Hyperbaric oxygen therapy is a recognised accepted adjunct to surgical debridements, antibiotic therapy and maximal goal-directed critical care therapy for infections of soft tissues resulting in necrosis. A number of clinical scenarios, specific lesions and syndromes have been described over the years, based on the affected tissues and location of infection, the etiologic organism or combination of organisms involved in the infection, and particular host immunologic and vascular risk factors. In all of these clinical situations, there appears to be the common denominator of the development of hypoxia resulting in necrosis.
Hypoxia is known to impair phagocytosis by polymorphonuclear leukocytes. After an infective process is initiated, metabolic products of aerobic and anaerobic metabolism tend to lower the oxidation-reduction potential (Eh), leading to a drop in pH, which creates a milieu for growth of strict and facultative anaerobic organisms. When the blood supply to the skin is affected by involvement within a phlegmon, with oedema and necrosis in the deep fascial layers in which they reside, the decreased perfusion pressure and ischaemia predispose to rapid progression and advancement of the infectious process within the skin and subcutaneous tissues, exacerbated by the dysfunctioning polymorphonuclear leukocytes. Local hypoxia occurs, with up-regulation of endothelial adherence molecules, resulting in leukocyte adhesion and endothelial cytotoxicity. Leukocytes may become sequestered in vessels, impairing local immunity, and incomplete substrate oxidation results in hydrogen and methane accumulation in the tissues. Tissue necrosis occurs, with purulent discharge and gas production. Quantities of gas within tissues are frequently seen in gas gangrene, crepitant anaerobic necrotising cellulitis, and necrotising fasciitis.
Hyperbaric oxygen therapy can reduce the amount of hypoxic leukocyte dysfunction occurring within an area of hypoxia and infection, and provide oxygenation to otherwise ischaemic areas, thus limiting the spread and progression of infection. The diffusion of oxygen dissolved in plasma in the circulation, where it is initially carried in large vessels, proceeds to areas of poorly perfused tissue, from regions of very high O2 saturation down a gradient to lower oxygen levels in tissue. Integrin inhibition decreases leukocyte adherence, reducing systemic toxicity.
In cases where the antibiotic being used requires oxygen for transport across cell walls, hyperbaric oxygen therapy can act to enhance antibiotic penetration into target bacteria. Enhancement of the post-antibiotic effect by hyperbaric oxygen has been demonstrated for amino-glycosides and Pseudomonas.
Clinical classification of the necrotising infections of soft tissues is easiest early in the course of infection, when anatomic levels of involvement of skin, superficial or deep fascia, and muscle involvement can be assessed either during exploration, on punch biopsy, or by radiologic investigation. However, as infection progresses, distinction between some of the clinical entities may become blurred as full thickness necrosis extends into muscle late, after having extended through skin, fat, fascia, and into muscle via direct extension of infection. At presentation, it may be difficult to differentiate these necrotising soft tissue infections one from another, or from Clostridial myositis and myonecrosis, until either Gram stain or cultures are available. Considering their historical differences and evolution, it remains useful to examine the separate categories of infection separately in order to anticipate pathways of extension of infection, anticipate complications, and identify when adjunct hyperbaric oxygen therapy should be considered.
Clinical Entities: Necrotising Fasciitis
Introduction: Necrotising fasciitis is an acute, potentially fatal infection of the superficial and deep fascia of the skin and soft tissues, which progresses to ischaemic dermal necrosis after involvement of the dermal blood vessels which traverse the fascial layers. The popular media refer to this entity as infection with "Flesh-eating bacteria.”
Necrotising fasciitis was initially described and named "haemolytic streptococcal gangrene” by Meleney in 1924. He described an illness characterised by gangrene of subcutaneous tissues, followed by rapid necrosis of the overlying skin from involvement of the blood vessels supplying the skin, which are found in the affected fascial layers. All his patients grew haemolytic streptococci on cultures, and the patients were all seriously ill. Surgical extirpation appeared to be the therapeutic approach. Reference to this entity as necrotising fasciitis appears around the time of the report by Wilson.
