The electronic searching identified 3332 eligible articles. From these, 144 full articles were retrieved for full-text screening. After review, eight studies were included and 136 were excluded for reasons reported in Fig. 1.

Fig. 1

PRISMA flow diagram of literature search and phases of study selection; n, number; CENTRAL, Cochrane Central Register of Controlled Trials

Characteristics of included studies

Of the eight included studies (n = 441 participants), four were randomised controlled trials (RCTs) [23,24,25,26] and four were prospective cohorts [27,28,29,30], reported across multiple publications between 2001 to 2021. Total study sample sizes ranged from 20 to 79 participants; for the RCTs, intervention arms ranged from 18 to 27 participants. Six studies recruited from clinical centres in the USA [23, 24, 27, 28] and of these, two also recruited participants from the UK [29] and Qatar [25]. One European trial recruited from diabetes-related foot clinics in public hospitals in the Netherlands and Germany [26], and one multicentre cohort study recruited from three diabetes-related foot ulcer clinics in Jordan [30].

Participant characteristics

An overview of RCTs and observational studies is presented in Tables 1 and 2 respectively. Most study recruits were male (333/441; 76%). Mean participant age, where reported, ranged from 53 to 64 years. Two clinical trials failed to describe age [23, 24] and six studies did not report ethnicity/race [23, 25, 27,28,29,30]. Of two studies reporting ethnicity, most recruited participants were White Caucasian (84/133, 63%), Hispanic (43/133, 32%), African American (4/133, 3%) or not described (2/133, 2%) [24, 26]. Six studies reported ulcer severity, recruiting participants with a University of Texas Ulcer classification of 1A [23, 27], 1A or 2A [24, 26], 1A up to 2D [29] or Wagner grade-1 classification [28]. One trial did not report foot ulcer severity [25] and the Jordanian cohort study only described ulcer severity in a subset of participants [30].

Table 1 Characteristics of Included Cohort StudiesTable 2 Characteristics of included randomised controlled studiesOffloading device treatment

Total contact casts [23,24,25,26, 28], and removable cast walkers [23, 25, 27, 30, 31] were the most popular devices for ulcer management, used in five studies each. Two studies investigated healing sandals [24, 29] and the remaining devices, used in one study each, included a half-shoe [23], shear-reducing walker [24], cast shoe [26], prefabricated forefoot-offloading shoe [26] with some ‘other’ devices not described [29]. One trial compared two different off-loading devices [25]; the remaining three trials each compared three different types of offloading devices.

Rehabilitation interventions

None of the eight studies delivered any form of structured or unstructured rehabilitation nor prescribed a physical activity intervention at any time during the period that offloading devices were worn. Furthermore, none delivered rehabilitation after offloading device removal to promote activity or support participants with returning to full mobilisation. All included studies used one or more measures of physical activity, captured either as a primary or secondary outcomes during ulcer treatment.

Overview of reported outcomes

All studies reported one or more physical activity outcomes, measured objectively using waist or ankle-worn pedometers or wearable sensors attached to clothes or embedded within offloading devices (Tables 1 and 2). Activity outcomes were reported as steps over time (hours, days, or weeks), daily stride count (distance of both right and left step), gait speed, percentage of time spent in different postures (sitting, lying, standing, walking) or in postural transition e.g. sit-to-stand and stand-to-sit. Activity monitors were used for different purposes, to capture periods of (in) activity, to determine adherence to wearing prescribed offloading devices [27] or to examine ulcer healing by activity status [28, 29]. Two studies reported activity outcomes only [27, 30]; the remaining six studies also reported ulcer healing or healing-related outcomes, either as proportion healed, change in size or area, and/or time to healing. None measured ulcer recurrence. Adherence to the offloading device was reported by five studies [24, 26, 27, 29, 30]. Other clinical and patient-reported outcomes included: wound/ulcer infection [24], participant satisfaction (foot care; activity levels; daily activities; offloading device comfort and satisfaction; sleep) [24], neuropathy and foot ulcer-related quality of life [29], complication rate [26], shoe peak pressure [26] anxiety/depression [29], and body posture [25]. Outcomes are reported separately by study design.

RCTs Physical activity (four trials; n = 242 participants)

All trials reported that participants wearing total contact casts were less active compared to those wearing other removable devices [23,24,25,26]. Physical activity after 4 to 12 weeks of ulcer treatment was lower amongst those wearing total contact casts compared to selected other offloading devices (half-shoe, forefoot offloading, healing sandals or removable cast), (SMD -0.45; 95% CI − 0.87 to − 0.04; p = 0.03; I2 = 56%; four trials, n = 242) (Fig. 2). Subgroup analysis of difference in physical activity after wearing a total contact cast versus a removable cast walker (two trials; n = 88) demonstrated higher levels of physical activity in those wearing a removable cast walker after 12 weeks (SMD -0.69; 95% CI − 1.32 to − 0.05; p = 0.03; I2 = 53%, n = 88) (Supplementary file 3).

