Effectiveness of Hyperbaric Oxygen Therapy in patients with Late Radiation Tissue Injury or Diabetic Foot Ulcer

Hyperbaric oxygen therapy (HBOT) is a treatment modality where a patient placed in a pressure chamber is breathing 100% oxygen and is exposed to elevated ambient pressure. There is great variations in the access to elective HBOT in Europe. In countries offering HBOT, LRTI of the head and neck region and DFU are the most common indications. These indications are characterized by chronic tissue damage and poor micro circulation. There are few effective treatment alternatives for these indications.  In Norway, all elective HBOT is provided by Haukeland University Hospital in Bergen. Approximately 200 patients are treated per year and the most common indication is LRTI, in particular of the head and neck region, but also LRTI of the pelvic region. In Norway, HBOT is not offered for treatment of LRTI associated lymphoedema or LRTI of the central nervous system. Patients with DFU may be offered HBOT if all other treatment has failed and macro circulation in the large blood vessels is maintained, but few patients with DFU are referred to HBOT.  HBOT is considered to be relatively safe, though with some acknowledged increased risks for adverse events such as oxygen toxicity, barotrauma, visual disturbances and claustrophobia.

The objectives of this report was to assess clinical outcomes of HBOT in terms of mortality, morbidity, quality of life, need for subsequent treatment, and adverse effects in patients with LRTI or DFU compared to standard treatment or placebo. We have based the assessment on guidance provided by our handbook and questions derived from the EUnetHTA Core Model®, developed by The European network for  health technology assessment (HTA).  In 2012, we performed exhaustive searches for systematic reviews (SRs) and included the most updated SRs of high quality fulfilling our inclusion criteria. Other SRs and HTAs published in 2005 or later were inspected for additional information. In April 2014, we performed a search for RCTs published in 2012 or later. For all relevant outcomes, data were extracted and analyzed using the RevMan5 program. We used GRADE (Grading of Recommendations, Assessment, Development, and Evaluation) to report our confidence in the evidence of all predefined outcomes.

In March 2012, we identified 55 potentially relevant systematic reviews (SRs) or HTA reports covering HBOT and LRTI and/or DFU. Of these, we retrieved 24 published in 2005 or later for full text examination. We included two Cochrane-reviews updated in 2012 as the most relevant SRs of high quality answering to our inclusion criteria. The Cochrane reviews had included 15 RCTs fulfilling our inclusion criteria. We found no additional relevant RCTs to be included by other SRs. In April 2014, we retrieved 26 studies in full text and identified two additional RCTs. The assessed documentation consists of 17 RCTs with 1117 participants. We found none of the included studies to reveal statistically significant results in disfavour of HBOT. In table S1 and S2 we have provided a short summary of our major findings. In cases where we found no grounds for conclusion, we have not revealed estimates of effect in this summary, but all estimates are presented in the result part of this SR.

Table S1 Summary of findings for HBOT compared to standard therapy or placebo in treatment of patients with LRTI compared to standard therapy or placebo

Table S2 Summary of findings for HBOT compared to standard therapy or placebo in treatment of patients with DFU

The grading of quality (GRADE) reflects our confidence in estimates. Differences in populations, outcomes and comparators provided limited grounds for meta-analysis. Downgrading form high quality was primarily based on few participants, few events and risk of bias in the included studies.

Compared to standard treatment, HBOT in treatment of LRTI of the head and neck region or DFU may provide more patients with resolution or improvement of symptoms.  Elective HBOT may provide that. For other outcomes or for patients with LRTI in other locations we were unable to conclude. The identified studies are small, with few events and risk of bias. Our confidence in estimates of clinical effectiveness and efficacy is limited. New studies may change the conclusions.

Our conclusion is in line with main conclusions of the included Cochrane-reviews as well as most other systematic reviews on the same topic.

We have based our assessment on RCTs. Some of the excluded SRs and HTAs had included studies of other design than RCTs. However, a check for additional evidence in these SRs and HTAs, revealed either few and small studies or studies with obvious risk of bias. Thus, we do not believe that inclusion of non-RCTs would make us more certain in our conclusions.

None of the identified RCTs were designed to address safety. Acknowledged adverse effects are claimed to be few. However, we have no estimates on what this means. For evaluation of safety it is relevant to assess well performed non-controlled studies.

The need for further research (additional evidence generation) depends on the identified knowledge gaps being critical for advice on technology usage. Larger and more standardized RCTs may have the power to reduce the uncertainty. For HBOT in treatment of DFU one ongoing RCT (DAMOCLES) may provide answers to some questions. However, ongoing RCTs on HBOT and LRTI are small and few and they will probably not contribute to greater confidence in results.

Due to lack of confidence in estimates of effect based on the included studies, our confidence in estimates of quality adjusted life years (QUALY) and economical models based on the results will be low as well.