Scolaris Content Display Scolaris Content Display

Combination drug therapy for low back pain

Esta versión no es la más reciente

Contraer todo Desplegar todo

Abstract

This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:

The primary objective is to investigate the effects of combination drug therapy in reducing pain and disability in patients with low back pain and/or sciatica presenting to primary care, compared to mono drug therapy, no/minimal treatment or placebo. A secondary objective is to investigate combination drug tolerability, participants’ rating of improvement and treatment satisfaction.

Background

Description of the condition

Low back pain is a highly prevalent condition, causing considerable disability and burden globally (Hoy 2014). Up to 70% of people will experience low back pain during their lifetime (van Tulder 2002) and although many patients with low back pain improve substantially within the first six weeks, some will still have pain and disability after one year (Menezes Costa 2012). Patients with low back‐related leg pain, such as sciatica experience intense radiating leg pain that may be accompanied by neurological signs (Koes 2006). Many patients with sciatica still have symptoms after two years and a quarter of those who recover from that episode of sciatica will have a recurrence within two years (Tubach 2004). Low back pain and sciatica are both associated with high healthcare costs, work absenteeism and economic burden (Hoy 2014; Stafford 2007).

Description of the intervention

Clinical guidelines for patients with low back pain and sciatica provide recommendations on analgesics, which are generally based on single ingredient medicines with few recommendations on combination drug therapy (Chou 2007). Combination drug therapy is commonly used in primary care in patients with chronic low back pain (Gore 2012; Taylor‐Stokes 2011) and in those with low back pain with a possible neuropathic pain component (Hall 2013). The use of combination therapy in patients with chronic low back pain increases as pain intensity increases (Taylor‐Stokes 2011). Studies have found that the most frequent combinations are opioid analgesics plus non‐steroidal anti‐inflammatory drugs (NSAIDs) or muscle relaxants (Gore 2012), and opioid analgesics are mostly prescribed in combination with paracetamol rather than as monotherapy (Williams 2010).

How the intervention might work

Combining two or more drugs may give greater pain relief (or equal pain relief with lower doses of each drug in the combination) compared to each drug given alone. This potentially can improve drug safety and tolerability. Obtaining greater or equal pain relief can be achieved with combination drug therapy when drugs have different modes of action or favourable pharmacokinetic properties, whereby the drug combination targets multiple pain mechanisms and produces additive or synergistic treatment effects. For example, opioid analgesics combined with paracetamol is thought to have synergistic effects (Miranda 2002) and combining drugs that target nociceptive and neuropathic pain may be beneficial in conditions such as low back pain where mixed pain mechanisms exist (Attal 2011; Freynhagen 2006).

Why it is important to do this review

There is limited evidence for the use of combination drug therapy in the management of low back pain and sciatica. Two previous systematic reviews on combination therapy in low back pain patients found that some drug combinations, such as pregabalin with celecoxib or opioid analgesics, were effective in reducing pain in patients with chronic low back pain compared to monotherapy (Morlion 2011; Romano 2012). However, these reviews were restrictive in their search strategies by language and date, no protocols were published, and they focused only on low back pain of chronic duration. The first review (Morlion 2011) was industry funded and the authors searched only one database. Furthermore, combination therapy may include a broader range of drugs, not considered in these previous reviews. For example, some studies investigating combination therapy in people with low back pain have used supplements such as vitamin B complex (Vetter 1988) and theramine (Shell 2012) in combination with an NSAID.

Combination drug therapy is used in primary care. Information about these medicine combinations, such as the amount of pain reduction, disability outcomes, and medicine safety over time, is clinically important. The current evidence on combination drug therapy in low back pain and sciatica remains unclear.

Objectives

The primary objective is to investigate the effects of combination drug therapy in reducing pain and disability in patients with low back pain and/or sciatica presenting to primary care, compared to mono drug therapy, no/minimal treatment or placebo. A secondary objective is to investigate combination drug tolerability, participants’ rating of improvement and treatment satisfaction.

Methods

Criteria for considering studies for this review

Types of studies

We will include randomised controlled, quasi‐randomised controlled and cross‐over trials (pre‐cross‐over data only) where group allocation occurred at random. These study designs minimise bias when evaluating the efficacy of interventions.

Types of participants

The population of interest will include participants of any background and age with non‐specific low back pain with or without sciatica. Pain may be (sub)acute (< 12 weeks) or chronic (≥ 12 weeks) in duration (Koes 2006). Trials that include participants with a combination of (sub)acute and chronic symptoms will only be included if the data are reported separately for each duration, or can be obtained. People with low back pain due to pregnancy, post‐surgery or specific causes such as neoplasm, metastasis, infection, osteoporosis, rheumatoid arthritis and fracture will be excluded.

