The aim of this study was to determine if there was adequate scientific evidence to support the concept of core stability. The initial problem encountered during this process was defining what core stability was. After considering a vast array of current definitions and analysing the anatomical, physiological and biomechanical evidence with respect to a number of well established theories and hypotheses there would appear to be sufficient foundation studies on which to build such a concept however this review of the evidence identified many areas where significant gaps in the knowledge base existed. For the core stability concept to progress there needs to be a universal definition for the term. This paper has argued that the anatomical region runs from the scapulothoracic joint superiorly to the pelvic girdle inferiorly.
Many anecdotal and theoretical models remain only partly vindicated by scientific research. Because the term core stability has become so popular as an expression and permeated our language both academically and socially it would appear to be sensible to perfect this concept before coining more jargon. It would be very easy to create more jargon however the core stability concept has appeared to grow organically and instead of reinventing the wheel, more research into developing this current concept would appear to be the most efficient use of future resources as directions for future research have now become more clear.
A finding from this study was the lack of efficiency and depth of searches available from the electronic databases used. Much of the material used in the evidence tables was sourced by manually searching contents pages, indexes and reference lists. An increasing number of journals are now available online in full text versions however it appears that many journals are still not listed with any databases. In a review of listings of sports medicine journals by Orchard and Blood (2002) they compared PubMed and SPORTDiscus listings for 23 journals. They found that all 23 were listed on SPORTDiscus compared with only 13 on PubMed. As some topics such as core stability transcend the boundaries of many disciplines it highlights the importance of performing multiple electronic searches and complementing technology with traditional manual data searches.
The SIGN 50 guidelines warrant further explanation of their criteria for grading guidelines. It is quite feasible that a key question with ample quality research may never receive an ‘A’ rated guideline. This tends to be related to questions pertaining to ‘raw’ science and not cause or effect from an intervention such as a treatment or a drug. In this present review Key Question 1 which reviewed the anatomical, physiological and biomechanical evidence behind the concept of core stability would be unlikely to ever receive an associated ‘A’ rated guideline. This may be viewed as a limitation of the SIGN 50 (2004) criteria however SIGN is trying to educate health professionals of the grading criteria to avoid the perception that a guideline without an ‘A’ rating may have a low level of importance. Harbour and Miller (2001) found that a common shortcoming of guideline users was to misinterpret the grade of recommendation as relating to its importance, rather than to the strength of supporting evidence. Due to the ratings schema where studies such as randomised controlled trials received the highest evidence rating much research into anatomy, physiology and biomechanics does not lend itself to these types of investigations so potentially their ‘gold standard’ level may only be a B or C. This view is supported by that of Harbour and Miller (2001) who found that grade A recommendations are relatively rare under the SIGN system and B will be the best achievable in many areas.
The guideline developed from key question 1 was, ‘Anatomical, biomechanical and physiological evidence exists that supports the concept of core stability’. This was arguably the most important key question as it examined the foundation of the core stability concept. If this question did not stand up to scrutiny then the concept would have no basis by which to exist. Whilst this review found good evidence to support the basic anatomical, biomechanical and physiological infrastructure there were significant areas that remain unclear. The functional role of the neural feedback subsystem is the area that lacks most in depth investigative analysis (Holm et al.,2002; McGill et al., 2003). This guideline was awarded a C grade primarily as there were a reasonable number of 2+ studies that related directly to the research question.
This review of the evidence proposed anatomical boundaries to define the parameters of core stability. As there were very few studies considering the direct anatomical, biomechanical and physiological principles of core stability it was necessary to extrapolate findings from related research. Determining the physical boundaries of the concept enabled appropriate evidence from secondary sources such as biomechanical models to be extrapolated as appropriate.
The guideline developed from key question 2 was, ‘Objective measures for measuring core stability are available and should be used to assist intra and inter tester reliability’. A ‘C’ grading was considered appropriate as the scientific evidence was reasonable and it did include studies with evidence levels of 2+. The current clinical emphasis appears to be the performance of graded functional tests with a degree of objectivity coming from devices such as pressure biofeedback units which provide a guide to general trends of overall performance rather than being diagnostic (Richardson et al., 1999; Sahrmann, 2002; Wisbey-Roth, 2000). This review did identify more accurate and specific tests of muscle function with proven reliability such as musculoskeletal diagnostic ultrasound (Hodges, 2005; Kermode, 2004) and EMG with indwelling fine wire electrodes (Richardson et al., 1999). More recent developments such as MRI scanning in functional positions (McGregor et al., 2001) and piezoelectric accelerometry (Brumagne et al., 1999) do provide measures of segmental movement and may become used more frequently especially in research settings.
