Criteria for considering studies for this review
Types of studies
We will include randomised controlled trials (RCTs) or cluster-RCTs. We will also include controlled clinical trials (CCTs) but conduct a sensitivity analysis including only RCTs.
Types of participants
All persons 18 years of age or older having a clinical diagnosis of fibromyalgia, as defined by the 1990 American College of Rheumatology (ACR) guidelines (Wolfe 1990).
Types of interventions
We will include studies with at least one treatment group in which any probiotic intervention as a mono- or mixed culture of live micro-organisms was applied to humans with fibromyalgia. We will compare the intervention with placebo, usual diet and other mind and body therapies.
Types of outcome measures
The primary outcome measures will be:
Pain (to include self-reported levels of pain measured by validated pain assessment)
Functional status (ability to complete everyday tasks, e.g. scores on the Fibromyalgia Impact Questionnaire) (Bennett 2005)
We will assess all outcome measures at the beginning (baseline), three and six months for short and medium term studies respectively, and at the end of the studies, ranging from one to twelve months.
Secondary outcome measures will include:
Quality of life
Tender point score (measured by dolorimetry or digital palpation)
Mood (depression, anxiety)
Number of withdrawals due to adverse events
Serious adverse events
In order to improve endpoint outcome measurements, we will use the OMERACT 9 (Choy 2009) core set of outcome measures (pain, physical function, patient global assessment, joint imaging, health-related quality of life measure, physician global assessment, tenderness).
Search methods for identification of studies
We report the search strategy developed for MEDLINE in Appendix 1. We will search the following electronic databases:
The Cochrane Central Register of Controlled Trials (CENTRAL) via OVID (The Cochrane Library, current issue)
MEDLINE via OVID (1948 to present)
CINAHL via EBSCOhost (1982 to present)
AGRICOLA via OVID (1970 to present)
Food Science Technology Abstracts via OVID (1969 to present)
Searching other resources
We will search the reference lists of all the included studies for additional articles. We will contact experts in the field to identify any other unpublished or published studies. We will identify unpublished or ongoing trials by searching clinical trials registers.
Data collection and analysis
Selection of studies
Two review authors (DKF and IP) will independently carry out selection of papers and decisions about eligibility. DKF is a nutritionist. If the relevance of a report is unclear, the full text will be reviewed, and all disagreements will be resolved by discussion and consensus with the review author team. Study authors will be contacted for clarification and to obtain additional data in order to perform the systematic review, whenever necessary. Studies will be translated into English when necessary.
Data extraction and management
At least two independent authors will extract relevant data from selected studies according to inclusion criteria. Disagreements will be resolved by discussion and consensus. Any issue which cannot be resolved by the review authors will be discussed by the entire review team. In the case of missing data or ambiguity, we will contact the authors of the original article for clarification. We will consult a statistician in cases of doubt about data extraction and analysis.
For each study, we will extract the patient and study characteristics (diagnosis), intervention and outcomes data. We will extract the raw data (means and standard deviations for continuous outcomes and number of events and participants for dichotomous outcomes) for outcomes of interest. We will pilot test and use predefined data extraction forms to collect data.
We will extract and enter the following data into a customised collection form:
1. Study ID number
2. Study design, date and duration of study; setting of the study
a) sample size
b) inclusion and exclusion criteria
c) demographic characteristics of participants: age, sex, country of origin, gender, co-morbidities
4. Intervention - details of the experimental and comparison interventions
Other data that we will extract include:
Assessment of risk of bias in included studies
Two independent authors will assess and record the risk of bias of included studies using the Cochrane Collaboration's tool for assessing risk of bias. The tool addresses six specific domains of bias:
Selection bias (random sequence generation, allocation concealment);
Performance bias (blinding of participants and personnel);
Detection bias (blinding of outcome assessment);
Attrition bias (incomplete outcome data);
Reporting bias (selective outcome reporting); and
For other bias, we will consider specifically contamination where control or placebo group intentionally or unintentionally received the intervention probiotics. Two review authors will independently assess the risk of bias for all included studies. In case of lack of important study information, we will contact authors using open-ended questions to obtain the information needed. To determine the risk of bias of each study, for each criterion we will evaluate the presence of sufficient information and the likelihood of potential bias. The judgement for each criterion will be assessed as 'Low risk', High risk', or 'Unclear risk' of bias. In a consensus meeting, disagreements among the review authors will be discussed and resolved. If consensus cannot be reached, a third review author will make the final decision.
Measures of treatment effect
The data extracted from the studies will be entered into Review Manager 5 (Review Manager 2011). We will complete a summary table describing the study characteristics. For continuous data, we will calculate the mean difference in end point scores between groups using the same self-report questionnaires (such as the Fibromyalgia Impact Questionnaire) with 95% confidence intervals. We will calculate standardised mean differences when different scales are used to measure end point scores. For dichotomous data, we will compile a 2x2 contingency table including the number of participants with each outcome event and risk ratios (RR) with 95% confidence intervals.
