2.1. Data Sources and Search Strategy

PubMed and Web of Science databases (including Current Contents Connect, Derwent Innovations Index, Medline, and SciELO Citation Index) were used to identify potential studies. The MeSH descriptors were used: “Hyperbaric Oxygenation”; “Amputation”; “Dibetic Foot”; “Rehabilitation.”

Thus, the following search string was employed: (“Hyperbaric Oxygenation” AND “Amputation” AND “Diabetic foot”), (“Hyperbaric Oxygenation” AND “Amputation” AND “Rehabilitation”).

Studies were included in the systematic review if they fulfilled the following criteria: (1) clinical trials in English or Spanish; (2) published between 2008 and 2023; (3) studies on the use of the HC in chronic diabetic foot, with the aim of treating the ulcer and reducing the amputation rate, as well as the effects of the HC in the treatment of chronic diabetic foot, the objective of treating the ulcer and reducing the amputation rate, as well as the effects on prosthetic fitting; (4) studies with a control group; and (5) studies in which the patients received previous treatment without favorable results. Moreover, the studies were excluded when (1) they were written in a different language from English or Spanish; (2) they were a review, study protocol, conference abstract, or a case report; (3) they did not involve HC; (4) they do not include the influence on amputation rate; (5) they deal with lower limb injuries other than ulcers; and (5) they were animal studies.

The search process ended in November 2023. Duplicated studies were excluded, and articles’ titles, abstracts, and full texts were carefully screened.

The study selection was performed by one author, M.D.A.-A., and checked by another, A.C.-P. In case of disagreement among the reviewers, it was resolved by consensus.

2.2. Risk of Bias Assessment

The Evidence Project tool [18] was employed to evaluate the risk of bias of the selected studies. This tool is composed of eight items that cover study design, the participants’ representativeness, and the equivalence of comparison groups. In this regard, the study design includes items referred to cohort, control, or comparison group and pre–post-intervention data. Participants’ representativeness includes items that analyze the random assignment of participants to the intervention, random selection of participants for assessment, and follow-up rate of 80% or more. Lastly, the comparison groups’ equivalence is assessed with items concerning the equivalent on sociodemographics and the equivalent at baseline. This scale allows to evaluate both randomized and nonrandomized trials.

2.3. Data Extraction

According to PRISMA methodology [14], participants, intervention, comparison treatments, outcomes, and study design (PICOS) data were extracted. Accordingly, information concerning participants’ characteristics, study design, sample size, age, severity of DFU according to the Wagner scale, and age of onset were exported from each article. Moreover, intervention characteristics such as intervention length, treatment frequency, duration of the sessions, and its description were analyzed. The extraction process was conducted by three authors (M.D.A.-A., L.G.-N., and A.C.-P).

Figure 1 shows the flowchart followed for the identification of relevant articles for the work, specifying those that were excluded for not being related to the objective of the work, for being duplicated, or for not having access to the full text of the article.

3.1. Study Selection

A total of 136 publications were identified in the electronic databases: 136 studies in PubMed. One hundred and twenty-six studies were excluded because they were reviews (17 studies), conference abstracts (8 studies), and protocols (7 studies); had no diabetic patients (89 studies); or were not written in English or Spanish (15 studies). Ten studies were assessed for eligibility. However, three studies did not fulfil the inclusion criteria since one was a case report; the other excluded studies were an observational study and a study not focused only on diabetic patients. Therefore, our systematic review included seven studies (Figure 1).

3.2. Characteristics of the Participants

The results obtained are summarized in Table 1, which includes the authors of each study and their year of publication, in chronological order, study design, sample size, age, and types of participants in the control group.

Table 1 shows the study design, sample size, age, severity of DFU according to the Wagner scale, age of DFU, pretreatment and duration of DFU disability level, variables, and results for each article. Table 1 also shows the variables and results analyzed for each study. A total sample size of 545 participants was included in this systematic review. The largest sample size is 120 patients and the smallest is 30 patients. The mean age was 62.95. The mean age was 62.29 in the experimental group (EG) and 63.57 years in the control group (CG). All patients had Type I or Type II diabetes and were classified between II and IV according to the Wagner scale.

Information was collected on the age of the DFUs and whether they had been previously treated. All the studies included in the review indicate that all the patients have had previous treatment and the age of the DFU ranged from 1 to 6 months. In the case of Igor et al. [19], there are already unilaterally amputated patients in which the parameters analyzed in the other clinical trials are not assessable.

3.3. Objectives of the Studies

The main objective of these studies [6, 20–24] was to evaluate the efficacy of HBOT + SC (standard of care) in chronic DFU versus SC. In addition, Clarke and Hussey [20], Fedorko et al. [22], and Duzgun et al. [24] analyzed whether amputations in patients with diabetes decreased as well as improvements in wound healing.

