Inglês (pdf)
Artigo em XML
Referências do artigo
Tradução automática
Permalink
INTRODUCTION
Diabetes is a major public health problem that affects more than 400 million adults worldwide1. Furthermore, the late referral of patients with foot ulcers to specialized diabetic foot services is also a common problem worldwide2,3.
Chronic wounds are defined as those that fail to heal through an orderly and timely series of events that produce durable structural and functional closure4. Furthermore, diabetic foot wound is one of the most frequent and serious complications associated with diabetes, affecting around 15% of all patients and resulting in around 50% of all lower limb amputations5.
Chronic wounds are mainly characterized by low partial oxygen pressure in the center of the wound, hindering healing6. In hyperbaric oxygen therapy (HBOT), the patients breathe in 100% oxygen at a pressure two– three-fold greater than the atmospheric pressure in a hyperbaric chamber, which increases oxygen tension in arteries and tissues7,8,9. Some studies show that HBOT can increase neovascularization, stimulate stem cells and growth factors, and inhibit the inflammatory response, thus helping wound healing10,11. While some researchers reported that HBOT had greater effectiveness in healing foot wounds in diabetic patients compared to the standard treatment12,13,14,15,16,17,18,19,20, others reported no such differences21,22,23. In addition, some studies reported reduced amputation rates with HBOT in diabetic patients14,24,25,26, while others showed no difference21,23,27.
Few trials evaluate the quality of life in diabetic patients with foot wounds treated with HBOT. Li G et al. reported no significant improvements in quality of life with this therapy28, while Londahl et al. stated that HBOT positively impacted these patients’ quality of life13.
The main objective of this study was to evaluate if adding HBOT to the standard chronic diabetic foot ulcer treatment improves wound healing more than the standard treatment alone. Secondary objectives were to evaluate whether HBOT reduces the number of major (transtibial and transfemoral) and minor (toes and forefoot) amputations and to test its effect on the quality of life in these patients.
METHODS
The study was approved by the research ethics committee of the ABC-SP Medical School (FMABC- SP), Brazil, and registered in the Brazilian Clinical Trials Registry (ReBEC) under number RBR-7bd3xy, with the protocol Does Hyperbaric Oxygen Therapy Facilitate the Healing of Chronic Foot Wounds in Diabetic Patients? This study followed the Declaration of Helsinki, and all participants signed an informed consent form.
This was a randomized clinical trial in which the Consort checklist was used. The study was conducted in Imperatriz, MA, Brazil, and included diabetic patients with chronic foot wounds classified as Wagner grades 2, 3, and 429 that persisted for more than 1 month. The patients were followed up at the SUS Diabetic Foot outpatient clinic and underwent HBOT sessions at the CicatrizAR Clinic, as described in the study protocol30.
Sample and study period
The sample consisted of diabetic patients with chronic foot wounds. From 2019 to 2023, half of them underwent the standard treatment with dressings, debridement, antibiotics, and load relief associated with HBOT (HBOT group), and the other half underwent only the standard treatment (control group).
The sample size was calculated using the formula to compare two independent groups according to qualitative variables31, with 95% confidence interval and 80% power. A wound healing rate of 90% was considered achievable in the HBOT group in 1 year, which we expected to be at least 20% higher than that in the control group (absolute difference, i.e., no more than a 70% wound healing rate was expected in the control group). Thus, the total sample size was estimated as 60 patients per group (120).
Eligibility criteria
The inclusion criteria were adult patients (aged > 18 years); stable clinical condition, type 1 and 2 diabetes; Wagner grades 2, 3, and 4-foot ulcers; ulcers persisting for more than one month without healing; study authorization; and patients of the SUS (Unified Health System).
The exclusion criteria were macroangiopathy (two absent distal pulses), absolute or relative contraindications to HBOT, chemotherapy with bleomycin, chronic obstructive pulmonary disease, previous spontaneous pneumothorax, chronic sinusitis, chronic otitis media, unstable angina, severe congestive heart failure, claustrophobia, severe dementia, depression, or history of seizures.
