs around 0.90), and were significantly greater than those of non-patented biomarkers APRI, Forns, FIB4). The combination of any of these tests was useless to improve diagnostic performance.
In addition to their diagnostic performance in the screening of fibrosis, non-invasive tests may be as useful as biopsy to predict liver-related mortality. A recent study compared the prognostic and diagnostic values of these biomarkers in 218 patients with ALD followed up for a median of 8.2 years. Five-year survival was inversely correlated with baseline Fibrotest® value: 98.7% if baseline Fibotest® between 0–0.31, 92.1% if 0.32–0.58 and 68.3% if 0.59–1 [19]. Among the different tests (FibrometerA®, Hepascore®, APRI, Forns, FIB4) Fibrotest® was the only biomarker strongly associated with overall survival in mul-tivariate analysis.
The estimation of fibrosis by measurement of liver stiffness is another alternative to biopsy. In patients with ALD, liver stiffness was correlated with the amount of fibrosis and ranged from 5.7kPa in patients with Metavir score F1, 8.3kPa in F2,17.5kPa in F3, to 40.9 kPa in F4. The optimal cutoff (kPa) for the screening of cirrhosis was 22.7kPa with positive predictive value at 0.85 and negative predictive value at 0.82 [20]. A study comparing the diagnostic performance of transient elastography (TE) to those of biopsy and biomarkers in patients with alcohol abuse demonstrated that TE has a good diagnostic performance [21]. Liver stiffness is correlated with fibrosis. A cutoff at 19.5 kPa identifies cirrhosis with a PPV of 68.6% and a NPV of 87.9%. In this cohort, the diagnostic performance of TE was not significantly different than those of Fibrotest® and Fibrometer® but higher than those of APRI and Hepascore®. Thus, TE is a useful tool for the screening for cirrhosis in populations with chronic alcohol abuse, and could identify patients requiring biopsy.
Screening for hepatocellular carcinoma
Chronic alcohol abuse (>50–80g/day during at least ten years) increases the risk of hepatocellular carcinoma (HCC) by five [22]. This risk does not decrease with abstinence. In alcoholic cirrhosis, the cumulative risk of developing HCC at five years is approximately 8% [23], with a yearly incidence rate of 1% in decompensated alcoholic cirrhosis [22].
Identification of patients at high risk for HCC, and its early detection are of crucial importance. In heavy drinkers, presence of cirrhosis, age >50 years, male sex, serum alpha fetoprotein (AFP) = 15ng/ml, HBsAg and antiHCV antibodies were independently associated with the occurrence of HCC [24–26]. A clinicobiological score identifies two groups at low (three-year cumulative incidence, 0%) and high risk of HCC (three-year cumulative incidence, 24%) [27]. The pre-neoplastic role of liver large cell dyspla-sia (LLCD) has been suggested [28]. A study estimated the cumulative incidence of HCC at three years at 38% in patients with LLCD and 10% in patients without [27]; another study found that LLCD was associated with a fivefold increase in the risk of HCC [29]. Nevertheless, in another study, LLCD was not an independent risk factor for HCC [30]. Taken together, these data do not support LLCD to be a direct precursor for HCC, but support its significance as a marker able to predict HCC development. Patients without the lesions are at low risk of HCC [31]. High grade dysplastic nodules are associated with a high risk of malignant transformation and are clearly classified as pre-neoplastic lesions [32]. Based on these results, biopsy would be useful to identify patients with a higher risk for HCC.
Screening for HCC is usually done with ultrasonography and concentration of AFP. Contrasting data have been reported on the effectiveness of ultrasonography for early detection of HCC: a French center observed that diagnosis of tumour less than 3 cm was made in only 21% of cases [33], but other studies detected small HCC in 76% of cases [34, 35]. A recent Markov model assessed that screening by AFP and ultrasonography was cost-effective [36]. In an RCT evaluating screening in hepatitis B versus no screening, the screening strategy decreased the mortality rate by HCC by 37% [37]. Regardless of these data, screening for HCC by biannual AFP testing and ultrasonography is recommended.
