Background
Hallux valgus (HV) is one of the most common chronic foot complaints presenting to foot and ankle specialists [
1], occurring when the hallux deviates laterally towards the other toes, and the first metatarsal head becomes prominent medially [
2]. As well as being a major contributor to the costs for forefoot surgery, HV has been linked to functional disability, including foot pain [
3], impaired gait patterns [
4], poor balance [
5], and falls in older adults [
6,
7].
Although HV has gained substantial attention in both historic and recent literature, several authors have highlighted the fact that a true prevalence estimate for HV is difficult to ascertain [
8,
9]. A wide range of prevalence estimates for HV has been presented in a multitude of independent reports. National health surveys in the United States have reported a prevalence of 0.9% across all age groups [
10], while a more recent survey in the UK reported a prevalence of 28.4% in adults [
9]. Research conducted in elderly populations has indicated prevalence rates as high as 74% [
11]. Individual studies have reported that HV is more common in female and elderly individuals [
9,
12]; however, there has been no synthesis of the literature to date or synopsis derived.
Due to the lack of firm epidemiological data relating to HV, it is difficult to estimate the impact that this condition has on the population; thus, in order to establish the need for future research, a better understanding of HV prevalence is warranted. To date there has been no published systematic review investigating the prevalence of HV and the influence of age and gender. Therefore, the aim of this systematic review and meta-analysis was to examine HV prevalence in the overall population and in age and gender subgroups.
Discussion
This review revealed a wide variation in HV prevalence estimates, and meta-analysis showed that systematic differences in these estimates were related to a number of factors, including method of HV diagnosis, gender, age, study quality, and sampling method. The finding that substantial differences may be related to the method of HV diagnosis (i.e. self-report or clinical examination) (Figure
3), confirms the results of a number of studies that have shown lower prevalence rates with the self-report methods commonly used in large-scale surveys when directly compared to clinical examination [
17‐
22]. Prevalence of HV may therefore be under-reported in epidemiological surveys that rely on self-report data.
Systematic differences according to gender and age were clearly demonstrated by our meta-analysis. The pooled estimate of HV prevalence in females (30%) was 2.3 times greater than the estimate for males (13%). This supports the observation of several individual reports that HV is more prevalent in females. For example, a recent large-scale epidemiological study of people older than 30 years reported a prevalence of 38% in women compared to 21% in men [
9], and another recent survey of older adults reported a prevalence of 58% in women and 25% in men [
12]. The trend for an increase in HV prevalence with age was also demonstrated by our data: 7.8% in juveniles (16 studies, n = 73,030), 23% in adults aged 18-65 years (15 studies, n = 23,790) and 35.7% in the elderly (37 studies, n = 16,001) (Table
1).
Variations in reported prevalence of HV in previous literature may also be explained by differences in study quality and methodological issues, particularly sampling bias (Figure
3). We identified a trend for higher prevalence estimates from studies with low quality scores on the EAI (score <0.91). Higher prevalence estimates were also reported by studies using convenience samples [
23‐
29] or biased samples of people seeking treatment for foot problems [
15,
30,
31], in comparison to those studies that used random sampling from the general population [
10,
16,
32‐
39]. Potential bias may be introduced by lower quality studies with sampling bias; however, as discussed previously, this trend may also be related to the fact that these "low" quality studies were mostly clinical studies that diagnosed HV rather than relying on self-report data.
Our findings should be considered in light of several limitations in the available literature concerning HV. One major concern is the lack of a clearly stated definition of HV in the majority of studies reviewed. Even in those studies where HV was observed on clinical examination, very few described a quantifiable method of measuring HV. Only 16% of studies in our review defined a diagnosis of HV using angular criteria measured clinically or on x-ray. A few more recent studies used the Manchester Scale, a categorical scale based on standardised photographs with four gradings to classify HV severity [
40‐
43]. Of those studies that collected self-reported prevalence data via interview or questionnaire, only a few provided participants with a definition or diagram of HV [
9,
35,
44]. In addition, there is confusion surrounding the interchangeable use of the terms "bunion" and "hallux valgus." In this review both terms were considered to represent HV; however, the term "bunion" strictly refers to the medial bursitis that may develop over the first metatarsal head as a result of irritation [
1]. Most included studies that used self-report data asked subjects about "bunions"; undoubtedly, a poor understanding of the terms used in a questionnaire or interview will result in inaccurate self-report data. Finally, there has been poor reporting of the reliability and validity of methods used to diagnose HV. Clearly, for accurate prevalence data to be collected and compared across different populations a consistent definition of HV and validated measurements should be employed.
Another consideration for our meta-analysis was the statistically significant degree of heterogeneity or variation across studies. Wide variations in sample populations meant that much of the retrieved data could not be pooled; however, pooling of estimates across age and gender subgroups was considered to be an important synopsis of the available literature pertaining to HV. Our subgroup meta-analysis was limited by the fact that not all studies reported HV prevalence by gender or age. Those studies that did report prevalence by age used a range of different age groupings, which rendered impossible further sub grouping the 18-65 years age bracket. Our analysis of potential sources of bias (Figure
3) was conducted to attempt to explain this variation between studies and highlight possible sources of heterogeneity.
Finally, insufficient data was available to examine the influence or adjust for other factors such as ethnicity, geographic location, shoe wearing or socioeconomic status on HV prevalence. Details of sampling frame and sample characteristics were also often poorly reported, as revealed by our quality assessment (Additional file
3 Additional File 3.xls). The vast majority of studies did not report on the presence of symptoms (i.e. pain or disability) related to HV, and therefore this factor could not be investigated by our review.
Having highlighted the limitations of the currently available epidemiological data relating to HV, further large-scale epidemiological studies are clearly warranted. Future studies should utilise rigorous methods, including random sampling from the general population and from different ethnic and socioeconomic groups. Validated tools should be used to diagnose HV, and results should be reported by gender and age as these factors are known to be associated with HV prevalence. Information relating to the presence of symptomatic versus asymptomatic HV would also be of great benefit in determining the impact of HV on the general population. Clear reporting of all these factors in future studies will provide an evidence base that will enhance our understanding of the impact of HV on the population and the health care system, and subsequently assist with the delivery of appropriate treatment. Due to its prevalence in the aging population, further research should focus on the impact of HV on mobility and quality of life in the elderly.
Acknowledgements
Funding/Support: SN is currently supported by an Australian Postgraduate Award Scholarship at The University of Queensland, and funding for the cost of language translation was provided by the School of Health and Rehabilitation Sciences, The University of Queensland.
Additional Contributions: We thank Dr Asad Khan, PhD (School of Health and Rehabilitation Sciences, The University of Queensland) for his statistical advice in this meta-analysis.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All authors contributed equally to the conception and design of this study. SN carried out literature searches, quality assessments, data extraction and statistical analysis and was responsible for drafting of the manuscript. MS also carried out quality assessments. MS and BV were responsible for supervision, including interpretation of data and critical revision of the manuscript. All authors read and approved the final manuscript.