The characteristic level of infection is at the deep fascia. Because infection with necrosis is noted to spread along fascial planes deep to the skin, it is not an uncommon event for there to be minimal skin signs early on. Pain out of proportion to findings could be an early tip off to the presence of deep fascial infection. Since blood vessels supplying overlying skin travel thru fascia, it is the involvement of these vessels by infection that leads to rapid progression to dermal necrosis. Microbiologically, groups A, C, or G beta-haemolytic streptococci can be isolated from tissue specimens in 50 to 90% of case series, with one or two more organisms often also accompanying the streptococci in up to half the cases. The occurrence of Staphylococcus aureus plus anaerobic streptococci is also known as Meleney's synergistic gangrene. Commonly isolated organisms include Enterobacteriaceae, Enterococci, Bacteroides species, Peptococcus species. Candida species have also been reported. Necrotising fasciitis is also reported to be caused by community-acquired strains of methicillin-resistant Staphylococcus aureus (CA-MRSA) alone. In many cases, infection is poly-microbial, with Enterobacteriaceae and anaerobes frequently isolated.
The most common risk factors associated with necrotising fasciitis are traumatic breaks in the skin, most commonly lacerations, insect bites, burns, deep abrasions, puncture wounds, or following surgery, particularly those involving bowel perforations. Diabetes appears to be a strong risk factor, as are obesity, alcoholism, smoking, and intravenous drug abuse. Reports of necrotising fasciitis as a result of infection of otherwise typical lesions of chickenpox have been published. An association with the use of non-steroidal anti-inflammatory agents has also been suggested. NSAIDs are cyclo-oxygenase inhibitors and may have an adverse effect on neutrophil killing and cell-mediated immunity. NSAIDs are reported to inhibit monocyte superoxide production.
Most common sites of occurrence of necrotising fasciitis are the lower extremities, while an increased incidence in the upper extremities is seen in the parenteral drug abuse population. However, any location of the body can be affected, including the abdominal wall of neonates, in association with omphalitis. Involvement of the scrotum and perineum in the male is known as Fournier's Gangrene, which is essentially necrotising fasciitis of the superficial perineal fascia, also known as Colles' fascia; which can spread infection to the penis and scrotum via Buck's fascia or Dartos' fascia; or Scarpa's fascia, which connects to, and can spread infection to, the abdominal wall. Perianal or perirectal infection may also spread into these areas, and undrained or inadequately drained perirectal abscesses are often cited as a source of Fournier's Gangrene. Perineal necrotising fasciitis can also occur in the female. Diabetes mellitus remains a strong risk factor in this particular form of necrotising fasciitis as well. Fournier's Gangrene is more likely to have multiple mixed organisms cultured, particularly Enterobacteriaceae, Group D streptococci, and anaerobic organisms, such as Bacteroides fragilis.
The patient with necrotising fasciitis will typically present with an acute combination of pain and swelling, which may or may not be accompanied by fever and chills. There may already be a focus of cellulitis apparent, but in some instances early on, there may be very few skin changes. In some patients, there may be pain out of proportion to the skin findings, which may not be unexpected considering that the initial level of infection is the fascia, not necessarily the skin. In others, manifestations of a large phlegmon may be quite obvious, although at times the area of infection may have been assumed to be cellulitis and not a more serious form of infection. Pain may proceed to numbness, as a result of compression of nerves which also pass through the fascia. With time however, the infection will rapidly proceed to cause areas of blistering and bullae formation. Hints of darkening of the skin may appear as perfusion decreases, until obvious areas of dermal ischaemia appear, making the skin appear dusky, greyish or frankly black. Upon exploration of the process, a clinical diagnosis can be confirmed at the time of biopsy or debridement, when the fascia is grossly observed by the surgeon to be necrotic, and will give way easily to a probing finger or surgical clamp, giving the sensation of "thunking” of the skin against the underlying muscle layers, instead of remaining tight and crisply defined. It has been suggested that limbs of patients with necrotising fasciitis, as opposed to those with cellulitis only, may be observed to have markedly reduced tissue oxygen saturations as measured by near-infra-red spectroscopy throughout the involved site, with oxygen saturations in the 52% ±18% range, compared to control measurements of 86% ±11% in uninvolved sites.
