Clinical pharmacy column
 
Incretin-based therapies for type 2 diabetes with nonalcoholic fatty liver disease: a systematic review and meta-analysis                      
Qin Hu1, Hulin Tang2*, Hong Shao1*
1. Department of Pharmacy Administration and Clinical Pharmacy, School of Pharmaceutical Sciences, Peking University Health Science Center, Beijing 100191, China
2. Department of Pharmacy, Peking University Third Hospital, Beijing 100191, China

 
Abstract: We conducted a systematic review and meta-analysis of randomized controlled trials (RCTs) to determine the effectiveness and safety of incretin-based therapies (IBTs) for the treatment of type 2 diabetes (T2DM) with nonalcoholic fatty liver disease (NAFLD). Electronic databases such as the Cochrane library, EMbase, PubMed, and three Chinese databases were searched for RCTs that compared IBTs with other treatments or placebo for T2DM with NAFLD. Two reviewers independently assessed the risk of bias, extracted, and analyzed the data. A meta-analysis was performed using Revman 5.2. Publication bias was evaluated. Seven RCTs involving 532 patients were ultimately included. The results of meta-analysis (random-effects model)revealed that IBTs had a significant reduction in serum ALT (WMD –12.30, 95% CI –17.53~–7.06) and BMI (WMD –2.64, 95% CI –4.35~–0.94). However, there was no significant difference in other outcomes including HbA1c, AST, TC, TG and HOMA-RA. IBTs were well tolerated by patients but the evidence was limited. The significant decrease in hepatic biochemical markers following treatment with IBTs, as well as improvements in BMI, suggested that IBTs may be an effective option for T2DM with NAFLD.        
Keywords: Incretin-based therapies, Dipeptidyl peptidase-4 inhibitor, Glucagon-like peptide-1 receptor agonist, Type 2 diabetes, Nonalcoholic fatty liver disease, Meta-analysis, Randomized controlled trial 
CLC number: R977                Document code: A                 Article ID: 10031057(2016)320609
 