Unit of analysis issues
For cluster-randomised trials (cluster-RCTs) we will follow the methods for adjusting for clustering as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). If a cluster-RCT has properly accounted for the cluster design it can be included in a meta-analysis by simply using the effect estimate and its standard error (SE) and using the generic inverse variance method in Review Manager 5. If an appropriate analysis has not been performed, we will incorporate the cluster-RCT into a meta-analysis (if relevant) by using an 'approximate method'.
The approximately correct analyses will only be performed if the following information can be extracted:
the number of clusters (or groups) randomised to each intervention group; or the average (mean) size of each cluster;
the outcome data ignoring the cluster design for the total number of individuals (for example, number or proportion of individuals with events, or means and standard deviations); and
an estimate of the intracluster (or intraclass) correlation coefficient (ICC).
The aim of the approximately correct analysis is to reduce the size of each trial to its "effective sample size" (Rao 1992). The effective sample size of a single intervention group in a cluster-randomised trial is its original sample size divided by a quantity called the ‘design effect’. The design effect is: 1 + (M – 1) ICC, where M is the average cluster size and ICC is the intracluster correlation coefficient. A common design effect is usually assumed across intervention groups. For dichotomous data both the number of participants and the number experiencing the event will be divided by the same design effect. We will use the generic inverse variance method (that uses the effect estimates and standard error). For continuous data only the sample size need be reduced; means and standard deviations should remain unchanged.
Studies with multiple treatment groups
For studies where there are multiple treatment groups, we will combine groups to create a single pairwise comparison as recommended by the Cochrane Handbook (16.5.4) (Higgins 2011). We will combine all relevant experimental intervention groups (probiotics or probiotics and food) of the study into a single group, and all relevant control intervention groups (placebo or treatment as usual) into a single control group. For dichotomous outcomes, both the sample sizes and the numbers of people with events will be summed across groups. For continuous outcomes, we will combine means and standard deviations using the formula specified in Appendix 2, according to the Cochrane Handbook (22.214.171.124) (Higgins 2011).
Multiple time points
We will report data at all collection time points during and after the intervention period (follow-up). Where data allow, we plan to group time points as follows: ≤ 3 months of probiotic treatment, 3 to 6 months of probiotic treatment and > 6 months of probiotics treatment.
Dealing with missing data
Where change scores are not available we will seek them from the authors. Failing that, we will use imputation methods recommended in the Cochrane Handbook (16.1.2) (Higgins 2011). We will be imputing the missing data with replacement values, and treating these as if they were observed. For dichotomous data we will use last observation carried forward and for continuous data we will account for missing data by imputing the mean.
We will request further details from authors in cases where published data are incomplete, perform sensitivity analyses to assess how sensitive results are to reasonable changes in assumptions made, and address the potential impact of missing data on the findings of the review in the Discussion section.
Assessment of heterogeneity
Where appropriate, we will assess heterogeneity of the data using the I2 statistic (Higgins 2003). A value greater than 50% is deemed to represent substantial heterogeneity. Where substantial heterogeneity is detected, and there are sufficient studies available, we will conduct subgroup analyses in an attempt to explain the heterogeneity. We will also conduct meta-regression where appropriate, with the help of the Cochrane Musculoskeletal Group statistician.
Assessment of reporting biases
We will assess publication bias using the funnel plot or other corrective analytical methods depending on the number of included trials of this review. We will explore reasons for any asymmetry in the funnel plot.
We will use a fixed-effect model unless statistically significant heterogeneity exists between studies. If heterogeneity is found a sensitivity analysis will be completed and followed by a random-effects model for meta-analysis if appropriate.
Subgroup analysis and investigation of heterogeneity
Conducting a large number of subgroup analyses increases the likelihood of false positive results and therefore it is important to carefully select the relevant characteristics to be investigated in advance (Higgins 2011). We plan to conduct the following subgroup analyses a priori in order to explore possible effect size differences:
Intervention - duration of treatment.
Characteristics of participants - severity of baseline disease; age; disease duration; sex.
For this review, we will assess results separately at three, six and 12 months, by duration of disease (as determined by average duration upon analysis), and by study eligibility criteria (probiotics).
Data will be analysed by different types of probiotic agents and how they effect on subgroup analyses.
Intervention – different types of probiotic agents in intervention group, and different control;
Different doses of probiotic agents;
Different preparations of probiotic agents;
Different duration of treatment;
Different follow-up period; and
Study design - inclusion of RCTs only and inclusion of CCTs as well.
Sensitivity analyses will be carried out to assess the robustness of results to inclusion of only studies at low risk of bias and to assess the robustness of results to variations in any estimated values of ICC.
'Summary of findings' table
We will include a 'Summary of findings' table to present the main findings of our review using GRADEprofiler (GRADEpro). It will provide key information concerning the quality of evidence, the magnitude or the effect of probiotics versus placebo or other mind and body therapies, and the sum of the available data on main outcomes (pain, fatigue, quality of life etc.).
We will include the following outcomes in the 'Summary of findings' table:
Tender point score
Withdrawals due to adverse events
Serious adverse events