In the neonate, necrotising fasciitis of the abdominal wall can be seen as a complication of omphalitis in 10 to 16% of cases, and appears to carry over a 50% mortality rate even when treated with aggressive debridement of involved skin, subcutaneous tissue and fascia.
A number of diagnostic observations have been made to enable confirmation of the diagnosis of necrotising fasciitis. Frozen section soft-tissue biopsy early in the evolution of a suspect lesion may provide definitive diagnosis. CT scan findings are also revealing. Asymmetrical fascial thickening that was at least twice the contralateral side and associated with fat stranding was seen in 80% of 20 patients with necrotising fasciitis. Gas tracking along fascial planes was seen in 55%, characteristically did not involve muscle and was not associated with abscess formation. The authors note that the areas of black, gangrenous skin were far smaller than the widespread infection in the underlying fascial planes. Also of note was that 7 of the 20 patients had associated deep space abscesses requiring immediate surgical drainage, which demonstrates the need for CT studies to assess extent of disease, particularly in patients who do not appear to be responding to therapy.
Magnetic resonance imaging (MRI) also demonstrates the extent of affected tissue well, is able to differentiate fluid and gas through differential signal intensities, and is useful in differentiating cellulitis from necrotising fasciitis, after injection of gadolinium contrast. But in a study of 15 patients, MRI overestimated the extent of deep fascial involvement in one patient who only had cellulitis, following IM injections which showed up on MRI as thickening of both superficial and deep fascia of the deltoid muscle.
Cultures of deep tissue at the time of debridement for aerobes, anaerobes and fungi, are imperative as up to 75% of patients in some series' have demonstrated polymicrobial aetiologies. Fungal cultures are particularly important in the immunocompromised, diabetic and cancer populations and in patients who have not responded to standard anti-bacterial antibiotics.
Amputation rates of 26% up to 50% are reported in cases of necrotising fasciitis of the extremities, without hyperbaric therapy. Mortality in reported series range from 16.9% up to 66% without the use of hyperbaric oxygen. Mortality is often associated with delayed diagnosis, underlying immunocompromise, and underlying heart disease, degree of leukocytosis, septic shock and severe underlying metabolic abnormalities.
In the neonate, necrotising fasciitis is reported as a complication of omphalitis, balanitis, mastitis, postoperative complication, and foetal monitoring. 4 of 6 cases found in a literature review who received hyperbaric oxygen therapy survived, while the overall mortality rate was 39/66 (59%). In a group of neonatal omphalitis patients with abdominal wall necrotising fasciitis reported from Children's Hospital in Los Angeles, 7 out of 8 cases died, for a mortality rate of 87% without hyperbaric oxygen therapy. In a series of 32 cases of omphalitis from Seattle over a 10-year period, 7 developed necrotising fasciitis, and 5 of the 7 died. The 2 patients who did survive, out of the 4 who had hyperbaric oxygen treatments, were noted to have resolved their systemic sepsis more rapidly, and had healthier granulation tissue on the perimeter of the debridement. Neither survivor treated with hyperbaric oxygen required any further debridements before their wounds were closed.
Gozal et al. treated necrotising fasciitis patients with combined antibiotics, radical surgery and hyperbaric oxygen, and reduced the historic mortality rate from 38% to 12.5%. Of 29 patients reported retrospectively by Riseman et al, 12 were treated by surgical debridement and antibiotics only, and 17 received hyperbaric oxygen treatments in addition. Both groups had similar parameters of age, race, sex, wound bacteriology and antimicrobial therapy. Body surface area was also similar. However, perineal involvement (53% vs. 12%) and septic shock (29% vs. 8%) were more common in the hyperbaric group, yet the overall mortality was significantly lower at 23%, versus 66% in the non-hyperbaric oxygen treated group. Additionally, only 1.2 debridements per patient in the hyperbaric treatment group were performed, vs. 3.3 debridements per patient in the surgery plus antibiotics-only group.