 
1. Introduction
The rate of Diabetes Mellitus (DM) is still on the rise around the world. According to epidemiological estimates, the prevalence of DM will increase from 285 million in 2010 to 439 million in 2030[1]. According to data from the Edinburgh type 2 diabetes research center, nearly 50% of patients with type 2 diabetes mellitus (T2DMs) suffer from non-alcoholic liver disease (NAFLD)[2], and its prevalence in the Asia-Pacific region is about 28%–55%[3]. NAFLD is a type of metabolic irritable liver injury, which is associated with genetic factors (adiponutrin C3), insulin resistance (IR), adipose cytokinesdisorder, and so on[4]. The spectrum of the disease includes isolated nonalcoholic simple fatty liver (NAFL), nonalcoholic steatohepatitis (NASH), liver cirrhosis, and hepatocellular carcinoma (HCC)[5]. Although the pathogenesis has not been fully elucidated, insulin resistance is considered to be the main characteristic of NAFLD, which independently increases the risk of both T2DM and cardiovascular disease. A primary objective for NAFLD treatment is prevention and treatment of the metabolic syndrome and associated complications such as glycemic control, which can be treated with hypoglycemic agents, so as to improve quality of life and prolong survival. Insulin sensitizers (metformin, pioglitazone, rosiglitazone) may also be effective in controlling glucose and lipid metabolism[6].
Recently, incretin-based therapies (IBTs) have appeared as a new class of oral hypoglycemic drugs, including dipeptidyl peptidase-4 inhibitor (DPP-4i) and glucagon-like peptide-1 receptor agonist (GLP-1RA). The first approved IBTs was exenatide, a GLP-1RA. It was approved in April 2005 by the FDA and in November 2006 by the EMA. Afterwards, sitagliptin, the first DPP-4i, was also approved by the FDA and EMA in October 2006 and March 2007, respectively. By 2015, a total of 7 DPP-4is and 6 GLP-1RAs were approved for use in treating T2DM. IBTs plus insulin is a superior option for treatment intensification in T2DM compared with insulin alone[7], and IBTs have showed promising effects on improving serum enzyme levels and histological lesions in patients with NAFLD[8]. A systematic review and meta-analysis carried out by Carbone et al showed a significant decline in ALT following treatment with IBTs in patients with NAFLD[9]. In addition, a retrospective study by Ohki reported that liraglutide and sitagliptin could effectively control blood glucose levels as well as improve liver inflammation and fibrosis[10]. However, it is unclear whether these agents can be effective options for these patients, and the role of IBTs on T2DM with comorbidities such as NAFLD needs further study. Therefore, our aim was to examine the effectiveness and safety of IBTs for the treatment of T2DM with NAFLD using data from randomized controlled trials (RCTs).
2. Materials and methods
A systematic review of published works was performedaccording to the methods of Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).
2.1. Search strategy
Published articles were systematically searched (starting from inception to May 31, 2015) from The Cochrane Library, EMbase, PubMed, and three Chinese literature databases (China National Knowledge Infrastructure (CNKI), Wan Fang and Chinese biomedical database) by using the following research terms: (“dipeptidyl peptidase-4 inhibitors” OR “DPP-4i” OR “gliptins” ORsitagliptin” OR “saxagliptin” OR “linagliptin” OR “alogliptin” OR “vildagliptin” OR “dutogliptin” OR “omarigliptin” OR “glucagon-like-peptide-1 receptor agonists” OR “GLP-1 RAs” OR “exenatide” OR “liraglutide” OR “dulaglutide” OR “albiglutide” OR “lixisenatide” OR “semaglutide”) and (“non*alcoholic fatty liver” OR “nonalcoholic fatty liver” OR “nonalcoholic fatty liver” OR “NAFL*” OR “NASH” OR (non*alcoholicAND steato*hepatitis) OR (nonalcoholic AND steato* hepatitis) OR (nonalcoholic AND steato*hepatitis)). In addition, the references for the included RCTs and the review were also manually searched for additional potential RCTs. No language, sample size, or publication status restrictions were applied to the search.
2.2. Study selection
Only RCTs that reported the effectiveness and safety outcomes of patients treated with IBTs for T2DM with NAFLD were eligible for our meta-analysis. The study selection was carried out according to the following inclusion and exclusion criteria: (1) Subjects: The subjectswere diagnosed with DM according to Diagnostic Criteria for Diabetes (1999, WHO) and diagnosed with isolated fatty infiltration of hepatocytes (steatosis), nonalcoholic simple fatty liver (NAFL), nonalcoholic steatohepatitis (NASH), liver cirrhosis, and/or NAFLD by the Practice Guidelines of NAFLD (2006 or 2010, China; 2007, APAC; 2010, Italy; 2012, U.S.A)[1115]. Due to the high prevalence of undiagnosed NAFLD in T2DM, an existing clinical diagnosis of NAFLD was not necessary for inclusion[16]. No gender, age, or ethnicity restrictions were applied to the subjects of study. (2) Intervention and comparison: Monotherapy or combination therapy with IBTs as an intervention group was compared to blank, placebo or active treatment as a control group. In addition, the background therapy was similar between the intervention and control groups. (3) Outcomes of interest: the primary outcomes included glycosylated hemoglobin, HbA1c%; index of insulin resistance (HOMA-IR); safety outcomes and alanine aminotransferase (ALT). The secondary outcomes included aspertate aminotransferase (AST); body mass index (BMI); triglycerides (TGs) and total cholesterol (TC). The exclusion criteria was considered as follows: a clear history of hepatitis B, hepatitis C, or other viral hepatitis, autoimmune liver disease, metabolic liver disease, hereditary liver disease and drug-induced liver injury; alcohol intake>140 g/week (female>70 g/week); patients with heart disease, renal insufficiency, history of malignant tumors, hemodynamic instability and pregnant women; patients taking lipid-lowering or weight-reducing drugs; patients with different background treatments. In addition, non-RCTs, reviews, retrospective studies, and case reports were excluded.
2.3. Data extraction and risk of bias assessment
According to the pre-design data extraction table, two reviewers (Q.H and H.T) independently extracted relevant information and assessed the risk of bias. The extracted data included the characteristics of each study (author, years), population (numbers of patients, diagnosis), drug regimens, time of follow up, and clinical outcomes of the two groups in each study. Otherwise, the risks of bias of the included RCTs were assessed according to Cochrane risk of bias tool[17]. The following 7 domains were judged as “Low”, “Unclear” and “High”: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other biases. Any discrepancies were resolved by discussion or consulting with a third review author (H.S) if required.
2.4. Statistical analysis
The meta-analysis was performed using Review Manager for Windows (version 5.2). The Weight Mean Difference (WMD) and odds ratio (OR) with 95% confidenceinterval (CI) were calculated for continuous and dichotomous variables, respectively. Statistical heterogeneity across studies was assessed by χ2 test or I2 test. P values less than 0.05 or I2 values more than 50% indicated significant heterogeneity. A random effects model was used throughout all the meta-analyses for the heterogeneity across the RCTs. A subgroup meta-analysis was performed and stratified by DPP-4i and GLP-1RA respectively. Adverse events were directly examined by qualitative analysis instead of quantitative analysis for a few sample sizes. Publication bias was evaluated using funnel plot.
3. Results
3.1. Study selection and study characteristics
Three hundred and forty-four citations were retrieved according to the search strategy. Only 67 articles were considered as potential RCTs upon retrieval of full texts for further evaluation. After excluding the duplications and reviewing the titles and abstracts, 7 RCTs were included. The search flowcharts and study selection with exclusion criteria are summarized in Figure 1.