Fig. 2

Forest plot comparing physical activity by type of offloading device at 12 weeks. SD, Standard Deviation; Std, Standard; IV, Weight Mean Difference; CI, Confidence Interval; TCC: Total contact cast

Najafi et al., [25] reported that walking was very low as a proportion of total daily activity (< 5% of time) – most participants spent their time lying or sitting, regardless of whether they were wearing a total contact cast (91% time sedentary) or a removable cast walker (83% time sedentary). Armstrong et al. [23] found no difference in activity levels between those wearing total contact vs removable casts over 12 weeks but failed to report activity data for all three intervention arms. One trial [26] reported no difference in daily stride count between custom-made offloading devices and a prefabricated forefront offloading shoe, although incomplete data were reported by treatment arm and attrition was high, with only 34/60 (57%) of participants providing activity data at 12 weeks, although authors reported that activity monitors were only given to half the recruited sample.

Ulcer healing (four trials, n = 242 participants)

Overall, odds of ulcer healing were higher in participants wearing non-removable total contact casts compared to all other removable devices combined after 12 weeks of treatment (OR 2.69; 95% CI 0.97 to 7.45; p = 0.06; I2 = 64%; four trials, n = 242) (Fig. 3). Subgroup analysis examining differences in ulcer healing between total contact casts versus removable cast walkers also demonstrated greater odds of healing in those wearing a removable cast walker after 20 weeks (OR 3.93; 95% CI − 1.48 to 10.47; p = 0.006; I2 = 0%; 2 trials; n = 88) (Supplementary file 4). Trial findings of ulcer healing outcomes at 20 weeks by different off-loading devices are reported in Supplementary materials (file 4). One trial [26] reported intention-to-treat and per-protocol data; a in a separate post-hoc sensitivity analyses including only per-protocol findings, this changed the strength but not direction of the estimate of effect at 12 weeks (OR 3.91; 95% CI 2.10 to 7.26; p = < 0.001; I2 = 0%; four trials, n = 235) (Supplementary file 5) and 20 weeks (OR 3.99; 95% CI 2.11 to 7.57; p = < 0.001; I2 = 0%; four trials, n = 231 (Supplementary file 6). Although all trials described ulcer size/area at baseline, only two reported area data at follow-up; one trial reported cumulative wound survival graphically (Table 2).

Fig. 3

Forest plot of ulcer healing after wearing an offloading device at 12 weeks. TCC, Total Contact Cast; CI, Confidence Interval; M-H, Mantel Haenszel

Observational studies Physical activity (four studies; n = 196 participants)

Four cohort studies measured physical activity for one to 13 weeks after offloading device application. Although one study was described as cross-sectional, recruited participants were monitored for 1 week, it was included as a cohort design [30]. Overall, findings from the observational studies suggested low levels of activity whilst people wore devices. In a cohort with 13 week follow-up [28], 40 participants were advised to be completely non-weight-bearing for the first 48 hours after total contact cast application, although most walked on the cast during this early period (median 808 steps over 48 hours). Total step count was low over the remaining observation period (median total 2083 steps/day). A USA/UK cohort study (n = 79) reported higher daily activity levels in those fitted with a removable cast walker, being active for over 6 h per day (mean 6.7 hours; SD 3.8) over 6 weeks, with adherence data suggesting the device was worn over half of time whilst active (59% (SD 22%)) [29]. In contrast, a cohort study with 20 participants found that even when monitored for only 1 week, adherence to a removable cast walker was poor [27]. Participants were less active whilst wearing their cast (mean total daily steps 345 (SD 219) vs 874 (SD 828) when not wearing cast; p = 0.01) with only 28% of total daily activity occurring while the cast boot was worn. Similarly, the Jordanian cohort study of 57 participants wearing removable cast walkers also reported poor adherence of 34% of activity time, when monitored for only 1 week [30].

Ulcer healing (two studies; n = 119 participants)

In the multicentre USA/UK cohort, only 19/79 (24%) ulcers healed over 6 weeks, although mean ulcer size reduced from 230mm2 to 106mm2 (p = 0.001) [29]. Ulcer healing was associated with better adherence to the offloading device although sample sizes were too small for meaningful analyses [29]. Ulcer healing was higher in the cohort study with longer follow-up, with 32/40 (80%) healed after 13 weeks of total contact cast treatment [28].

Ongoing studies

Registered clinical trials and unpublished studies on the EU Trial and ClinicalTrials.gov registers were searched. Six studies of interest were identified on ClinicalTrials.gov (NCT04280016, NCT04310137, NCT04085926, NCT05236660, NCT04460573, NCT04257565). All are currently open to recruitment with one or more pre-specified activity-related outcomes (Supplementary file 7).

Risk of methodological bias

Overall methodological quality of the RCTs was poor, with three of four trials being graded at high risk of bias overall and the more recent European trial judged as having some concerns [26] (Supplementary file 8). Only two trials had low risk of bias associated with the randomisation process [25, 26]. Two trials described that outcome assessors or data collectors were blinded to treatment allocation [25, 26] and one trial mentioned blinding but did not describe details [24]. All four trials were at risk of bias in selection of the reported results, due to missing outcome data. Similarly, methodological quality of the observational studies was poor - one was at moderate risk of overall bias [28] and three studies were judged as serious r