On review of the evidence there is a reasonable foundation of tests for measuring components of core stability function such as muscle activation patterns (Ali, et al., 2001; Koumantakis et al., 2001), fatigability (Roy et a., 1989 and 1995) and alignment (Toppenberg and Bullock, 1986; Bullock-Saxton, 1993) however the clinician must understand that there is not one test or unit of measure that satisfies the criteria as a single objective measure of core stability function. On considering a hypothetical mathematical equation for core stability one would have to include such a vast array of variables with many different units of measure that it is doubtful the resultant measure of core stability would have any useful relevance. This review found one attempt at creating a core stability like equation in the literature. Richardson et al., (1990) used an equation to consider stability however when extrapolated into the core stability concept it would appear to be deficient of key variables that contribute to the core stability principle such as muscle function, pain and neuromuscular dynamics (Butler, 1991; O’Sullivan et al., 1997a and 1997b; Sahrmann, 2002).
Many studies used EMG data to monitor muscle activation patterns (Cholewicki et al., 1997; Brown et al., 2003; Stanton et al., 2004; Davidson and Hubley-Kozey, 2005). The majority used surface EMG data that was normalised however this cannot account for inaccurate electrode placement and cross-talk between electrodes that could interfere with accuracy of readings and therefore interpretation of results. The use of fine wire electrodes directly into deeper muscles improved the validity of EMG readings (Richardson et al., 1999) however the practical application is generally limited to experimental studies rather than clinical use. Of note Hodges and Gandevia (2000b) did report some limitations in measuring EMG activity using intramuscular techniques. Of further interest concerning the use of surface EMG were the number of studies that measured rectus abdominus activity as a measure of stability function (Cholewicki et al., 1997; Brown et al., 2003; Stanton et al., 2004). Stability theorists such as Bergmark (1989), Gibbons and Comerford (2001a and 2001b) and Richardson et al. (1999) classify rectus abdominus as a global mobiliser therefore being on the third tier of the stability scale following local and global stabilising muscles. In defence of some of these studies their research objective may not have been directly core stability related, for example evidence was extrapolated from studies on exercise and back pain (Burnett et al., 2004). It is noteworthy however that rectus abdominus muscle activity may provide poor evidence for core stability function. Conversely, lack of rectus abdominus activity during established exercises that focus on deeper stabilising muscles (e.g. abdominal hollowing as described by Richardson et al. (1999)) could potentially provide an indication of the functioning of the deeper muscles involved in spinal stabilisation. When measuring variables indirectly as in the above example using surface EMG there is considerable margin for error especially as it has been established that individual EMG characteristics have been shown to vary in normal populations performing identical movement tasks (Basmajian and De Luca, 1985).
While there are some well established objective measures for measuring components of core stability the challenge still remains on how best to apply the current mix of static and dynamic tests. Many of the tests simply measured general trends of core stability function rather than isolated specific lesions or erroneus variables.
The guideline developed from key question 3 was, ‘Core stability has a direct influence on human performance’. The anecdotal evidence for this guideline was vast however there have been very few scientific studies performed to directly measure this influence. The guideline was therefore awarded a ‘D’ rating. Potentially, there is extensive scope to improve this guideline rating. The converse may also be true and future studies may not be able to find a direct link between performance and core stability. This is an exciting area of potential research as many training programmes assume this concept to be steadfast and base much of their training protocols on this possible association.
Sporting competition and the desire for individuals to strive for athletic perfection would appear to be the catalyst for sports trainers to constantly seek new directions in training methods and techniques (Boyle, 2004; Cook, 2003). Some of their strategies have no scientific basis and become passing trends or fads that come and go before any research is performed on them. Others survive the practical testing ground of sporting fields and gymnasiums and continue to be developed and refined based on experience, intuition and results. It appears at this stage that scientific evidence often starts to catch up and evaluate techniques and concepts. Pilates is a good example of an exercise concept that evolved organically for 40 years before scientific research confirmed the benefits of much of the original programme (Stott, 2002). Science has also optimised the Pilates concept with the integration of the neutral pelvis concept into many of the exercises (Blount and McKenzie, 2000; Stott, 2002).
