Discussion
Both insoles were comparable in reducing peak pressure (for all foot types) and rate of forefoot loading whilst increasing total contact area, but the custom-made functional insoles were more expensive than the prefabricated insoles. Peak pressure is the measure that has been traditionally used in studies evaluating the effect of offloading. However, in addition, we used four other measures. For one of these, forefoot pressure time integral, the custom-made insoles performed significantly better, and these differences were maintained over the six month study period.
Total peak pressure was selected as the traditional measure of the effect of footwear and insole efficacy to ease comparison of results between studies. Forefoot pressure time integral was selected to reflect the increased risk of ulceration over the forefoot. More recently it has been reported that peak pressure and pressure time integral are inter-dependent and that within clinical trials significant differences in patterns found between the two parameters are generally minimal [
20,
21]. Thus, there is little value in routinely reporting both measures [
20,
21]. We acknowledge, therefore, that forefoot peak pressure would have been an equally suitable measure and accept that our findings are likely to have been the same if forefoot peak pressure had been selected [
20,
21].
No other randomised controlled trial has been identified to indicate that custom-made functional insoles are significantly more effective in reducing forefoot pressure time integral or forefoot peak pressure than prefabricated insoles when used to reduce ulcer risk in neuropathic diabetic feet. Reductions in peak pressure observed at insole issue in our study (31-37%) are, however, comparable to peak pressure reductions demonstrated by other non-randomised studies (32%) [
22,
23]. At six month follow-up, peak pressure reductions observed in our trial (Prefabricated 31%, Custom-made functional 37%) appear better than findings presented elsewhere. Mohamed and colleagues [
22] compared a custom plastazote orthosis to a custom aliplast/plastazote orthosis in two groups of eight people with diabetic neuropathy and reported peak pressure reduction of 26% at three months. Lobman and colleagues [
23] compared 18 participants with neuropathy provided with custom made EVA insoles in therapeutic shoes with 63 controls and reported a mean reduction in peak pressure of 28% at six months. The apparent enhanced performance across time of the insoles evaluated by our trial may reflect the greater durability of materials selected. Mohamed and colleagues [
22] describe making modifications to the plastazote insoles after only one month of use to compensate for material compression.
Reduction in peak plantar pressure is believed to result from a corresponding increase in total surface area. We found that although both insoles increased total contact area significantly, that change had reduced by 50% at six months follow-up. If peak pressure was strongly associated with total contact area, a reduction in total contact area would generate an associated increase in peak pressure. Conversely, the results of our study found that at six months follow-up, whilst the effect of the insoles on total contact area had decreased, the effect on peak pressure stayed constant.
The absence of association between peak pressure and total contact area in this trial may be explained by considering the dynamic nature of the data. Most peak pressure sites were located over the forefoot during propulsion. In contrast, the insole generated the greatest increase in surface area beneath the medial longitudinal arch, at midstance. Thus, the effect of increased total contact area on peak pressure at midstance may not carry over into propulsion. The mechanism by which insoles reduce plantar peak pressure is unclear and in need of further investigation.
Two non-randomised studies confirm an increase in total contact area when insoles were worn [
24,
25]. Albert and Rinoie [
26] recorded a 5-10% increase in total contact area when a rigid custom-made device was worn; this increase was constant over the 3-month study period. Raspovic and colleagues [
24] used the F-scan® to evaluate change in contact area, testing a range of customised accommodative insoles in a sample of 8 high risk people with diabetic neuropathy and a history of ulceration. A 19% increase in total contact area was found when wearing the insole, no follow-up evaluation was undertaken. In comparison, our study reported a substantially greater total contact area increase of 32% and 29% for the custom-made functional and prefabricated insoles at issue, reducing to 15% at six months.
Differences between study results may be attributed to differences in methodology. Raspovic and colleagues [
24] investigated the effect of previously worn non-casted custom-made insoles on total contact area. Compared to our study insoles, the profile of the non-casted device may be less contoured, thus reducing the comparative total surface area. Equally the 19% increase in total contact area reported by Raspovic and colleagues [
24] may reflect the time from issue; their findings are comparable with the 15% increase reported by our study at six month follow up.
Albert and Rinoie [
25] included only people with diabetes and pronated feet. The contact area beneath pronated feet is high, thus reducing the capacity for an insole to further increase total contact area. A greater effect therefore might be predicted in a sample population not limited by foot type, such as that recruited for our trial (44 out of 119 had pronated feet in our sample). Furthermore, unlike our semi-rigid insoles, the insole investigated by Albert and Rinoie [
25] was of rigid construction, unlikely to flex under load. Differences in insole flexibility offer further explanation for differences in change in total contact area reported between studies.
The custom-made functional device fabricated for clinical use is commonly prescribed to include the biomechanical features best suited to foot type (pronated or supinated). Likewise, the custom-made functional device designed for this trial incorporated biomechanical features specific to foot type in conjunction with features considered safe for the insensate foot. More than one third of study participants randomised to the custom-made functional insole group (n=25/60) received an insole intended to benefit the pronated foot. Although direct comparison between studies is not possible, the 37% reduction in peak pressure achieved by the custom-made insole in this trial appears similar to the clinical benefit achieved by the rigid device evaluated by Albert and Rinoie.
The outcome measure, duration of load as a percentage of stance at the site of peak pressure was included in recognition of the contribution toward ulceration that increased load times are thought to play [
26,
27]. No other study considering this measurement was found in the literature. The effect of wearing insoles on duration of load as a percentage of stance was variable and inconsistent between participants and when the same participant was measured on different occasions. Duration of load as a percentage of stance over the site of peak pressure cannot easily be altered by insole therapy and does not appear a suitable measurement to assess the effect of insoles in people with diabetes and neuropathy. However, given the limited research using this outcome measure, further work is required to determine whether these findings are representative of a general trend before a final conclusion can be drawn.