Types of interventions

We will include studies that administered two or more different drugs compared to a single drug that formed a part of the combination, a placebo or no/minimal treatment (e.g. advice). Drugs will be classified according to the Anatomical Therapeutic Chemical (ATC) classification system advocated by the World Health Organisation’s Collaborating Centre for Drug Statistics Methodology (WHO 2013) or, if no ATC code is available, drug(s) will be categorised by the pain mechanism the drug targets. Drugs may include prescription, over‐the‐counter, complementary or alternative medicines and supplements, at any dose and duration. The route of administration may be local, systemic (oral or parenteral) or transdermal, as long as the intervention was provided in primary care.

Types of outcome measures

Primary outcomes

  • Pain intensity measured by a self‐reported outcome measure (e.g. visual analogue scale or numerical rating scale).

  • Disability measured by a self‐reported outcome measure (e.g. Oswestry Disability Index, Roland‐Morris Disability Questionnaire).

Secondary outcomes

  • Adverse events. This will be separated into:

    • Serious adverse events, defined as events that were life threatening or resulted in death, hospitalisation, significant incapacity, congenital anomaly or birth defects.

    • Non‐serious side effects. This will include health outcomes, e.g. nausea.

  • Participants’ rating of improvement (e.g. Likert scale).

  • Participants’rating of treatment satisfaction (e.g. Likert scale).

Follow‐up time points of outcomes will be categorised as immediate (≤ 2 weeks), short (> 2 weeks but ≤ 3 months), intermediate (> 3 but < 12 months) or long (≥ 12 months) term. The short follow‐up period will be considered the primary outcome time point. If multiple time points fall within the same time period, we will use one time point closest to 2 weeks, 7 weeks, 6 months and 12 months for each follow up period.

Search methods for identification of studies

Electronic searches

We will use the latest search strategies developed by the Cochrane Back and Neck Review Group (Furlan 2009).

We will search the following databases from inception to current:

  • Cochrane Central Register of Controlled Trials (CENTRAL, Cochrane Library)

  • MEDLINE (OvidSP)

  • MEDLINE In‐Process & Other Non‐Indexed Citations (OvidSP)

  • EMBASE (OvidSP)

  • Cumulative Index to Nursing and Allied Health Literature (CINAHL, EBSCO)

  • PsycINFO (OvidSP)

  • Web of Science (Thomson Reuters)

  • ClinicalTrials.gov

  • World Health Organization International Clinical Trials Registry Platform (WHO ICTRP)

We will search PubMed using the strategy recommended by Duffy 2014 to identify studies not in MEDLINE and the Cochrane Back and Neck Group's Trials Register through the Cochrane Register of Studies (CRS) for studies not in CENTRAL.

See Appendix 1 for the draft MEDLINE search strategy. It will be translated as closely as possible across the other databases. There will be no language or publication restrictions.

We will conduct additional electronic searches to identify other potentially‐relevant studies by searching the International Pharmaceutical Abstracts database (OvidSP), clinical trial registries, including pharmaceutical industry trial registers, and grey literature databases (Open Sigle and Grey Literature Report).

Searching other resources

We will check reference lists of eligible studies and relevant reviews for additional citations, and communicate with content experts to identify any missing studies. If necessary, we will contact authors to retrieve study information to determine eligibility.

Data collection and analysis

Selection of studies

Two review authors from a panel of four (SM, CL, RK, RP) will independently screen identified titles and abstracts to determine eligibility. Potentially‐eligible studies will have the full text independently appraised to determine inclusion into the review. Duplicate studies will be removed upon agreement. Disagreements will be resolved by discussion first, then arbitration by an independent third review author (CM). When articles are written in languages that cannot be read by the authors, the authors will seek colleagues to assist with reading the article. Review authors will not assess for inclusion any studies to which they have contributed. The flow of studies will be summarised in a study flow diagram, following the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) statement (Liberati 2009).

Data extraction and management

Two review authors from a panel of five (SM, CL, RP, BK, RK) will independently extract data. Standardised and piloted data extraction forms will be used. The following information will be extracted:

  • Bibliometric data (e.g. authors, year of publication, language, funding sources, if prospectively registered)

  • Study characteristics (e.g. study design, method of randomisation, sample size)

  • Participants (e.g. age, gender, duration of low back pain, type of pain)

  • Interventions and controls (e.g. drug class, dose, mode of delivery, treatment period)

  • Outcomes and results (e.g. pain score, adverse events, number of drop outs)

We will contact study authors to clarify any uncertainties or obtain additional information. Disagreements will be resolved by discussion first, then arbitration by a third author (CM).

Assessment of risk of bias in included studies

The quality of included studies will be assessed using the 'Risk of bias' assessment tool as recommended by The Cochrane Collaboration (Higgins 2011) and the Cochrane Back and Neck Review Group (Furlan 2009) (Appendix 2). Two review authors from a panel of five (SM, CL, RP, BK, RK) will independently assess risk of bias. We will consider the domains of selection bias, performance bias, detection bias, attrition bias, reporting bias and any other biases identified. Each domain will be scored as either 'low', 'unclear' or 'high' risk of bias and tabulated. If additional information is required, we will contact the study authors. If the authors cannot be contacted, or if the information is no longer available, the criterion will be scored as 'unclear'. A study with low risk of bias will be defined as having low risk of bias in six or more domains without any serious flaws (Furlan 2009). Disagreements will be resolved by discussion first, then arbitration by a third review author (CM).