Not much of the core stability concept has its birth place from scientific foundation studies. It is important that the clinician is able to distinguish between concepts based on solid grounding and those from less structured backgrounds. Many hypotheses and practices are eventually vindicated by scientific research which is often lagging. As a minimum criteria it is important that clinicians and trainers at least have a structured justification for methods they practice until a related evidence based option is available.
The guideline developed from key question 4 was, ‘Core stability training should be part of any injury prevention programme’. This is another area where direct evidence was poor whereas anecdote and expert opinion was overwhelmingly positive. As the majority of the evidence was derived from expert opinion which has an Evidence Level rating of 4 the guideline was awarded a ‘D’ rating. The simple biomechanical and physiological principles that justify this guideline appear to be commonsense however the supporting evidence has been slow to follow.
The guideline developed from key question 5 was, ‘Core stability training can decrease the recurrence of lumbo-pelvic dysfunction’. This question provided studies with the highest ratings of evidence levels in this review. This included studies rated between 1+ and 3 and the extrapolated evidence from 3 systematic review papers. On considering the trend of evidence and the extrapolated opinions of some 1+ studies this guideline was awarded a B grade. There is still further scope for this guideline to receive a higher rating.
Low threshold stabilisation exercises (Davidson and Hubley-Kozey, 2005; Richardson et al., 1999) combined with an isolated approach for retraining specific local stabilising muscles has become the leading active treatment option for lumbopelvic dysfunction based on evidence. This approach has been shown to have good longterm followup results (Hides et al., 2001). However, it would appear that the quest to find the most optimum form of exercise type for the core stability muscles has not been exhausted. Whilst it would appear that the low threshold exercises complement the research findings from muscle fibre type studies (Jorgenson et al.,1993) there are many potential treatment paradigms that have not been studied. Even within the context of low threshold stabilisation exercises there are no gold standard variables for exercise type, frequency, duration or intensity. The general approach appears to mimic the guidelines for endurance strength training which tends to be 4 sets of 12-15 repetitions ranging from 2-5 times per week (Astrand and Rodahl, 1986). Other researchers have suggested that a rounded strength training programme should devote 15% to eccentric training, 10% to isometric training and 75% to concentric training (Astrand and Rodahl, 1986; Hakkinen and Komi, 1981). This sort of regimen is purported to produce maximal power (Hakkinen and Komi, 1981) which is not what is required of the local and global stabilising systems (Gibbons and Comerford, 2001a and 2001b) however may be of benefit to the global mobilisers depending on task requirements. This approach is possibly redundant when considering functional core stability training programmes that have an emphasis on regaining control over the body when placed in challenging situations of progressive instability (Boyle, 2004). There is a distinct lack of evidence for functional training programmes and this review found no long term outcome studies. This is in contrast to the popularity of this system especially when applied to sports rehabilitation (Cook, 2003; Haynes, 2004; Lawrence, 2003; Liebenson, 2003).
On reviewing the available evidence it would appear that a sound approach to lumbopelvic rehabilitation would be to commence with low threshold stabilising exercises (Richardson et al., 1999) and to progress to more functional training that involves more complicated movement patterns to enhance neural control systems (Boyle, 2004; Liebenson, 2003). Objective measures should be used to determine when to progress the degree of difficulty of exercises such as EMG (Elfving et al., 2003; Sung et al., 2001), musculoskeletal diagnostic ultrasound (Hodges, 2005; Kermode, 2004) or a graded schema of exercise tests (Sahrmann, 2002; Wisbey-Roth, 2000). This treatment schema has not been validated in this combined format however future investigations may provide further direction for treatment approaches.
The guideline developed from key question 6 was, ‘core stability can be improved with specific exercises’. Some components of this section are well investigated with evidence levels ranging between 1+ and 3. There was an abundant supply of papers rated as evidence level 4. Due to the inconsistencies this guideline was rated a ‘C’. There is scope for this guideline to be rated higher in future. With a combination of more robust study designs and a more consistent approach to what core stability is there is potential for this guideline to eventually receive an ‘A’ rating.