This study found that wearing insoles reduced rate of forefoot loading in the diabetic neuropathic foot, although the effect was small. It has been suggested that rate of forefoot loading maybe as important as the magnitude of load in ulcer formation [
28], although an association with ulceration has not been determined. The only other study that we have found that assessed change in rate of loading in the evaluation of insoles for the management of the neuropathic foot [
24], was non-randomised and recruited only eight participants so was likely to be under-powered.
Neither the ADDQoL nor the Bristol Foot Score showed any difference between the two insoles. Compared to other populations with diabetes, this sample of 119 participants with diabetic peripheral neuropathy had worse health-related quality of life. A survey of 795 outpatients attending annual review at a UK hospital diabetes clinic reported mean weighted impact score for health-related quality of life of −1.98, which compares to a mean weighted score for participants in this trial of −2.38 [
29]. No other study has employed the ADDQoL to capture the impact of diabetic peripheral neuropathy on health-related quality of life.
The estimated direct cost of treating a diabetic ulcer over a 12 year period is £27,000 [
30]. Assuming insoles are replaced annually and footwear two yearly, the estimated cost of insole provision over a 12 year period was calculated at £4771 and £3500 for the custom-made functional and prefabricated insole respectively. Therefore, each diabetic foot ulcer prevented by insole provision offers a potential cost saving of approximately £23,000. Full utility analysis using longitudinal measures of time to foot ulceration or life expectancy is needed to provide robust economic analysis of insoles for diabetic neuropathic foot ulceration.
There is no clear evidence that the custom-made functional or prefabricated insole is best practice for all diabetic neuropathic feet, however both are of value for reducing ulcer risk in people with diabetes and neuropathy. Both custom-made functional and prefabricated insoles were equally effective in reducing peak pressure; therefore the less expensive prefabricated insole is likely to be most cost effective. Practitioners tasked with accessing the diabetic foot for insole provision should, where appropriate, consider prescribing the more cost effective prefabricated insole. The custom-made functional insole was found slightly more effective than its cheaper counterpart in reducing forefoot pressure time integral. The clinical significance of reducing forefoot pressure remains undetermined. Further research is needed to determine; (i) which parameter is more important in predicting neuropathic foot ulceration, (ii) the magnitude of reduction deemed clinically sufficient to produce a symptom change, and (iii) confirmation of costs over time.
Of the 104 participants completing our study, 42 (40%) reported full compliance equating to wearing the insoles and shoes for a minimum of 7 hours a day, 7 days per week for a period of six months [
31]. Chantelau and Haage [
31] reported that participants with diabetic neuropathy and a history of foot ulceration, wearing protective shoes for >60% of the daytime reduced ulcer relapse rate by 50%. Thus, for the purposes of this study, 60% of daytime wear of insoles and footwear for the duration of the six month study were considered compliant. Footwear compliance within this trial, although apparently low, is favourable compared to compliance rates reported elsewhere [
31,
32]. Three participants did not wear the insoles because they felt unstable when walking with them. However, the most common reason given by participants for not wearing the insoles and shoes for more hours per day, was that they were being removed whilst indoors, despite footwear education to the contrary. Improving insole and footwear compliance in patients presenting with diabetic neuropathy is crucial to the success of diabetic foot health maintenance. Insole compliance may be improved if patients were provided with both indoor and outdoor footwear to accommodate insoles.
This study needs to be considered in light of a few limitations. Firstly, although all the tests were pre-specified, by exploring four outcomes, in addition to peak pressure, we increased the possibility of finding a significant result by chance. However, even adjusting the p-value for multiple testing, the difference in pressure time-integral is significant. Secondly, the clinical environment we used to collect data may not fully replicate the day-to-day conditions within which the insoles are required to function. The choice of footwear used to accommodate insoles can affect function. Therefore, care must be taken not to generalise the findings of this trial beyond the type of therapeutic footwear provided within the study. The six month follow up in our trial gave limited information regarding insole durability. In practice, frequency of insole replacement often extends beyond six months and is usually determined by physical signs of insole wear or changes in foot health. Finally, application to clinical practice may be enhanced if future studies were specific not only to foot pathology but foot type (pronated or supinated). The implication that forefoot pressure time integral and other selected variables are symbolic of ulceration risk is too simplistic and should be approached with caution; the aetiology of diabetic neuropathic ulceration is multi-factorial and complex, therefore whilst clearly relevant to ulceration risk, kinetic parameters are merely surrogate measures of internal tissue stress. Moreover, the actual pressure threshold generating internal tissue stress above which ulceration is inevitable is undetermined, therefore estimates of clinical significance are difficult to predict. Further studies are required using a randomised controlled trial design to assess insoles used for the prevention of diabetic neuropathic foot ulceration, particular attention should be given to the comparison of insole type using incidence of ulceration as a primary measure of outcome.
Competing interest
The authors declare that they have no competing interest.
Authors’ contributions
JP conceived and conducted the trial, extracted and analysed the data and produced the initial draft manuscript. RJ, ES and GB critically reviewed the design, development and progress of the trial. RJ and ES reviewed the manuscript for academic content and GB reviewed the discipline specific content. DZ helped with the re-analysis of data dealing with reviewers’ concerns on initial submission. All authors read and approved the final manuscript.