In contrast, Igor et al. [19] evaluated the effects of HBOT + SPR (standard prosthetic rehabilitation) in patients with unilateral lower limb amputation.

3.4. Evaluated Measures

Different measures were assessed before and after intervention. All the studies included in this review [6, 19–24] evaluated sociodemographic characteristics of the patients: age, sex, duration of diabetes, arterial hypertension, body mass index, and glycosylated hemoglobin. In addition, all patients were classified according to DFU severity using the Wagner scale. In this regard, it was not applicable to Igor et al.’s study [19] and no results have been reported. The most studied outcome measures were the characteristics of the ulcers (location, size, infection, etc.) [6, 21, 22], the rate and need for amputation [21–24], ulcer healing [20, 22, 23], and the number of ulcers that healed [20, 22, 23]. Finally, the following variables were analyzed in a single study: markers of inflammation [21], quality of life [20], absence of major amputation [20], total or partial closure of DFUs without surgery [24], and mortality rate [23].

3.5. HC Parameters Used and Intervention Characteristics

The duration of HC intervention ranged from 2 to 8 weeks. Three interventions were performed for 2 and 4 weeks, while Clarke and Hussey [20] and Fedorko et al. [22] for 6 and 8 weeks. On the other hand, the number of sessions ranged from 20 to 40 sessions. The study that performed the fewest sessions was Chen et al. [21] with a total of 20 sessions, while Salama et al. [6], Clarke and Hussey [20], and Duzgun et al. [24] performed up to 40 sessions. Two studies did not indicate the frequency of sessions.

With regard to the characteristics of the HC sessions, all the studies included in the review used a single-place chamber, except for three studies [19, 20, 22], which used a multiplace chamber. Absolute atmospheric pressure ranged between 2.4 and 2.5 atm of pressure for 85–90 min with decompression periods of 5–15 min. The longest intervention was 120 min [22], while those of Igor et al.’s study [19] and Salama et al.’s study [6] were the shortest with a time of 60 min at 1.7 and 2.5 atm of pressure, respectively.

3.6. Effects of HC on the Variables Evaluated

Table 1 depicts the variables evaluated and the effects produced by HC therapy.

The application of HC in patients diagnosed with diabetes obtained positive effects on the healing and severity of DFUs [21, 23, 24], decreased amputation rate [16, 21, 24], and improved quality of life. In this sense, Chen et al. [21] obtained statistically significant results in terms of DFU severity (p = 0.010) and amputation rate (p < 0.05). On the other hand, Fedorko et al. [22] did not obtain significant differences between groups for amputation rate (p = 0.771) nor in DFU healing (p > 0.491). Clarke and Hussey [20] only obtained significant improvements with respect to the CE group in amputation rate, but there were no significant changes in complete DFU healing.

Thus, the study of Löndahl et al. [23] did determine the complete healing of patients treated with HC (p = 0.009 and p = 0.014) in the study of Salama et al. [6].

On the other hand, Duzgun et al. [24] obtained positive effects for the group that used CH, thus reducing the need for surgical interventions.

Regarding DFU surface area, there were significant changes at the end of treatment in the CH group in terms of DFU surface area (p = 0.001), while there were no significant changes in the conventional group (p = 0.126).

Finally, Igor’ et al. [19] obtained a significantly higher increase in the percentage of arterial hemoglobin saturation in the group that received HC (p = 0.009). There were significant increase in pulse palpation frequency (p = 0.015), significant decrease in residual limb complications (p = 0.024) and residual limb strength (p = 0.000), and improvements in the Narang scale (P =0.038) and locomotor capacity index score (p = 0.048) in comparison with the control group. On the other hand, there were no significant changes for the increase in gait velocity and stride length over time on the 2-min gait scale (p = 0.081).

3.7. Risk of Bias

The mean score of the risk of bias analysis with the Evidence Project tool was 6 and scores ranged from 5 to 7 (Table 2). Higher scores corresponded to randomized controlled trial (RCT) studies (7/8) [21, 24] where assignment to experimental groups was randomized. Item by item analysis showed that assessment of the quality of the study design (Items 1, 2, and 3) was satisfactorily reached by all the studies. However, in the participants’ representativeness evaluation, more heterogeneous results were found. Item 4, which assessed the “random assignment of participants to the intervention,” was fulfilled by all the studies, while Item 5 (“random selection of participants for assessment”) was not reached for any of the studies whereas only two studies [21, 24] positively scored Item 6 (“follow-up rate of 80% or more”). Besides, in the equivalence of comparison groups, except two studies in Item 7, all the studies fulfilled the requirements. Item 8 (which referred to the “comparison groups equivalent at baseline on outcome measures”) was satisfactorily reached by all the studies.