Hyperbaric oxygen treatment in the management of necrotising soft-tissue infections: results from a Danish nationwide registry study

Abstract

Objectives: Application of hyperbaric oxygen (HBO₂) treatment in the multidisciplinary setting of necrotising soft-tissue infection (NSTI) is debated as a considerable number of studies are of low quality with marked prognostication bias due to inadequately addressing disease severity. The objective of this study was to associate HBO₂ treatment with mortality in patients with NSTI including disease severity as a prognostic variable.

Design: Nationwide population-based register study.

Setting: Denmark.

Participants: Danish residents with NSTI patients between January 2011 and June 2016.

Primary and secondary outcome measures: Thirty-day mortality was compared between patients receiving and patients not receiving HBO₂ treatment using inverse probability of treatment weighting and propensity-score matching with predetermined variables (age, sex and weighted Charlson comorbidity score, presence of septic shock and Simplified Acute Physiology Score II (SAPS II)).

Results: A total of 671 NSTI patients were included with a median age of 63 (52-71), 61% male sex, 30% had septic shock and a median SAPS II of 46 (34-58). Patients who received HBO₂ treatment (n=266) were younger and had lower SAPS II, but a larger fraction had septic shock compared with patients not receiving HBO₂ treatment. Overall, all-cause 30-day mortality was 19% (95% CI 17% to 23%). The statistical models were in general acceptably balanced with covariates reaching <0.1 absolute standardised mean differences and patients receiving HBO₂ treatment were associated with lower 30-day mortality (OR 0.40, 95% CI 0.30 to 0.53, p<0.001).

Conclusions: In analyses using inverse probability of treatment weighting and propensity score analysis, patients treated with HBO₂ treatment were associated with improved 30-day survival.

Keywords: adult intensive & critical care; infection control; infectious diseases.

Figure 1

Balance in covariates figured as absolute standardised mean differences according to the unadjusted, nearest neighbour matched (with replacement), 1:1 matching using a within calliper of 0.25 (without replacement) and IPTW models. Dotted line represents the 0.1 threshold. IPW, inverse probability weight; IPTW, inverse probability treatment weighting; NN, nearest neighbour; SAPS II, Simplified Acute Physiology Score II.