Throughout the period of one year, approximately 50% of recreational runners will sustain an injury that disrupts their training regimen 12. Intrinsic risk factors shown to consistently correlate with running injury include previous injury 34567 and limited running experience 38910. Support for an association between foot morphology and specific running-related injuries has also been shown in several clinical studies. For example, the association between pes cavus and lower extremity stress fracture is well supported 71112131415, while further evidence demonstrates an association between pes planus and medial tibial stress syndrome 1113161718. An association between foot posture and plantar fasciitis in the running population is less convincing 7192021, however Irving et al. 22 recently found pronated foot alignment to be a risk factor for this condition. An association between foot pronation and patellofemoral pain has also been suggested in the literature 7112324, however this relationship has been contested by several prospective and cross-sectional cohort studies 252627282930.

Patellofemoral pain syndrome is reported to be the most commonly encountered running-related injury 13132, accounting for between 11% and 16% of conditions 3132. The incidence of stress fracture has been found to vary from 4% to 15% of running injuries, with the tibia, navicular, and femur being the most common sites 313233. Medial tibial stress syndrome and plantar fasciitis have been found to have similar rates of incidence, accounting for between 4% and 8% of running-related injuries 3132.

The clinical effectiveness of foot orthoses has been demonstrated in clinical trials for either the prevention 34353637 or treatment 383940414243 of the running-related conditions described above. However at the time of writing, no systematic review evaluating the mechanism of action of foot orthoses during running had been published. This limitation has consequences in relation to dispensing foot orthoses, as without an understanding of the intervention effect, the presumed action may not be produced as intended.

Several literature reviews have evaluated the effects of foot orthoses on lower extremity position and movement (kinematics) without a systematic search strategy 44454647. These reviews have described the effect of foot orthoses on kinematic variables to be small and non-systematic. As a result, research into the biomechanics of foot orthoses has increasingly focussed on lower extremity force and pressure (kinetics). However, the effect of foot orthoses on kinetic variables during running had not been systematically evaluated at the time of writing. The aim of this literature review was therefore to systematically collect all published research in this topic, and critically evaluate the methodology and experimental findings.

A systematic search of electronic databases including Medline (1966-present), CINAHL, SportDiscus, and The Cochrane Library occurred on 7 May 2008. The search terms foot orthotic$, foot orthos$s and insole$ were used in conjunction with the terms kinetic$, biomechanic$, running, and force$ in various combinations (Table 1). The search strategy was limited to articles published in the English language. Targeted searching of relevant journals also occurred following bibliographic review of retrieved articles.

Articles accepted for inclusion were required to be published in peer-reviewed journals, and report the findings of original experimental or quasi-experimental research. Articles were excluded according to the following criteria:

• Measurement of effects during walking

• Kinematic variables exclusively investigated.

• Orthoses other than foot orthoses exclusively investigated.

Electromyographic (EMG) studies were also excluded, as they measured biophysical variables that are distinct from kinetics. Furthermore, in accordance with definitions contained in the Australian Podiatry Council's 'Guidelines on Orthotic Therapy' 48, studies investigating the effects of insoles limited to cushioning properties were also excluded. Titles and abstracts of all citations generated by the search were assessed by one author according to the inclusion and exclusion criteria above, with articles printed in full-text as required.

All articles accepted for review underwent methodological assessment to evaluate the research quality. This process occurred in accordance with the Quality Index described by Downs & Black 49, in which a systematic checklist is used to evaluate the external and internal validities of clinical trials. This checklist was adjusted to exclude 12 questions deemed to be less relevant to the articles assessed in this review, resulting in the retention of 15 questions (Table 2). Additionally, the protocol and specific design features of each study were extracted, with particular focus on condition randomisation, participant running experience, condition acclimatisation and orthosis design.

Following methodological assessment, articles were grouped and discussed according to kinetic outcome variables. However the evaluation was unable to be conducted as a meta-analysis, due to heterogenicity in experimental designs.

The search process generated a total 1801 citations for initial screening (Table 1), of which 1770 were excluded on review of title and abstract. 31 articles were printed in full-text for further consideration, of which 10 were eligible for final inclusion (Table 3). All included articles were published between 1991 and 2008, and reported the findings of laboratory-based research with a repeated measures design.

The mean Quality Index Score for the articles was 64 % (SD = 10.5), demonstrating limited overall quality (Table 2). The majority of studies demonstrated inadequacy in selecting a representative sample and in the description of participant characteristics, while none attempted to blind subjects or investigators. Four studies included participants without reporting estimates of weekly running mileage. Additionally, several trials failed to randomise the sequence in which conditions occurred, thereby exposing the findings to order effects.

Despite the limitations described above, all studies provided adequate descriptions of outcome variables and orthosis design parameters, and reported findings with estimates of random variability. Furthermore, all used standardised footwear and running speeds during experimental conditions.

The studies included in this review are of low methodological quality, with the most confounding error being the lack of randomisation to the order of conditions. The most widely reported kinetic outcomes were loading rate and impact force, however the effect of foot orthoses on these variables remains unclear. In contrast, current evidence suggests that a reduction in the rearfoot inversion moment is the most consistent kinetic effect of foot orthoses during running.

This systematic review has evaluated the evidence surrounding the effects of foot orthoses on lower extremity kinetics during running. The findings demonstrate systematic effects that may inform the direction of future research in this field, as further evidence is required to define the mechanism of action of foot orthoses during running. Continuation of research in this field will enable targeting of design parameters towards biomechanical variables that are supported by evidence, and may lead to advancements in clinical efficacy.

The authors declare that they have no competing interests.

AM conceived the study design, conducted the systematic review, interpreted the findings and drafted the manuscript. CP reviewed the manuscript and provided academic support throughout.

AM is an Honours student within the Department of Podiatry, La Trobe University. CP is a Senior Lecturer within the Department of Podiatry, La Trobe University.