Figure 1.
Flow chart of the identification of eligible studies.
 
All seven RCTs were published in 2014[1824]. A summary of the included RCTs is presented in Table 1. Of the seven RCTs, 4 RCTs evaluated the GLP-1 RA (exenatide or liraglutide), and 3 RCTs evaluated the DPP-4i (sitagliptin or linagliptin). A total of 532 T2DMpatients with NAFLD were involved. There were 264 in the intervention groups and 268 in the control groups. The patients ranged from 45 to 65 years of age. The period of the treatment was between 8 weeks and 16 weeks. Conventional treatment regimens was permitted to continue in addition to GLP-1 RA or DPP-4i therapy.  

Table 1. The characteristics of included studies

T:
treatment group; C: controlled group; NR: not reported; Outcomes: HbA1c%; HOMA-IR; ADR rate; ALT, alanineaminotransferase; AST, aspartateaminotransferase; BMI; TG, triglycerides;TC, totalcholesterol. 
3.2. Risks of bias of included studies
Seven domains were used to evaluate the risk of bias in the 7 included RCTs according to the Cochrane risk of bias method. Only 2 RCTs described the method of sequence generation and incomplete outcome data. The allocation concealment and blinding of the 7 RCTs were determined to be unclear because no detailed information about these two domains were mentioned. None of the trials used intention-to-treat (ITT) analysis by imputing the missing data, and little detailed information could be obtained from the original authors. Overall, the risk of bias of the RCTs included in the meta-analysis was high or unclear.
3.3. Quantitative synthesis
3.3.1. HbA1c
Five RCTs reported the HbA1c (%) level changes before and after treatment[1821,24]. There was no significant difference between IBTs and other treatments (WMD –0.24, 95% CI –0.65~0.17, P = 0.26), but significant heterogeneity was found (I2 = 75%, P = 0.003). Considering the heterogeneity between DPP-4i and GLP-1RA interventions, a subgroup meta-analysis was performed for GLP-1RA and DPP-4i. The results of the sub-analysis showed that no significant difference in HbA1c (%) reduction was observed between the GLP-1RA groups and the control groups (WMD –0.05, 95% CI –0.23~0.12, P = 0.60) and between the DPP-4i groups and the control groups (WMD –0.40, 95% CI –1.55~0.75, P = 0.68). The results are present in Figure 2.


Figure 2.
Forest plot for meta-analysis of mean difference between IBTs and other treatments in terms of change in HbA1c. IBTs, incretin-based therapies; CI, confidence interval; df, degrees of freedom; HbA1c, glycated hemoglobin; IV, inverse variance; SD, standard deviation. 
3.3.2. ALT
Six RCTs provided the outcome of ALT level changesbefore and after treatment[1821,23,24]. There was a significant difference between IBTs and other treatments (WMD –12.30, 95% CI –17.53~–7.06, P<0.00001), with significant heterogeneity (I2 = 85%, P<0.00001). The results of the subgroup meta-analysis for DPP-4i and GLP-1RA showed that a significant difference in ALT reduction was observed between the GLP-1RA groups and the control groups (WMD –17.99, 95% CI –26.76~–9.22, P<0.00001), but there was no significant difference between the DPP-4i groups and the control groups (WMD –6.90, 95% CI –16.25~2.44, P = 0.15). The results are presented in Figure 3.  