Randomization
A randomized, parallel, two-arm, non-blind, controlled clinical trial. Randomization was done by the research coordinator using a simple 1:1 draw (for each pair of patients, one participant selected the number 1 or 2, being allocated to one group, and the second patient to another group). Randomization did not use computer- generated random numbers. The patients were selected at the SUS diabetic foot clinic and underwent regular foot wound follow-up. They were evaluated according to the inclusion criteria and invited to participate in the research if the criteria were met. The researchers explained how the study would be carried out. After 1 week, patients interested in participating in the study signed the informed consent form, which the research coordinator collected.
Interventions
The patients of the HBOT group were evaluated at admission after 10, 20, 30, and 35 HBOT sessions. After 6 months and 1 year, the control group was also assessed at equivalent periods (upon admission, after 2, 4, 6 and 7 weeks, 6 months, and 1 year) to clinically evaluate the ulcers and perform specific measures using the software ImageJ, developed at the National Institutes of Health (NIH, Bethesda, Maryland). Wound progression and particular treatments such as appropriate dressing, antibiotic therapy, or the need for surgical intervention were assessed weekly. The HBOT sessions were conducted 5 days a week, carried out with patients in a multiplace chamber at 2.5atm absolute (ATA) and 100% O2, with 10min for compression and decompression, effective sessions of 90min. Both groups received dressings with Exufiber Ag® foam (Molnlycke, Gotemburgo, Suécia) and Mepilex® (Molnlycke, Gotemburgo, Suécia), which were chosen according to wound characteristics, with a silver dressing for infected lesions and a protective polyurethane foam for wounds in granulation. With regard to load relief, patients in both groups wore orthopedic Baruk shoes.
The SF-36 quality of life questionnaire was administered to both groups at admission and after a 3-month follow-up. Patients of both groups underwent laboratory tests at admission, such as a haemogram, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), creatinine, fasting blood glucose, and glycosylated hemoglobin.
Primary outcome
Wound healing was assessed by evaluating the diameter of the lesions using specific software and periods, as described in the interventions. Healing was achieved when there were no more skin lesions (Wagner grade 0), and the primary endpoint was a binary result.
Secondary outcomes
Amputation rates and the reduction of lesions that do not heal were assessed, with statistical significance evaluated between the groups. The domains of the SF-36 quality of life questionnaire were also evaluated in both groups. The data from this questionnaire upon admission and after 3 months of follow-up were compared.
Data analysis
The collected data were stored in a Microsoft Excel 2016 spreadsheet format. After checking for errors and inconsistencies, a descriptive analysis was carried out using central tendency measures and variability, in addition to absolute and relative frequencies for all sociodemographic and clinical variables between groups.
The Chi-square test was used to verify the association between sociodemographic and clinical variables in the groups.
Quantitative variables were compared between and within groups by the Shapiro-Wilk normality test and the Levene’s homogeneity of variance test to evaluate the assumptions, otherwise corresponding non-parametric tests would be performed. The Student’s t-test was used to evaluate increased SF-36 domains between groups. The Wilcoxon test compared the domains before and after the experiment and within each group. The Mann–Whitney test was used to evaluate laboratory test results in the group.
Generalized Estimating Equation (GEE) models were calculated to compare the wound area in the control and HBOT groups at each moment and within each group over time (weeks). The Bonferroni test was used for comparisons significantly different at 5%. The GEE model was proposed by Zeger and Liang32, being suitable for continuous responses and repeated measurements and reflecting the relationship between variable and independent responses considering the correlation between measurements at each moment33. All analyses were performed at 5% significance in the IBM SPSS® program, Version 24.0, 2016 (IBM, Armonk, NY, USA).
RESULTS
The study had 43 eligible participants, of whom 30 patients were randomized: 13 in the HBOT group and 17 in the control group, as shown in the Consort flow diagram in figure 1.
