Smoking & risk of advanced liver fibrosis among patients with primary biliary cholangitis: A systematic review & meta-analysis

Primary biliary cholangitis (PBC) is a chronic autoimmune disease of the liver characterized by inflammation and intrahepatic bile duct destruction, resulting in intrahepatic cholestasis1. PBC is a relatively uncommon disease with the reported prevalence of only 20-400 cases per million persons in Northern Europe and North America23. The precise aetiology of PBC is unknown but is believed to be an interplay between genetic and environmental factors45. Patients with PBC may present with abnormal liver chemistry tests without any symptoms, symptoms of cholestasis (pruritus, yellow eyes and fatigue) or signs and symptoms of cirrhosis67.

Cigarette smoking is a known cause of several preventable non-communicable diseases such as coronary artery disease (CAD), cerebrovascular disease, chronic obstructive pulmonary disease (COPD) and malignancy89. The impact of smoking cessation on the prevention of those diseases is substantial. For instance, a study of postmenopausal women found that smoking cessation can decrease the risk of stroke by almost 40 per cent10. Recent studies found that smoking may also have deleterious effects on the liver because of the increased oxidative stress burden and lipid peroxidation, which may lead to hepatic injury and fibrosis11. The effect may be more pronounced among patients who already have chronic inflammation in the liver, including patients with PBC, although clinical data from epidemiologic studies are still limited121314. The current systematic review and meta-analysis was conducted with the aim to comprehensively analyze the association between risk of liver fibrosis and history of smoking among patients with PBC.

Material & Methods

Information sources and search strategy: A systematic literature search was carried out using the MEDLINE and Embase databases from inception to February 2019 to identify original studies reporting the relationship between history of smoking and risk of advanced liver fibrosis in patients with PBC. The systematic literature review was independently conducted by three investigators using the search strategy that included the terms such as, ‘primary biliary cholangitis’, ‘primary biliary cirrhosis’, ‘smoking’, and ‘cigarettes’ (Supplementary Table I). A manual search for additional potentially relevant studies was also carried out using the references of the included studies as well as some selected review articles. This study was conducted in accordance with the Preferred reporting items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines (checklist available as Supplementary Table II). EndNote X7 (Clarivate Analytics, Pennsylvania, United States) was used for study retrieval.

Selection criteria: Only cross-sectional studies that recruited patients with PBC with data on (i) smoking status, and (ii) presence or absence of advanced liver fibrosis for each participant were selected. Odds ratios (OR) with 95 per cent confidence intervals (CI) or sufficient raw data to calculate the same for this association should have been reported. Inclusion was not restricted by study size. When more than one study using the same database/cohort was available, only the study with the most comprehensive data/analyses was included. Retrieved articles were independently reviewed to determine their eligibility by the same three investigators. Any discrepancy was resolved by discussion. The modified Newcastle-Ottawa scale was used for quality assessment of the included studies as described previously15.

Data abstraction: A structured data asbtraction form was used to extract details such as title of the study, publication year, name of the first author, calendar year(s) when and in which country the study was conducted, number and demographic data of participants, definiton of advanced liver fibrosis, method(s) used to evalaute liver fibrosis, definition of positive history of smoking (i.e., definition of ever-smokers), method(s) used to determine smoking status, adjusted effect estimates with 95 per cent CI as well as covariates that were adjusted for in the multivariable analysis.

To ensure the accuracy, this data extraction process was independently performed by two investigators. The data abstraction forms were cross-checked by the senior investigator. Any data discrepancy was resolved by referring back to the original articles.

Statistical analysis: Data analysis was performed using the RevMan 5.3 software (Cochrane, London, UK). Adjusted point estimates for the association between ever-smoker status and advanced liver fibrosis were extracted from each study and combined together using the generic inverse variance method as described earlier16, to assign the weight of each study in the pooled analysis inversely.

Random-effects, rather than fixed-effects model, was utitlized for the meta-anlayses as the assumption of the latter that all studies, regardless of study design and participants, should produce the same result is almost always not true for clinical research. Cochran’s Q test and I2 statistic were used to determine the between-study heterogeneity. This I2 statistic quantified the proportion of total variation across studies due to true heterogeneity rather than chance. A value of I2 of 0-25 per cent represents insignificant, 26-50 per cent represents low, 51-75 per cent represents moderate heterogeneity and more than 75 per cent represents high heterogeneity, respectively17. If enough number of eligible studies were identified, visualization of funnel plot was used to assess for the presence of publication bias.