Summary of screening in heavy drinkers
At present, non-invasive methods are available and will be helpful to define the best strategy to identify significant fibrosis in heavy drinkers in routine practice. Patented blood biomarkers (Fibrotest®, Fibrometer® and Hepascore®) and TE are simple, non-invasive, cost-effective and reproducible. As observed in HCV patients, the combination of serum biomarker tests and TE need to be tested in patients with ALD for the screening of fibrosis. This practice can select candidates for liver biopsy and at risk of development of portal hypertension and HCC. HCC screening should be done using biannual AFP dosage and ultrasonography, and the presence of high grade dysplasia at biopsy is a strong predictive marker for high risk of malignant transformation. However, future studies are required to validate these screening strategies.
Treatment
In heavy drinkers, pharmacological treatments and liver transplantation (LT) have been tested to improve survival of heavy drinkers with severe liver injury such as severe alcoholic hepatitis (SAH) or cirrhosis. Except for SAH, the usefulness of pharmacological treatments for controlling the alcohol-induced liver injury is still unsettled.
Pharmacological treatments
Anti-fibrotic therapies
Colchicine Colchicine is an anti-fibrotic therapy which has been widely evaluated in the treatment of alcoholic or nonalcoholic fibrosis. In the first RCT evaluating its effect on long-term survival of patients with alcoholic cirrhosis, five and ten-year survival rates were significantly higher in the colchicine group (75% and 56%, respectively) than in the placebo group (34% and 20%, p < 0.001) [38]. Nevertheless, in patients with AH, two other studies did not observe any effect on short-term survival [39,40]. A recent RCT did not observe any difference between the colchicine and the placebo groups in terms of liver-related mortality: 32% vs 28% [16]. The ineffectiveness of colchicine was confirmed by a meta-analysis combining all the RCTs evaluating colchicine [41]. A1a
Propylthiouracil Two RCTs did not document any effect of propylthiouracil on short-term survival [42, 43]. Meta-analysis of these RCTs confirms the lack of benefit on survival, with a mean difference of 1% (CI: 7–9%) between propylthiouracil-treated and controlled patients.
The effect on long-term survival was analyzed in an RCT with 310 alcoholic patients receiving propylthiouracil (n = 157) or placebo (n = 153) for two years [44], which found that two-year mortality was lower in propylthiouracil group than in placebo group (13% versus 26%, p < 0.05). Nevertheless, this study was limited by two biases: the use of “per protocol analysis” and the cumulative drop-out rates in both groups were approximately 60%.
A systematic review of six RCTs including 710 patients demonstrated no survival benefit of propylthiouracil compared with placebo (odds ratio 0.99) [45]. Moreover, propylthiouracil was associated with a non-signifiacnt trend towards an increased risk of adverse events and serious adverse events (leukopenia). Taking into account these data, propylthiouracil is not considered as an effective therapy in alcoholic cirrhosis. A1a
Other drugs d-penicillamine, vitamin E, (+)-cyanidanol-3, thioctic acid, malotilate and amlodipine have been evaluated in RCTs, but none improved survival [46–51]. A1d
Silymarin The first RCT suggested an improvement in long-term survival of patients with cirrhosis [52]. However, another larger RCT involving 200 patients did not confirm any effect on survival [53]. A1 d
S-adenosyl methionine In an RCT, there was a trend toward greater decrease in overall mortality/ liver transplantation in S-adenosyl methionine than in placebo patients (16% vs 30%, p = 0.077) [54]. A sensitivity analysis restricted to patients with Child A or B classes observed a significant lower mortality/liver transplantation in S-adenosyl methionine patients (12 vs 29%, p = 0.025). In summary, because of the absence of statistical difference on overall patients, the benefit of S-adenosyl methionine is still unknown and future studies are needed. A1d
Phospatidylcholine This has demonstrated a trend towards a decrease in transaminase and bilirubin levels in heavy drinkers, in a trial on alcoholic cirrhosis. Nevertheless, there was no improvement on survival or histology at 24 months [55].
Therapies for acute severe alcoholic hepatitis
Alcoholic hepatitis (AH) occurs in approximately 20% of heavy drinkers. Several investigators have studied SAH intensively, due to its significant early mortality (50–75% in severe forms) [56]. Despite the impact of abstinence on survival in ALD [57–59], pharmacological treatments are required. The treatment of SAH remains controversial and is one of the main challenges in ALD.