Clearly a goal when making the diagnosis of necrotising fasciitis is to make it as early as possible so as to be able to start appropriate treatments and avoid rapid spreading and the onset of sepsis. Time is tissue. The main differential diagnoses includes standard cellulitis, which may be a precursor of necrotising fasciitis in some cases; and erysipelas, with its erythematous well-delineated border. Additional entities which should be considered include Clostridial myositis and myonecrosis; non-Clostridial myositis and myonecrosis; toxic shock syndrome, which may accompany necrotising fasciitis; Zygomycotic gangrenous cellulitis; mixed aerobic/anaerobic necrotising cellulitis; toxic epidermal necrolysis (TEN), also known as Lyell's Disease, usually due to exposure to particular medications; and Staphylococcal Scalded Skin Syndrome, also known as Ritter's Disease, due to exfoliative toxins produced by Staphylococci, with the latter two entities being most common in neonates and children under 5 years of age. In the neonate with omphalitis, violaceous discolouration of the skin appears to be a strong marker for the emergence of necrotising fasciitis. Vibrio vulnificans infections cause blistering infection quite commonly, and are seen in patients who have either been swimming in waters, along the Gulf of Mexico, or have been eating shellfish from that area. Aeromonas infections also occur following open wounds acquired in sea water. Cutaneous anthrax may present with a blackened central area and surrounding oedema.
Numerous studies have continued to demonstrate the beneficial effect of hyperbaric oxygen therapy in the management of necrotising fasciitis. Wilkinson and Doolette reported a 5-year retrospective cohort Australian study of 44 patients with necrotising soft tissue infection, between 1994 and 1999, looking at the primary outcome of survival to hospital discharge, and secondary outcomes of limb salvage and long-term survival after hospital discharge. Logistic regression analysis determined the strongest association with survival was the intervention of hyperbaric therapy (p=.02). Hyperbaric oxygen therapy increased survival with an odds ratio of 8.9 (95% confidence interval, 1.3-58.0) and a number of 3 needed to treat to benefit. Hyperbaric oxygen therapy also reduced the incidence of amputation (p=.05) and improved long-term outcome (p=.002). In the series by Escobar et al. there were no further amputations beyond those already done prior to transfer, once hyperbaric oxygen therapy was initiated in their series of 42 patients. The negative study by Brown at al. which purports to be a multi-centre retrospective review of treatment at 3 facilities over 12 years, of 54 patients, had numerous discrepancies in the demographics of their two groups. Half of the hyperbaric-oxygen-treated group of 30 patients, all from one institution, were noted to have Clostridial infections, while the non-hyperbaric treated group had only 4 of 24 patients (17%) with Clostridial infection. 6 of the 30 in the hyperbaric group are noted to have the diagnosis of Clostridial myositis and myonecrosis, whereas only one of the non-hyperbaric oxygen treated patients were so diagnosed. Hence this clearly shows the same diseases were not being compared in this study. Additionally, as is pointed out in a subsequent letter to the editor, 80% of the patients received 4 or fewer treatments, the remaining 20% received between 5 and 7 treatments, and the timing of these treatments is not specified. If the guideline of treating three times in the first 24 hours were followed, and then twice per day until the patient is stable and shows no relapse of toxicity between treatments, the gas gangrene patients in this study were treated for less than a day and a half, which is a shorter period of time than most other studies, and the others were treated for around 2 days. In the Wilkinson study, patients received a median of 8 treatments, which is more than that received by the patient with the greatest number of treatments in Brown et al. The authors state that the mortality difference between the two groups (9/30, or 30% of the hyperbaric group, versus 10/24, or 42% in the non-hyperbaric group) was not statistically significant. Thus the Brown et al study should not be used as an argument that the use of hyperbaric oxygen for truncal necrotising fasciitis is "controversial,” because these mortality statistics are not comparable, with a different mix of diagnoses in the two, compounded by the fact that the numbers themselves are small, resulting in a study that had insufficient power to demonstrate a statistically significant result. Furthermore the study does not add to the literature of necrotising fasciitis involving the limbs and other non-truncal sites.