Considering the basic characteristics, there was no significant difference between groups, both including type 2 diabetes patients, most of whom already undergoing previous amputation, with a predominance of Wagner grade 3 wounds. Laboratory tests showed no difference between groups; therefore, infectious or inflammatory changes or even decompensated diabetes showed no difference between groups, as shown in table 1.
Table 1 Baseline Characteristics of HBOT and Control Groups
| Control group (n = 17) | HBOT group (n = 13) | p-value | |||
|---|---|---|---|---|---|
| Sex | |||||
| Men | 10 | 58.8% | 8 | 61.5% | 0.88* |
| Women | 7 | 41.2% | 5 | 38.5% | |
| Age | |||||
| Mean | 56.00 | 55.00 | 0.66** | ||
| Standard deviation | 13.00 | 8.00 | |||
| BMI | |||||
| Mean | 25.03 | 27.72 | 0.27** | ||
| Standard deviation | 6.44 | 6.11 | |||
| Diabetes | |||||
| Type 1 | 0 | 0.0% | 0 | 0.0% | **** |
| Type 2 | 17 | 100.0% | 13 | 100.0% | |
| Medications | |||||
| Oral antidiabetic agents | 12 | 70.6% | 10 | 76.9% | 0.70* |
| Insulin | 5 | 29.4% | 3 | 23.1% | |
| Comorbidities | 15 | 88.2% | 10 | 76.9% | 0.41* |
| Hypertension | 9 | 52.9% | 4 | 30.8% | 0.23* |
| Cardiovascular disease | 1 | 5.9% | 0 | 0.0% | 0.37* |
| Renal disease | 1 | 5.9% | 0 | 0.0% | 0.37* |
| Diabetic retinopathy | 4 | 23.5% | 2 | 15.4% | 0.58* |
| Previous stroke | 1 | 5.9% | 0 | 0.0% | 0.37* |
| Previous AMI | 1 | 5.9% | 0 | 0.0% | 0.37* |
| Previous amputation | 10 | 58.8% | 8 | 61.5% | 0.88* |
| Wound classification | |||||
| Wagner grade 2 | 7 | 41.2% | 3 | 23.1% | 0.34* |
| Wagner grade 3 | 9 | 52.9% | 10 | 76.9% | |
| Wagner grade 4 | 1 | 5.9% | 0 | 0.0% | |
| Haemoglobin (g%) | |||||
| Mean | 10.56 | 10.98 | 0.44** | ||
| Standard deviation | 1.58 | 1.23 | |||
| White blood cell count (mm3) | |||||
| Mean | 8,281.76 | 8,714.62 | 0.84*** | ||
| Standard deviation | 1,710.71 | 3,386.04 | |||
| ESR (mm) | |||||
| Mean | 18.01 | 33.12 | 0.18*** | ||
| Standard deviation | 17.89 | 40.68 | |||
| CRP (IU/m:) | |||||
| Mean | 3.32 | 5.41 | 0.08*** | ||
| Standard deviation | 2.85 | 6.24 | |||
| Fasting blood glucose (mg/dl) | |||||
| Mean | 167.80 | 162.05 | 0.84** | ||
| Standard deviation | 73.91 | 74.08 | |||
| Creatinine (mg/dl) | |||||
| Mean | 1.50 | 0.97 | 0.09** | ||
| Standard deviation | 1.06 | 0.25 | |||
| Glycosylated haemoglobin (%) | |||||
| Mean | 8.72 | 9.48 | 0.39** | ||
| Standard deviation | 2.37 | 2.43 | |||
*Chi-square test. **Student’s t-test. ***Mann-Whitney test. ****Value could not be calculated.
Source: the authors (2024).
Wounds were reduced over time in both groups, while the reduction was stronger in the HBOT group with a statistically significant difference between 1- and 6-month follow-up periods, which suggests that HBOT increases wound healing. However, this difference was not significantly reproduced at the end of the follow-up period, as shown in table 2.