Evaluation of any effect on short-term survival requires identification of patients with significant risk of death at one or two months. Until the Maddrey function became available, no reproducible objective criterion existed to predict the risk of early death. Prior to the era of the Maddrey Discriminant Function (DF), survival in untreated control arms ranged from 0% to 81% [56, 60–71]. Maddrey et al. described a DF [56] and later modified it [61]. This modified DF identifies patients at high risk of early mortality: 4.6 (PT-control time in seconds) + bilirubin (in mg/dl). In the absence of treatment, the spontaneous survival of patients with a DF ≥ 32 fluctuated between 50% and 65% [61, 72, 73].
Insulin-glucagon infusion [74–76] and nabolic-andro-genic steroids [71, 77–79] are not effective in SAH. A systematic review confirmed that anabolic-androgenic steroids did not have any benefit on mortality and were associated with serious adverse events [80]. A1a Furthermore, their use is questionable when considering the potential associated risk of HCC.
Corticosteroids
Thirteen RCTs have evaluated corticosteroids in patients with AH [56, 60–71]. These yielded inconsistent results, attributed to the wide differences between studies, since survival of placebo groups ranged from 0% to 81%. The last two RCTs included only patients with either a DF ≥ 32 or spontaneous encephalopathy [60, 61]. Survival in corticos-teroid groups was significantly higher than in placebo groups: 94% vs 65% at 28 days (p = 0.006) [61] and 88% vs 45% at 66 days (p = 0.001) [60]. The response to steroids was lower in case of bleeding [81]. The survival benefit of steroids is restricted to short term [82].
A recent meta-analysis using multivariate statistics to adjust for confounding variables between corticosteroid and control groups concluded that corticosteroids are ineffective [83]. Nevertheless, this was problematic in identifying the effect of steroids because of the limited number of trials (n = 12) and the inclusion of lower-quality trials [84]. Moreover, meta-analysis of published results is not designed to identify the effect of treatment in distinct patients [85]. In the particular setting of AH, this method cannot pool the results restricted to patients with DF ≥ 32, as most of the previous RCTs did not supply the survival data on this subgroup.
To conclude this controversy, individual data of the most recent RCTs were combined [72]. The analysis was restricted to patients with a DF ≥ 32. The findings were: (1) at 28 days, corticosteroid patients had significantly better survival than placebo patients: 84.6 ± 3.4% vs 65.1 ± 4.8%, p = 0.001; (2) steroids, age and creatinine were independently associated with survival at 28 days; (3) steroids induced rapid improvement in liver function by seven days of treatment [72]. The final argument for the benefit of corticosteroids is that all the studies observed that two-month survival of patients treated with corticosteroids was approximately 80% [60, 61, 86–89]. A1a
To progress in the management of SAH, a criterion for early identification of non-responders to steroids has been proposed [87]. This criterion, “early change in bilirubin levels (ECBL)” is defined as a bilirubin level at seven days lower than these on the first day of treatment [87]. However, the ECBL was highly specific but not sensitive enough for predicting death. A specific prognostic model, the so-called Lille model, was developed to identify subjects early on who are unlikely to survive and propose new management based on this specific model [90]. Patients above the ideal cutoff of 0.45 showed a marked decrease in six-month survival as compared to others: 25 ± 3.8% vs 85 ± 2.5%, p < 0.0001. Using the 0.45 Lille model cutoff, close to 40% of patients do not benefit from corticosteroids, which is higher than the 25% previously identified by ECBL [90]. In non-responders (i.e. 0.45 ≥ Lille model) corticosteroids should be interrupted at day seven. A recent study evaluating a two-step strategy consisting of early withdrawal of corticosteroids and a switch to pentoxifylline for 28 additional days in non-responders observed that non-responders to corticosteroids do not obtain any benefit from an early switch to pentoxifylline. Thus, the issue of management of non-responders remains unsettled.
In summary, even if steroids improve short-term survival in SAH, new treatments are required to improve the probability of being alive within the year following the onset of the disease.
New therapies
Extracorporeal liver support
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