Figure 3.
Forest plot for meta-analysis of mean difference between IBTs and other treatments in terms of change in ALT. IBTs, incretin-based therapies; CI, confidence interval; df, degrees of freedom; ALT, alanine aminotransferase; IV, inverse variance; SD, standard deviation.  
3.3.3. AST
Six RCTs reported the outcome of AST level changesbefore and after treatment[1821,23,24]. There was no significant difference between IBTs and other treatments (WMD –10.90, 95% CI –21.97~0.17, P = 0.05), but significant heterogeneity was detected (I2 = 98%, P<0.00001). The subgroup meta-analysis performed for GLP-1RA and DPP-4i showed that a significant difference in AST reduction was observed between the GLP-1RA groups and the control groups (WMD –7.16, 95% CI –12.88~–1.44, P = 0.01), but there was no significant difference between the DPP-4i groups and the control groups (WMD –14.07, 95% CI –29.13~0.99, P = 0.07).
3.3.4. TG
Five RCTs reported TG changes before and after treatment[1821,23]. There was no significant difference between IBTs and other treatments (WMD –0.08, 95% CI –0.20~0.04, P = 0.20) without significant heterogeneity (I2 = 0%, P = 0.49). In addition, a subgroup meta-analysis for GLP-1RA and DPP-4i showed that no significant differences in TG reduction was observed between the GLP-1RA groups and the control groups (WMD –0.05, 95% CI –0.18~0.07, P = 0.39) and between the DPP-4i groups and the control groups (WMD –0.44, 95% CI –0.93~0.04, P = 0.07).
3.3.5. TC
Five RCTs reported TC changes before and after treatment[1821,23]. There was no significant difference between IBTs and other treatments (WMD –0.03, 95% CI –0.19~0.14, P = 0.77) without significant heterogeneity (I2 = 0%, P = 0.98). The results of the subgroup meta-analysis for GLP-1RA and DPP-4i showed that no significant difference in TC reduction was observed between the GLP-1RA groups and the control groups (WMD –0.04, 95% CI –0.22~0.14, P = 0.67), and between the DPP-4i groups and the control groups (WMD 0.10, 95% CI –0.44~0.64, P = 0.72).
3.3.6. BMI
Six RCTs reported BMI changes before and after treatment[1822,24]. IBT could significantly reduce the BMI than other treatments (WMD –2.64, 95% CI –4.35~–0.94, P = 0.002), and no significant heterogeneity was detected(I2 = 97%, P<0.00001). In addition, a subgroup meta-analysis for GLP-1RA and DPP-4i showed that there was significant difference in BMI reduction between the GLP-1RA groups and the control groups (WMD –3.50, 95% CI –6.43~–0.56, P = 0.02), and between the DPP-4igroups and the control groups (WMD –1.85, 95% CI –2.70~–1.01, P<0.00001).The results are presented in Figure 4. 