Table 2 Wound area comparisons over time
| Wound area | Control group (n = 15) Mean ± standard deviation | HBOT group (n = 13) Mean ± standard deviation | p-value |
|---|---|---|---|
| Initial wound area (cm2) | 20.93 (22.49) | 22.23 (17.85) | 0.87 |
| Wound area after 10 HBOT/2 weeks (cm2) | 16.37 (19.10) | 12.60 (9.75) | 0.12 |
| Wound area after 20 HBOT/2 weeks (cm2) | 12.06 (14.41) | 7.01 (6.28) | 0.04 |
| Wound area after 30 HBOT/2 weeks (cm2) | 10.14 (13.29) | 3.49 (3.97) | 0.02 |
| Wound area after 35 HBOT/2 weeks (cm2) | 8.17 (12.28) | 2.13 (3.13) | 0.03 |
| Wound area after 6 months (cm2) | 2.35 (5.65) | 0.82 (2.47) | 0.04 |
| Wound area after 1 year (cm2) | 0.99 (3.01) | 0.32 (1.14) | 0.12 |
| p-value | < 0.001 | < 0.001 |
Different letters in the column indicate different mean wound areas over time according to the Bonferroni test at 5% significance.
Assessing wound healing, we had a positive trend for the HBOT group, with a statistically significant difference below 6 months, but not maintained up to 1 year, as shown in table 3. However, the mean time needed for healing was less than half in the HBOT group, (20.33 weeks for Clinical Group vs. 9.67 weeks for HBOT Group), with statistical significance (p = 0.005).
Table 3 Wound healing by group
| Time | Clinical Group (n= 15) | HBOT Group (n=13) | p-valor* |
|---|---|---|---|
| n (%) | n (%) | ||
| Wound healing _10 OHB/2 weeks(cm2) | 0/15 (0.00) | 0/13 (0.00) | ** |
| Wound healing _20 OHB/4 weeks (cm2) | 0/15 (0.00) | 3/13 (23.1) | 0.04 |
| Wound healing _30 OHB/6 weeks (cm2) | 0/15 (0.00) | 5/13 (38.5) | 0.01 |
| Wound healing _35 OHB/7 weeks (cm2) | 1/15 (6.67) | 6/13 (46.2) | 0.02 |
| Wound healing _6 months (cm2) | 10/15 (66.7) | 11/13 (84.6) | 0.10 |
| Wound healing _1 year (cm2) | 11/15 (73.3) | 12/13 (92.3) | 0.19 |
*Chi-square test. **No statistics were calculated because wound healing is a constant.
Both groups improved their quality of life domains; however, this improvement was greater in the HBOT group, with statistically significant differences in almost all domains, except mental health, which did not occur in the control group, as shown in table 4.
Table 4 Comparison of SF-36 Score Before and After the Intervention in HBOT and Control Groups
| Control group (n = 17) | HBOT group (n = 13) | |||||
|---|---|---|---|---|---|---|
| Before Mean | After Mean | p-value | Before Mean | After Mean | p-value | |
| Functional capacity | 10.88 | 40.00 | 0.003 | 16.54 | 85.38 | 0.001 |
| Physical limitation | 1.47 | 20.59 | 0.02 | 0.00 | 75.00 | 0.001 |
| Pain | 45.47 | 60.76 | 0.08 | 43.31 | 95.54 | 0.001 |
| General status | 27.12 | 45.27 | 0.02 | 32.46 | 67.69 | 0.006 |
| Vitality | 29.71 | 33.24 | 0.11 | 41.15 | 72.69 | 0.008 |
| Social aspects | 12.50 | 57.50 | 0.01 | 9.62 | 83.65 | 0.002 |
| Emotional aspect limitations | 1.96 | 35.29 | 0.006 | 10.26 | 100.00 | 0.001 |
| Mental health | 35.29 | 25.59 | 0.80 | 43.85 | 59.23 | 0.07 |
| GENERAL | 20.55 | 49.54 | 0.001 | 24.65 | 79.90 | 0.001 |
Wilcoxon Test. Source: the authors (2023).
In terms of amputations, there were only two minor amputations in the control group (2/15 patients, 13.3%) and none in the HBOT group, but without statistically significant difference between groups – p-value 0,60.