Fortunately, Fournier's Gangrene cases in the literature are usually studied and reported as a distinct group. Hollabaugh et al. reported a retrospective series of 26 cases from the University of Tennessee's five hospitals. Of the 15 patients with identifiable sources for their infections, 8 had urethral disease or trauma, 5 had colorectal disease, and 2 had penile prostheses. All patients were managed with prompt surgical debridement and broad-spectrum antibiotics. Procedures performed included urinary diversion, faecal diversion, and multiple debridements. Fourteen of the twenty-six were additionally treated with hyperbaric oxygen. The group treated with hyperbaric oxygen had a mortality rate of 7%, versus 42% in the group not receiving hyperbaric oxygen (p=.04), with a combined overall mortality rate of 23%. The one patient who died while receiving hyperbaric oxygen therapy had been progressing well without evidence of ongoing infection, but suffered an acute MI not thought to be related to the underlying disease process. In the non-hyperbaric group, deaths were usually attributed to ongoing or fulminant sepsis. Relative risk for survival was 11 times greater in the group receiving hyperbaric oxygen therapy. This study did not show a decrease in the number of debridements by HBO2 therapy, but was confounded due to the larger number of patients who died and thus were not able to get further debridements. Delay to treatment was not a factor in the different groups.
Additional series include that of the Genoa, Italy group who treated 11 patients without any deaths, and all delayed corrective procedures healed without infectious complications. Another 33 patients were reported in a series from Turku, Finland. These patients were treated at 2.5 ATA, in conjunction with antibiotics and surgery. 3 patients died, for a mortality rate of 9%. Hyperbaric oxygenation was observed to reduce systemic toxicity, prevent extension of the necrotising process, and increased demarcation, improving overall outcomes. 2 of the 3 patients who died were moribund upon arrival to their facility. Management included diverting colostomies for those patients with a perirectal or perineal source, and orchiectomy, although sometimes reported in all series, is not routinely done since the blood supply to the testes is from the spermatic vessels which do not perfuse the scrotum and penis. Suprapubic cystostomy was indicated and performed when the source of the infection was genitourinary.
Due to the difficulty in making direct comparisons of clinical series', a Fournier's Gangrene Severity Index Score was developed in order to assess a number of variables rather than the presence of the disease itself. The score uses degrees of deviation from normal of physiologic variables to generate a score that correlates with patient mortality. It is clear that the amount of disease, related by some to body surface area of involvement, may be a significant variable. The Duke University analysis of 50 consecutive patients seen at their institution over a 15 year period had a 20% overall mortality. Three statistically significant predictors of outcome were identified when examined using univariate analysis: extent of infection, depth of the necrotising infection, and treatment with hyperbaric oxygen. However the same data using multivariate regression analysis identified only the extent of the infection as the only statistically significant independent predictor of outcome in the presence of other co-variables. Patients with disease involving a body surface area of 3.0% or less all survived. The numbers of patients with disease extent greater than 3%, where hyperbaric oxygen would thus be expected to play a role, became smaller, and with small numbers of patients, the power of the study to demonstrate a significant response was not present. Using multivariate analysis, the p value for statistical significance for hyperbaric oxygen treatments was equal to .06.
With such strong case series evidence of reductions in morbidity and mortality for necrotising fasciitis and the subset of Fournier's Gangrene, it is difficult to envision ever seeing a controlled, double-blinded study of hyperbaric oxygen therapy.