Despite improvements in critical and surgical care over the last decade, our data shows that hyperbaric oxygen continues to be associated with significantly increased survivability in necrotizing soft tissue infections. After adjustments for confounders, the addition of HBOT to surgery was associated with decreased mortality, decreased amputation rates, and more discharges to home. However, our analysis revealed the apparent tradeoff of HBOT being associated with longer total lengths of stay and slightly higher hospital costs. In the subgroup of patients with sepsis, decreases in mortality, amputations, and non-home discharges persisted; however without the increases in length of stay and cost. One decade prior to our study, Soh et al investigated an earlier version of the NIS and similarly reported a significant reduction in mortality associated with HBO in NSTI (OR 0.49, 95%-CI 0.29-0.83) [6]. More recently Hedetoft et al’s work with the Danish national registry reported improvements in 30-day mortality (OR 0.54, 95% CI 0.33 to 0.91, p=0.02) and 90-day mortality (OR 0.61, 95% CI 0.39 to 0.97, p=0.03) [7]. Despite these findings and many other studies detailing significant improvements in patient outcomes with the addition of HBOT to NSTI, utilization rates remain remarkably low nation-wide.
There are several proposed and well-studied mechanisms by which HBOT may be effective adjective therapy in the management of NSTI. Traditionally HBOT was advocated and used mainly for clostridial infections (gas gangrene) since high oxygen tension in the tissues creates an inhospitable environment for the obligate anaerobes involved in the deadliest form of NSTI [14]. In the case of clostridial infections, there is persuasive evidence that hyperbaric oxygen therapy at sufficiently high partial pressures halts the production of alpha-toxin, buying precious time for antibiotics to take effect [15]. Modern understanding of host-pathogen interaction has advanced the use of HBOT beyond gas gangrene to various complicated chronic and acute infections [16]. Tissue hypoxia below 30mmHg impedes granulocyte and macrophage function, a situation commonly seen in NSTI as infection progresses [17]. Hyperbaric oxygen can assist by intermittently increasing the oxygen diffusion gradient, enabling normal function of the innate immune system and antibiotics as well as improving survival of threatened cells [18, 19, 20]. Despite improvements in the understanding and management of NSTI, mortality is still very high and survivors are left with debilitating amputations, tissue defects, and scars [21].
Within the subgroup of patients with NSTI and sepsis, the association with HBOT and decreased mortality, amputations, and non-home discharges persisted while the signals for both longer length of stay and increased costs were no longer seen. Several iterations of work on the role of HBOT in sepsis has found that in mice models, survivability is significantly increased with early HBOT. The proposed mechanism involves the ability of HBOT to down-regulate inflammatory cytokines such as TNF-alpha, IL-6, and modulate IL-10 leading to a reduction of the systemic inflammatory response. Indeed, cultured macrophages similarly show a decreased cytokine release when exposed to hyperbaric oxygen conditions. [22, 23, 24]. These findings strengthen the argument for the utilization of HBOT early in the disease course, especially in the most severely ill patients.
The increased cost associated with HBOT should be weighed against the financial burden of mortality and morbidity. During the Sars-COV2 (COVID19) pandemic and the resultant strain on medical resources, a renewed interest was placed on appropriate resource allocation and re-evaluating the value of a statistical life (VSL) through the lens of finite medical resources. While the VSL is not a perfect corollary as its main utility is comparison of wage paid versus occupation risk, it is nevertheless a useful benchmark when comparing, in broad strokes, the financial viability of a potentially lifesaving treatment. For the purposes of risk analysis, the United States government values a statistical life at $7.4 million USD [25]. By this framework, an intervention that reduces risk by half, for instance, would be considered cost-effective if it cost less than $3.7 million dollars. In addition to mortality, amputation is a common complication of NSTI leading to significant short and long term morbidity for patients. Gomez et al looked at the overall cost of amputations from workplace accidents and reported that, depending on anatomy of amputation and number of amputations, costs may be as high as $46,000 and result in several months of lost work days [26]. Our study found an association with significant decrease in the odds of both mortality and amputation among NSTI patients treated with HBOT. Though our study found an estimated increase in costs of $7,800 in patients who received HBOT, and while a detailed financial analysis is outside the scope of this publication, the relatively small difference in hospital costs for a significant reduction in death and amputations make a strong case for the cost-effectiveness of HBOT in NSTI.
Our study does have several limitations inherent to its nature as a large retrospective review of a national database. Temporal relations within specific hospitalizations are difficult to ascertain, for example timing of HBOT from presentation, timing of surgery, etc. Data was not available with regards to HBO treatment protocol (duration, treatment depth, number of treatments completed) nor whether chambers were multiplace or monoplace. In fact, this shortcoming may have actually served to dampen the positive results of HBOT, as there may be patients with only one treatment or delayed HBOT included in the treatment group. With the retrospective nature of this work, it is impossible to guarantee standardization of any treatment modality (HBOT, surgery, antibiotics, etc) between patients. Sepsis diagnosis is based on ICD-10 as registered; the register does not contain any information on adherence to current sepsis guidelines but the overall low mortality rate suggests that most patients did not have septic shock [11]. From experience and a previous prospective cohort study, we know that there is an inherent selection bias for the administration of HBOT depending on multiple factors such as availability of ICU resources and hyperbaric chambers at the receiving hospital, and hemodynamic stability of the patient [27]. Our data does not include information on disease severity, such as sequential organ failure assessment score (SOFA) or hemodynamic parameters. Patient groups were heterogeneous in several aspects, though best efforts were made to account for this through statistical normalization and modeling. Lastly, complications of HBOT can be vague, nebulous, and shared with several diseases or iatrogenic processes making it difficult and ill-advised to draw useful conclusions regarding their rate or severity from such a large database.
Our study has several strengths worth highlighting. The NIS allows us to amass a cross-section of real patients from across the country to evaluate real-world application of HBOT in NSTI rather than animal models or restrictive homogenous patient populations. Our robust multivariate model allows adjustment for numerous potential confounders between our two patient groups. Lastly, we follow in the footsteps of Soh et al [6] and by employing the same large, nationwide database of hospital admission for NSTI we are able to evaluate the association between HBOT and NSTI through time and advances in surgical and critical care.
Necrotizing soft tissue infections can be rapidly progressive, life threatening, and often leave patients with chronic disfigurements. While the medical field agrees regarding the critical role of prompt antibiotics and surgical debridement, disagreement as to the role of HBOT remains pervasive [3, 4]. Our study represents the largest single study of the role of HBOT in NSTI to date. We found that after adjusting for relevant confounders, HBOT added to surgery and antibiotics was associated with decreased risk of mortality, amputations, and non-home discharges with the trade-off of slightly increased costs and longer lengths of stay. In a subset of patients with NSTI and sepsis however, improved rates of mortality, amputations, and non-home discharges persisted, without the signal for increased costs or lengths of stay. To gain a more comprehensive understanding of the role of HBOT in NSTI, with the ultimate aim of improving patient outcomes and changing evidence-based clinical practice, a well-designed randomized controlled, multicenter trial is warranted.