DISCUSSION
There are current controversies regarding best strategies to improve foot wound healing in diabetic patients; therefore, this study was designed to compare the existing standard treatment for chronic foot ulcers in diabetic patients, i.e., dressings, debridement, antibiotics, and load relief and the combination treatment with HBOT. Some studies demonstrated greater effectiveness of adjuvant HBOT compared to placebo12,13,14,15,16,17,18,24, while other researchers reported no benefit21,22,23,34. According to some authors, HBOT also reduces the likelihood of amputation14,24,25; however, others have shown no benefit21,23,27.
A systematic review of randomized clinical trials published by the Cochrane Collaboration35 reported significant short-term wound healing improvement (i.e., 6 weeks), but found no statistically significant difference in long-term wound healing and major or minor amputation rates regarding HBOT, suggesting the need for more randomized studies to clarify these doubts.
This randomized trial showed a very similar distribution of basic characteristics between groups, with prevalence of men and type 2 diabetes, corroborating other randomized studies15,17,21,36; however, the mean age was lower than that in these related studies. Regarding the foot wound classification in diabetic patients, Wagner grade 3 is the predominated type, as seen in other studies17,21. About 60% of patients had a previous amputation, and only 6.6% of amputations occurred during the study, which suggests that the best study period to evaluate the possible amputation reduction would be the acute phase of foot wounds in diabetic patients and not in the chronic and outpatient phase, where the chance of this outcome is lower.
Previous randomized clinical trials are heterogeneous in several aspects, such as follow-up time, number of HBOT sessions, pressurisation time, pressure level, sample size, and use of placebo or just the standard treatment in the control group. Thus, we will analyse these different aspects in trials conducted in this century.
Abidia A et al. (2003) published a double-blind randomized-controlled trial with 16 patients with diabetes and Wagner grade 1 and 2 ischaemic ulcers treated in a multiplace chamber and followed up for 1 year. They used a protocol of 30 HBOT sessions at 2.4 ATA for 90 min, concluding that HBOT improved wound healing in this group of patients, despite the data presented in table 2 of their article, but they suggested the need for a more robust multicenter study to confirm their findings16. This occurred in 2017, when Santema KTB et al. presented the results of the DAM02CLES study, conducted in 24 services of Netherlands and Belgium, which assessed ulcers in ischaemic diabetic patients for 1 year. The authors initially planned to analyze 226 patients but, for logistical reasons, reduced the sample size to 120 patients, with 95 participants assessed after 30–40 HBOT sessions at 2.5 ATA for 90 min, and the control group is the standard treatment. They concluded that HBOT did not improve healing or reduce amputations in this population36.
Kessler L et al. (2003) published a trial with 27 non-ischaemic diabetic patients with Wagner grade 1–3 ulcers followed up for 4 weeks. The control group was the standard treatment, using a protocol of 20 HBOT sessions at 2.5 ATA for 90 min. They concluded that HBOT doubled wound healing in that sample; however, this difference was only observed until the second week15. Duzgun AP et al. (2008) presented a study with the same follow-up and control group as Kessler L et al. but with 100 patients and a different HBOT protocol, which included two sessions a day alternating with one session on the following day, at 2.0 ATA for 90 min. They reported a positive outcome for wound healing and amputation reduction37.
A double-blind, randomized controlled trial by Londahl M et al. (2010) analyzed 75 patients with Wagner grade 2–4 ulcers followed up for one year, who underwent 35–40 HBOT sessions at 2.5 ATA for 85 min. They concluded that HBOT helped heal foot wounds in diabetic patients17. In 2011, another publication by the same authors reported a positive quality-of-life outcome in that trial13. Fedorko L et al. (2016) conducted a double- blind, randomized controlled trial with 87 patients with Wagner grade 2–4 ulcers and a 3-month follow-up. They used 27–30 HBOT sessions at 244 kPa for 90 min, concluding that HBOT did not decrease the indication for amputations in diabetic patients with chronic ulcers21. In another publication, they also presented a quality-of- life outcome, reporting no improvement28. The trial by Fedorko L et al. was severely criticized. Lindahl et al. (2016) published a comment in the same journal as the original article contesting several items, especially their conclusion38.