Results

Two-hundred and ninety-six potentially eligible articles were identified as per the described search strategy (107 from MEDLINE and 189 from Embase). After the exclusion of 105 duplicated articles, 191 articles underwent title and abstract review. A total of 170 articles were excluded at this stage since these did not fulfill the eligibility criteria based on type of article, study design, population or measured outcomes, leaving 21 articles for full-text review. Eighteen of these were excluded after the full-length review as these did not report the outcome of interest. Finally, three cross-sectional studies121314 with 544 participants were included in the meta-analysis. It should be noted that the study by Zein et al14 consisted of two cohorts that were recruited from different centers. The effect estimates for each cohort were reported separately and, therefore, were both included in the meta-analysis. The literature review and selection process are depicted in Fig. 1. The characteristics and quality assessment of these studies are detailed in the Table. In brief, all included studies diagnosed PBC based on clinical presentation, serology, and histopathology. The definition of ever-smokers was consistent across the studies (defined as current or history of smoking of ≥5 packs at any time during the patient’s lifetime up to the time of PBC diagnosis)121314.

Close

Fig. 1

Literature review process.

Export to PPT

Risk of advanced liver fibrosis among patients with primary biliary cholangitis (PBC) who were ever-smokers versus patients who were non-smokers: The pooled analysis found a significantly increased risk of advanced liver fibrosis among patients with PBC who were ever-smokers compared to patients who were non-smokers with the pooled OR of 3.00 (95% CI, 1.18-7.65) as shown in Fig. 2. Statistical heterogeneity was high with I2 of 89 per cent.

Close

Fig. 2

Forest plot of the meta-analysis.

Export to PPT

Discussion

As per our acknowledge, this study is the first systematic review and meta-analysis that summarizes all available data on the association between smoking status and risk of advanced liver fibrosis among patients with PBC. The pooled analysis found a three-fold increased risk of advanced liver fibrosis among patients with PBC who were ever-smokers compared to patients without history of tobacco exposure. The mechanism behind the increased risk is not known with certainty. Possible explanations are discussed below.

First, smoking has been shown to alter the balance of T helper cells, Th1 and Th2, and several cytokine levels, including IL-5 and IL-131819. IL-13 has been implicated in progression of fibrosis in animal studies and smoking can increase the production of IL-1311202122. It has been demonstrated that IL-5 can augment the progression of liver fibrosis by up-regulating activity of IL-1323. Th1 cells that are inducible by smoking24 have been shown to accelerate the progression of fibrosis by activating hepatic stellate cells to secrete more profibrogenic markers through the IFN-γ/STAT pathway25.

The second possible explanation involves pro-angiogenic factors. A study in patients with chronic hepatitis C virus infection found that the level of pro-angiogenic factors, such as vascular endothelial growth factor (VEGF) and VEGF-D, are higher among smokers compared to non-smokers and the higher level was independently associated with advanced fibrosis26. Tissue hypoxia induced by smoking is the likely cause of the higher level of these factors.

Third, smoking can lead to insulin resistance as demonstrated by Houston et al27 in the CARDIA study. Furthermore, studies have also demonstrated that insulin resistance is associated with higher prevalence of severe hepatic fibrosis in patients with NAFLD2728. Some other studies have suggested that hyperinsulinemia can stimulate influx of fatty acid to the liver, leading to hepatic triglyceride accumulation293031. This excessive fatty deposition will cause cellular injury through oxidative stress and hepatocyte apoptosis, which will eventually lead to hepatic fibrosis32.

The present study has some limitations that may affect the validity of the results. First, statistical heterogeneity was high in this meta-analysis. We believe that the difference in background populations and methods used to evaluate liver fibrosis were the main source of the between-study variation. In addition, there was variation in adjustment of the effect estimates as two studies1213 adjusted their effect estimates for sex and alcohol consumption while one study14 did not. Second, a formal assessment for the presence of publication bias could not be performed due to the limited number of included studies. Therefore, it is possible that publication bias in favour of studies that showed positive association may have been present and may have skewed the pooled result. Third, all of the studies were conducted in Western countries and the results may not be generalizable to other populations. Fourth, subgroup analysis comparing heavy, regular, occasional, and ex-smokers could not be performed compared to non-smokers as the included studies did no provide such data. Similarly, there was no subgroup data to perform subgroup analysis based on sex and age. Lastly, this was a systematic review and meta-analysis of observational studies. Thus, it is still possible that the observed association was not causal but was a function of a confounding effect. Other factors related to the smoking habit, but not smoking itself, could still be the actual etiology of the increased risk. This limitation is true for all observation studies but is of particular concern for the current study because only minimal to none adjustment for potential confounders was performed by the primary studies.

In summary, the current study demonstrated that smoking is associated with a significantly higher risk of advanced liver fibrosis among patients with PBC. Further prospective studies are required to determine whether this association is indeed causal.