Figure 4.
Forest plot for meta–analysis of mean difference between IBTs and other treatments in terms of change in BMI. IBTs, incretin-based therapies; CI, confidence interval; df, degrees of freedom; BMI, body mass index; IV, inverse variance; SD, standard deviation.  
3.3.7. HOMA-IR
Only two RCTs reported the outcomes of HOMA-IR changes before and after treatment[18,20]. The results of the study by Fan et al. showed that there was no significant difference between GLP-1RA and metformin in HOMA-IR reduction (WMD 0.04, 95% CI –0.11~0.19)[18].The RCT carried out by Song et al. reported that there was a significant difference between DPP-4i and glipizide in HOMA-IR reduction (WMD –0.70, 95% CI –1.25~–0.15).
3.4. Safety outcomes
Only one study reported adverse drug reaction (ADR)rates in the IBT treatment and control groups[20]. Five studies provided information on adverse effects at treatment[1822]. Four studies described gastrointestinal events, including nausea, vomiting, diarrhea, and abdominal pain[1820,22]. There exists a higher incidence of gastrointestinal side effects in the IBT group compared to the insulin treatment group[19]. Gastrointestinal side effects were found 2–3 weeks after treatment[18]. Two studies evaluated the incidence of hypoglycemia with consistent results, and both RCTs showed that the risk of hypoglycemia was significant lower in the IBTs group than in the insulin group. None of participants withdrew due to adverse effects[19,21].
3.5. Publication bias
A selectivity funnel plot was generated to assess the likelihood of publication bias, which showed that the primary outcome (mean ALT reductions) in the six groups were nearly symmetrically distributed around the synthesis estimate.
4. Discussion
4.1. Summary of evidence
Previous meta-analysis was performed by Carbone et al[9]and included retrospective studies and non-controlled tests, which only evaluated the outcome of ALT level in patients with NAFLD with or without definite diagnose of T2DM. Our meta-analysis used glycemic control measurements (HbA1c), hepatic parameters (ALT and AST), and metabolic parameters (TG, TC and BMI) as primary and secondary outcomes,and we comprehensively evaluated the clinical effectiveness and safety of the IBTs for the treatment of T2DM with NAFLD. The meta-analysis showed that for the primary outcome, a significant decline in serum ALT following treatment with IBTs, especially GLP-1 RA, could better reduce hepatic enzyme levels (ALT and AST) compared to other treatments (metformin, glipizide and placebo). There was no significant difference in the change in blood glucose index (HbA1c) between IBTs and other treatments. The metabolic indices (TG and TC) were not found to be significantly different between IBTs and others. With respect to BMI improvement, there was a significant difference between IBTs and other treatments: GLP-1 RA was better than insulin and metformin, whereas DPP-4i was similar to the control. In addition, the existing data showed that IBTs were well tolerated.
4.2. Overall completeness and applicability of evidence
There is an increasing incidence of diabetes and NAFLD with elevated liver enzymes. Because most of the oral hypoglycemic agents (biguanides, sulphonylurea, thiazolidines) are metabolized in the liver and often result in liver damage, coexistent disease is more difficultto treat compared to just the single disease and NAFLD is widely found in the T2DM population[19]. As a novel type of anti-diabetes drug, IBTs were approved for T2DM with little evidence about the therapeutic effect on hepatic insufficiency.
Recently, some research has demonstrated that IBTs were somewhat effective in liver function recovery. The LEAD project (Liraglutide Effect and Action in Diabetes) had carried out the second project studying the efficacy of liraglutide for treating fatty liver illnesses. The results by Armstrong et al.indicated that 1.8 mg liraglutide per day was not only safe but could also improve hepatic enzyme elevation in patients with T2DM[25]. A systematic review, including 15 RCTs, demonstrated that lixisenatide could increase the proportion of T2DM patients with normal ALT[26]. In addition, research by Blaslov et al. found that exenatidecould improve the amount of fatty liver and liver biochemical indicators in patients with T2DM[27].
The mechanism for GLP-1 RA acting on liver lipid metabolism is still unsolved. One study suggested that it related to weight loss[28], whereas another study noted that patients gained weight after hypoglycemic therapy and intrahepatic lipid (IHL) levels were reduced[29].The study by Cuthbertson et al. examined the IHL levels in 25 patients with T2DM after 6 months of therapy with GLP-1 RA (including exenatide and liraglutide), and the results showed that the change of IHL corresponded with HbA1c, but was not relevant to weight, abdominal subcutaneous adipose tissue (SAT), and visceral adipose tissue (VAT)[16]. 
NAFLD is closely associated with metabolic syndromes that are complicated by hypertension, hyperlipidemia, obesity, and insulin resistance[9,30]. Obesity is also a prevalent complication in T2DM[31], and certain oral hypoglycemic drugs, such as metformin, also have the advantage of promoting weight reduction. Our meta-analysis revealed that the improvement in BMI followed by IBTs could have some metabolic action that may be helpful in treating T2DM with NAFLD.