Subsequently, new trials followed up the participants only throughout the HBOT sessions, reporting improved foot wound healing in diabetic patients. Chen CY et al. (2017) analysed 38 patients with Wagner grade 1–3 ulcers undergoing 20 HBOT sessions at 2.5 ATA for 120 min18. Salama SE et al. (2019) analysed a final sample of 30 patients with Wagner grade 2–3 ulcers undergoing 20–40 HBOT sessions at 2.5 ATA for 60 min (20). Kumar A et al. (2020) analysed 54 patients after 36 HBOT sessions at 2.4 ATA for 90min19.
As for other studies, some included patients with Wagner grade 5 ulcers, which has a primary indication at least for major amputation at below-the-knee level39,40. Margolis DJ et al. (2013) published the most significant cohort, a multicenter study analyzing 6,259 patients: 5,466 (87.3%) in the non-HBOT group and 793 (12.7%) in the HBOT group. That study used a propensity score to obtain a distribution as close as possible to the one used in randomized studies; however, table I of that study showed that Wagner grade ≥ 3 ulcers had a percentage distribution of 18.4% in the non-HBOT group and 45.7% in the HBOT group. This difference may have influenced the results, which showed no healing improvement or amputation reduction, with worse outcomes and up to three times more chance of amputation or non-healing in the HBOT group, not corroborating other studies22. Thus, we can observe several differences between the studies that make it difficult to compare them. Furthermore, some studies showed a distribution of patients with more severe injuries and a higher score in the Wagner classification in the HBO group, which may have interfered with their results.
In our trial, the analysis of foot ulcer improvement in diabetic patients showed a positive trend, with a statistically significant difference below six months in the HBOT group, but with no statistical difference after one year, corroborating other studies15,16,18,19,20,37. As for reduced amputations, there was no major amputation in the groups, only two minor amputations in the control group, with no statistically significant difference. As previously mentioned, we had insufficient data to assess this outcome, and we believe that the most appropriate period for this assessment would be in the acute phase and in hospitalized patients, not in stable outpatients.
Regarding quality of life, there are only three trials; two reported no difference in the SF 36 in patients in the HBOT group compared to those in the control group28,16. However, our study showed improved quality of life in the HBOT group, corroborating the study by Londahl et al., who also found improved quality of life with HBOT13. However, in our research, there was a statistically significant improvement in the quality of life of patients undergoing HBOT in almost all domains of the SF-36, except mental health. At the same time, Londahl et al. also had an improvement related to this domain, possibly due to a more extended follow-up period13. Thus, social interaction during treatment sessions and improved healing of lesions impact the quality of life of these patients, as cited by Londhal et al.13
In our study, there were no adverse events, such as hypoglycemia, dizziness, barotrauma, or fainting, as occurred in the studies by Londahl et al. and Fedorko L et al.17,21 These adverse events may have occurred in these studies because they used placebo in the control group, thus exposing more patients to the hyperbaric environment. They also had a higher average age of patients than ours, which facilitates this occurrence.
Our main limitation was the small sample size due to recruitment interruption during the COVID-19 pandemic when the diabetic foot outpatient clinic was closed and used to support care for patients with COVID-19 for almost two years. Furthermore, when it was reopened, its pre-pandemic configuration was lost, and it did not have the same flow, hindering sample size recovery. In addition, this study was a non-blind RCT design for all participants, which is known to carry a greater risk of selection bias. However, since the essential characteristics of the groups were similar, we believe this bias was minimized.
CONCLUSION
Our findings suggest that HBOT has a positive impact on the healing of foot wounds in diabetic patients with less than 6 months of follow-up. However, data are insufficient regarding the method’s benefit in reducing minor or major amputations. In terms of quality of life, we obtained results that support the use of HBOT. Further studies with a larger sample could help clarify these doubts.