Measures of treatment effect

Analyses of treatment effects on the primary and secondary outcomes will be conducted separately for participants with (sub)acute (< 12 weeks) and chronic low back pain (≥ 12 weeks) at immediate (≤ 2 weeks), short (> 2 weeks but ≤ 3 months), intermediate (> 3 but < 12 months) and long (≥ 12 months) term follow‐up periods per drug combination and its comparator (e.g. NSAID versus combination drug therapy).

For dichotomous variables such as side effects we will report risk ratio (RR) or risk difference (RD) as the effect measure with 95% confidence intervals (CI). For continuous outcomes, such as pain intensity on a visual analogue scale or disability on the Roland‐Morris Disability Questionnaire, results will be reported as mean differences (MD) with 95% CI. Where possible, outcomes will be converted to a 0 to 100‐point scale to facilitate comparison and interpretability. Analyses will be conducted using Review Manager 5.3 (Review Manager 2014).

Unit of analysis issues

The unit of analysis will be each participant recruited into the included studies. Due to the review's design, each participant will have only been allocated to one intervention. Therefore we anticipate that unit of analysis issues could potentially arise from repeated observations. If this occurs (e.g. in adverse events) we will use only one outcome closest to our defined follow‐up time points (i.e. immediate (≤ 2 weeks), short (> 2 weeks but ≤ 3 months), intermediate (> 3 but < 12 months) and long (≥ 12 months) for (sub)acute (< 12 weeks) and chronic low back pain (≥ 12 weeks) participants).

Dealing with missing data

If relevant data are missing, we will first contact the authors for clarification and additional data. Failing that, for outcomes we will estimate the means and standard deviations (SDs) from graphs, if available. If SDs are not reported, we will attempt to estimate them from the CIs or other measures of variance. If SDs are missing for follow‐up outcomes, we will use the SD for that outcome at baseline. Finally, if no measure of variation is reported or available from authors, we will estimate the SD based upon other studies with a similar population and risk of bias.

Assessment of heterogeneity

Heterogeneity will be assessed by visual inspection of the forest plot (e.g. P values and overlapping CIs) and by the Chi2 and I2 tests, following the recommendations in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We will follow the recommended guide for interpretation of I2 as: 0% to 40%, might not be important; 30% to 60%, may represent moderate heterogeneity; 50% to 90%, may represent substantial heterogeneity; 75% to 100%, considerable heterogeneity (Higgins 2011). If sufficient data are available and studies are clinically and statistically homogeneous (I2 < 50%) the results will be combined in a meta‐analysis using a random‐effects model. If heterogeneity is present, we will not pool data but will instead present a narrative synthesis.

Assessment of reporting biases

We aim to perform a comprehensive literature search to reduce the possibility of reporting biases. Reporting bias is considered in Assessment of risk of bias in included studies. If enough studies are retrieved (> 10 studies) we will further examine reporting bias using funnel plots.

Data synthesis

The overall quality of evidence will be assessed for each combination therapy group and outcome. A GRADE approach (Guyatt 2008) will be used to report the overall quality of evidence as recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) and Cochrane Back and Neck Review Group method guidelines (Furlan 2009). The GRADE approach considers the domains of risk of bias, imprecision, indirectness, inconsistency and publication bias (Appendix 3). The quality of evidence may decrease from ‘high’ in each domain when the criteria are not satisfactorily met.

The five levels of evidence are:

  • High quality evidence: there are consistent findings among at least 75% of RCTs with low risk of bias, consistent, direct and precise data and no known or suspected publication biases. Further research is unlikely to change either the estimate or our confidence in the results.

  • Moderate quality evidence: one of the domains is not met. Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.

  • Low quality evidence: two of the domains are not met. Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.

  • Very low quality evidence: three of the domains are not met. We are very uncertain about the results.

  • No evidence: no RCTs were identified that addressed this outcome.

A 'Summary of findings' (SoF) table will be constructed using RevMan. The main comparison will be analgesic versus analgesic combination drug therapy for the primary outcome of pain, at the short‐term follow‐up (> 2 weeks but ≤ 3 months) for (sub)acute (< 12 weeks) and chronic low back pain (≥ 12 weeks). The analgesic medicines and their comparisons listed in the SoF table will be dependent on the review's findings.

Subgroup analysis and investigation of heterogeneity

If there are adequate studies, we will conduct subgroup analysis of drug combinations for participants presenting with low back pain with radiating leg pain versus participants with low back pain only.

Sensitivity analysis

If there are adequate studies, we will perform sensitivity analyses to explain any possible sources of heterogeneity between studies and to evaluate the robustness of our analysis. In addition, we will compare treatment effects in both primary outcomes between studies with high and low risk of bias, based on the criteria recommended by the Cochrane Back Review Group (van Tulder 2003) where studies of low risk bias have scored low risk of bias in six or more domains without any serious flaws.