For the patients with NAFLD and potential risk of diabetes, IBTs could treat hepatic insufficiency and controlblood glucose at the same time. A prospective trial involved 27 NAFLD patients with glucose intolerance indicated that 0.9 mg of liraglutide per day was well tolerated and affected liver function as shown by biochemical and histological metrics[32]. The study by Zhang et al. followed 200 patients with pre-diabetes and NAFLD, and it found that sitagliptin led to an improvement in blood glucose levels, BMI, and pancreatic function greater than that of conventional therapy. Moreover, the progression to diabetes was slowed, and NAFLD was markedly relieved[33]. Our meta-analysis only involved T2DM patients with NAFLD so as to obtain pertinence results for the targeted population. Based on our results, IBTs might be useful in clinical practice for T2DM patients with NAFLD.
4.3. Limitations
RCTs  included in our study were of poor quality with risk of bias, therefore evidence reliability was limited. We attempted to make our conclusions relatively reliable, but there were still some confounding factors. There was some heterogeneity across RCTs, which was partly decreased in our subgroup, and a random effect model was used. However, other sources of heterogeneity including inconsistent medication dosage and duration, inconsistent diagnoses of NAFLD within these studies, as well as differing disease statuses can lead to misclassification bias. In addition, small sample trials and poor methodological quality might cause bias. These results should be interpreted with caution. Therefore, multi-center, large sample, random, double blind, controlled clinical trials of IBTs for T2DM with NAFLD are still needed to verify our research results and to provide more credible evidence for clinical drug use.
5. Conclusions
The significant reduction in hepatic biochemical markers following treatment with IBTs, as well as improvements in BMI, suggested that IBTs may be appropriate for treating T2DM with obesity and NAFLDwith elevated liver enzymes. However, large, well-designed, and well-executed RCTs that involve more clinical endpoints are necessary to evaluate the role of IBTs in T2DM patients with NAFLD.
Acknowledgements
We thank all authors that contributed data or additional information about their studies. The authors have disclosed no relevant financial relationships.
References
[1] Shaw, J.E.; Sicree, R.A.; Zimmet, P.Z. Diabetes Res. Clin. Pract.2010, 87, 4–14.
[2] Blaslov, K.; Zibar, K.; Bulum, T.; Duvnjak, L. Clin. Res. Hepatol. Gastroenterol. 2014, 38, 61–63.
[3] Fan, J.G.; Saibara, T.; Chitturi, S.; Kim, B.I.; Sung, J.J. J. Gastroenterol. Hepatol. 2007, 22, 794–800.
[4] de Alwis, N.M.; Day, C.P. J. Hepatol.2008, 48, 104–112.
[5] Farrell, G.C.; Larter, C.Z. Hepatology.2006, 43, 99–112.
[6] Ali, R.; Cusi, K. Ann. Med.2009, 41, 265–278.
[7] Charbonnel, B.; Bertolini, M.; Tinahones, F.J.; Domingo, M.P.; Davies, M. J. Diabetes. Complications.2014, 28, 880–886.
[8] Olaywi, M.; Bhatia, T.; Anand, S.; Singhal, S. Hepatobiliary Pancreat. Dis. Int. 2013, 12, 584–588.
[9] Carbone, L.J.; Angus, P.W.; Yeomans, N.D. J. Gastroenterol. Hepatol.2015, 327329.
[10] Ohki, T.; Isogawa, A.; Iwamoto, M.; Ohsuqi, M.; Yoshida, H. Sci. World J.2012, 496–453. 
[11] Fatty Liver and Alcoholic Liver Disease Study Group of the Chinese Liver Disease Association Guidelines for diagnosis and treatment of nonalcoholic fatty liver disease.Chin. J. Hepat. 2006, 14, 161–163.
[12] Chalasani, N.; Younossi, Z.; Lavine, J.E.; Diehl, A.M.; Brunt, E.M. Gastroenterology.2012, 142, 1592–1609.
[13] Loria, P.; Adinolfi, L.E.; Bellentani, S.; Bugianesi, E.; Grieco, A. Dig. Liver Dis.2010, 42, 272–282.
[14] Fan, J.G.; Cai, X.B. Drug Evaluation. 2007, 4, 80–81.
[15] Fatty Liver and Alcoholic Liver Disease Study Group of the Chinese Liver Disease Association Guidelines for diagnosis and treatment of nonalcoholic fatty liver disease. Chin. J. Hepatol. (electronic edition). 2010, 2, 43–48.
[16] Cuthbertson, D.J.; Irwin, A.; Gardner, C.J.; Daousi, C.; Purewal, T. PLoS One. 2012, 7, e50117.
[17] Higgins JPT GSE Cochrane Handbook for Systematic Systematic Reviews of Interventions Version 5.1.0. The Cochrane Collaboration. Available: www.cochranehand-book.org. 2011.
[18] Fan, H.; Pan, Q.; Xu, Y.; Yang, X. Arq. Bras. Endocrinol. Metabol. 2014, 57, 702–708.
[19] Shao, N.; Kuang, H.Y.; Hao, M.; Gao, X.Y.; Lin, W.J. Diabetes Metab. Res. Rev. 2014, 30, 521–529.
[20] Song, X.X.; Jiang, T.; Kang, K.; Wen, Z. Chin. J. New Drug.2014, 23, 215–218.
[21] Bao, M.J.; Wu, J.C.; Yang, Y.; Zhang, L.; Zhu, X.J. Pract. J. Clin. Med.2014, 11, 70–72.
[22] Jin, H.L.; Zhao, H.Y.; Li, Y.H.; Jia, S.E.; Ma, L.N. Harbin Med. J.2014, 34, 279–281.
[23] Liang, L.F.; Xu, M.; Zhang, H.L.; Yang, H.; Zhang, E.S. J. Front. Med. 2014, 220–221.
[24] Wang, Y.; Dai, Y. China Pharma.2014, 23, 107–108.
[25] Armstrong, M.J.; Houlihan, D.D.; Rowe, I.A.; Clausen, W.H.; Elbrønd, B. Aliment. Pharmacol. Ther. 2013, 37, 234–242.
[26] Gluud, L.L.; Knop, F.K.; Vilsboll, T. Bmj. Open. 2014, 4, e005325.
[27] Blaslov, K.; Zibar, K.; Bulum, T.; Duvnjak, L. Clin. Res. Hepatol. Gastroenterol. 2014, 38, e61–63.
[28] Petersen, K.F.; Dufour, S.; Befroy, D.; Lehrke, M.; Hendler, R.E. Diabetes.2005, 54, 603–608.
[29] Belfort, R.; Harrison, S.A.; Brown, K.; Darland, C.; Finch, J. N. Engl. J. Med.2006, 355, 2297–2307.
[30] Karlas, T.; Wiegand, J.; Berg, T. Best Pract. Res. Clin. Endocrinol. Metab.2013, 27, 195–208.
[31] Xie, X.H.; Wang, F.; Cui, J.Y. J.Chin. Pharm. Sci. 2015, 24, 412–418.
[32] Eguchi, Y.; Kitajima, Y.; Hyogo, H.; Takahashi, H.; Kojima, M. Hepatol. Res. 2015, 45, 269–278.
[33] Zhang, L.X.; Wei, C.; Wei, X.D.; Liu, Y.; Chen, D.H. Chin. J. Modern Drug Appl. 2015, 9, 120–122.
  
