Effects of Inulin on the Plasma Lipid Profile of Normolipidemic and Hyperlipidemic Subjects

A Meta-analysis of Randomized Controlled Trials

Zhuang Guo; Xiao-Ming Liu; Qiu-Xiang Zhang; Feng-Wei Tian; Hao Zhang; He-Ping Zhang; Wei Chen


Clin Lipidology. 2012;7(2):215-222. 

In This Article


Study Selection

A search was conduced for studies that focused on the efficacy of inulin on the plasma lipid profile of subjects with hyperlipidemia (total serum cholesterol >200 mg/dl) or normolipidemia in online databases Embase, Web of Science, PubMed and Cochrane Controlled Trials Registry. The following headings and keywords were combined using the following Boolean operators: (inulin OR dahlin) AND (cholesterol OR HDL OR LDL OR triglycerides OR plasma lipids OR serum lipids). Each database was searched from inception to June 2011. Moreover, a manual search of references from reviews and research articles was performed to identify relevant clinical trials; no language restriction was applied.

Articles were eligible for meta-analysis if they fulfilled the following criteria:

  • The study design was a randomized control trial;

  • The mean total cholesterol, LDL-C, HDL-C and triglyceride concentration changes, along with standard deviation (SD), were reported for the intervention and control groups;

  • The products were enriched with inulin only and did not contain probiotic strains or other prebiotics;

  • The subjects had not undergone intestinal surgery.

Data Abstraction

Study characteristics, which included authors, publication year, study design (crossover or parallel), duration of intervention, sample size, subject characteristics (ratio of male and female subjects, BMI and age) and dosage of inulin, were extracted for quantitative and qualitative synthesis. In addition, the vehicle of inulin in the intervention was recorded; namely, whether subjects received inulin as a substitute for a portion of the sugar content of the study foods (biscuit, ice cream or breakfast cereals) or supplemental inulin powders as sweetener. Two investigators evaluated the articles for eligibility and abstracted the data independently. Any disagreement among the investigators in data abstraction was resolved by discussion. Mean study baseline and end points for the inulin and control diets were recorded for total cholesterol, HDL-C, LDL-C and triglycerides. These values were recorded in the form of mean ± SD or mean ± standard error (SE).

Statistical Analysis

Some articles reported lipid levels in mmol/l, which required conversion to mg/dl before meta-analysis. The conversion factor was 1 mmol/l = 38.61 mg/dl for cholesterol and 1 mmol/l = 88.50 mg/dl for triglycerides. The mean net changes (mean values ± SD) in the total cholesterol, HDL-C, LDL-C and triglycerides for each study were calculated. For parallel and crossover trials, net changes in the measures were calculated as follows: (measure at the end of follow-up in the intervention group – measure at baseline in the intervention group) – (measure at the end of follow-up in the control group – measure at baseline in the control group).

Variance in the mean net changes in the outcomes of each trial was calculated using SDs. If the SDs of the net changes in the measurements were not calculated, they could be calculated as follows (Equation 1):

N1 refers to the sample number of the control group, SD1 refers to the SD of the control group, N2 refers to the sample number of the trial group and SD2 refers to the SD of the trial group.

The amount of variation within the values was indicated as SEs in some articles, which required conversion to SDs as follows (Equation 2):

N refers to sample size.

We used the Review Manager 4.2 (The Cochrane Collaboration, Oxford, UK) to calculate a weighted mean difference (WMD) by conducting a random-effects meta-analysis. The heterogeneity of effect size among studies was tested; p < 0.05 was considered statistically significant for all analyses. In addition, to assess the potential for publication bias, funnel plots for each outcome were also examined.