It appears that core stability is a trainable variable that can improve with various exercise programmes such as spinal stabilisation exercises (Davidson and Hubley-Kozey, 2005; Sung, 2003); Swiss Ball exercises (Marshall and Murphy, 2005) and Pilates (Herrington and Davies, 2005). Improvement in core stability function does not necessarily equate with improved performance in other physical tasks (Stanton et al., 2004; Tse et al., 2005). Future research may identify methods to determine the most appropriate CST protocol for specific functional requirements. Early indications suggest that low threshold spinal stabilising exercises are the most appropriate for managing non-specific chronic low back pain (Hides et al., 2001; McGill et al., 2003).
Limitations of Study
This review found that there were very few studies specifically considering core stability as the principal form of intervention when considering aspects of human performance, injury and training protocols. This resulted in studies that investigated components of the core stability model having their results extrapolated to be considered in a wider context. An example of this extrapolation of findings comes from some exercise and back pain studies that were applied to the lumbopelvic dysfunction and core stability research questions (Dolan et al., 2000; Dumas et al., 1995a and 1995b; Hides et al., 2001). Where there is extrapolation of results the value of the evidence decreases (SIGN 50, 2004). As more studies specifically use core stability as the principal intervention the quality of evidence will improve assuming that consistent definitions for core stability are used.
The level of bias could be affected in this study by the small number of research teams focusing their ideas and opinions on the core stability concept. Currently the evidence base is relatively small with an even smaller number of researchers that are regularly published. Familiar names frequently appear in the literature and while this is not necessarily a problem if the research methodology is tight it tends to mean that some areas of the concept become highly examined and other areas remain unexplored to the same depth. This could lead to specific components receiving greater emphasis than warranted simply because background material is more substantiated. An example of this could potentially be in the isolationist versus functional training debate. (Siff, 2000). Much work has been done on teaching subjects and patients to achieve an isolated contraction of transverse abdominus and multifidus (Richardson et al., 1999) as part of their spinal stabilisation exercises. The efficacy of graded functional training in comparison to this type of isolated muscle exercise is poorly researched (Boyle, 2004; O’Sullivan, 2000; Siff, 2000). As research develops in this area significant changes in the application of core stability principals may occur and potential bias could be reduced.
This evidence based review was performed by a physiotherapist and it is acknowledged that this may have influenced the interpretation of the material. Bias could be further limited in this study by having professionals from other fields involved to provide a more thorough analysis.
It is imperative that future research uses consistent terms. This review found that the terms core stability and core strength were frequently used interchangeably even within the same studies. Core stability has not been defined ‘universally’ which has made comparison between studies extremely difficult and direct statistical analysis between studies impossible. For core stability to become a more credible concept there is a need for the physical component of the definition to be clearly defined in the literature. This study proposed defined anatomical regions and structures for what constitutes the physical component of core stability based on review of the anatomical, biomechanical and physiological evidence. Many studies and expert opinions have been too narrow with their definition and use of the term and have extrapolated their findings and the impact of these findings on ‘greater’ core stability without considering the impact on many of the anatomical variables that appear to contribute to the core stability concept.
Within the field of movement dysfunction studies new terms are frequently proposed for describing states of dysfunction and methods to correct or treat them (Gibbons and Comerford, 2001a and 2001b; Siff, 2000). Many of the terms would appear to have the same or similar meaning. At times it appears that considerable effort has gone into creating new labels for old concepts. While this is done in the business world to attract new customers to old stock in the science community it would appear to be an inefficient use of resources and potentially dilutes the credibility of some pre-existing concepts. It would appear that there is scope for a universally recognised database of accepted definitions. This could eventually have a similar influence as accepted SI units have as standardised units of measure. A web-linked database could be created to access industry accepted definitions and provide a forum to the scientific community whereby new definitions could be proposed and critically appraised by scientists worldwide. This would assist the quality of research in all fields of science by ensuring consistent definitions are used thus enabling more direct comparisons to be made between different studies. This would also have positive financial implications as research into specific areas could be more efficient. Through the use of consistent terms comparative statistical analysis between different studies could be more powerful thus increasing the potential to arrive at more statistically robust findings and conclusions efficiently.
To maintain clinical relevance the core stability clinical guidelines proposed in this review of the evidence would need to be regularly reviewed and updated as new evidence is borne.