 
  
基于肠促胰素药物治疗2型糖尿病合并非酒精性肝病的疗效和安全性的系统评价和meta分析
胡琴1, 唐惠林2*, 邵宏1*
1. 北京大学医学部药学院药事管理与临床药学系,北京100191
2. 北京大学第三医院药剂科,北京100191     
摘要: 本文旨在对基于肠促胰素药物治疗2型糖尿病合并非酒精性肝病的疗效和安全性进行系统评价和meta分析。全面检索Cochrane LibraryEMBASEPubMedCNKI、中国生物医学文献数据库CBM和万方中华医学会期刊数据库,纳入20157月之前发表的全部随机对照试验(RCT)。共纳入7RCT532名患者(干预组: 264,对照组: 268)。肠促胰素组显著降低患者谷丙转氨酶ALT (WMD12.30, 95% CI 17.53~7.06)BMI (WMD2.64, 95% CI 4.35~0.94)。两组对患者血红蛋白(HbA1c%)、天门冬氨酸转移酶(AST)、总胆固醇(TC)、甘油三酯(TG)和胰岛素抵抗指数(HOMA-RA)的改变无显著差别。有限的安全性资料显示不良反应可耐受。基于肠促胰素药物对ALTBMI的改善效果较好,安全性尚可,可作为2型糖尿病合并非酒精性肝病患者的治疗选择,但研究质量有限。 
关键词: 肠促胰素; DPP-4酶抑制剂; GLP-1受体拮抗剂; 2型糖尿病;非酒精性肝病; meta分析;随机对照试验
 
 
 
 
Received: 2015-11-14, Revised: 2015-12-28, Accepted: 2016-01-18.
*Corresponding author. Tel.: 13671391113, E-mail: hltang1985@bjmu.edu.cn, h_shao@163.com