Risk factors, microbiological findings and outcomes of necrotizing fasciitis in New Zealand: a retrospective chart review

We reviewed 299 NF patient charts of which 247 (82.6%) satisfied the case inclusion criteria. Conditions like diabetic gangrene of toe, foot and leg, cellulitis, abscess, plantar fasciitis, Dupuytren’s contracture, and limb loss due to invasive meningococcal disease were sometimes coded as fasciitis or NF. However, these were clearly not cases of NF and were excluded from the analysis. Operative and histology findings criteria were the two most commonly used criteria for assigning disease status. However, for more than half of the charts a histology report was not available. Many operative notes did not mention specific operative findings but just mentioned ‘NF found at operation’. Consequently, a very tight case definition could not always be applied.

Table 1 shows the age, sex and ethnicity distribution of the 247 NF cases fulfilling the case definition. Males comprised 61% and females 39% of the total. Amongst Europeans the highest number of cases was in the 70+ age group, while for Maori and people of Pacific and Other ethnicity, the number of cases was highest in the 50–69 year age group. The median age of European cases (61.4 years) was significantly higher than Maori (48.9 years) and Pacific (50.8 years) (p < 0.001).

Table 2 shows the ethnic distribution of selected predisposing conditions. Diabetes was the most common (32.0%) predisposing condition, followed by obesity (22.7%). Nearly a quarter of the patients (23.1%) were reported to have taken NSAID within the seven days preceding development of NF. About a fifth of the patients did not have any predisposing condition recorded in their charts.

Some conditions were comparatively more prevalent in particular ethnic groups, e.g., diabetes in Others and Pacific people, alcohol abuse in Europeans, obesity in Maori and Pacific people, coronary heart disease in Others, congestive heart failure in Pacific people, chronic obstructive pulmonary disease (COPD) in Europeans, cancers in Others and Europeans, steroid use in Europeans, NSAID use in Others and chronic renal failure in Maori. However, the ethnic difference in prevalence was significant only for diabetes (p < 0.01) and steroid use (p < 0.05). P-values in this and subsequent tables should be interpreted with caution where some cells have values <5. There was also a statistically significant difference (p 0.004) in the ethnic distribution of cases with ‘no condition documented’ (Table 2). From the available clinical notes it is not possible to say the cause of this association. The significantly greater age of the European cases may have contributed to the higher documented prevalence of co-morbidities for NF cases from that ethnic group.

The distribution of precipitating events by ethnicity is shown in Table 3. By far the most commonly reported precipitating event was accidental trauma (33.2%), followed by skin ulcer (15.8%) and surgery (11.3%). The type of trauma ranged from superficial graze to blunt trauma without skin break, laceration and penetrating injuries. This group also included insect and animal bites. Five patients had a history of insect bite (one mosquito, one sand fly and three unspecified insect bites) as precipitating events. One person developed NF following a dog bite on their leg.

Only four patients (1.6% of the total) developed NF directly following traditional Samoan tattooing. All of them were Samoan and one of them died. Not surprisingly, the p value of this category indicates statistically significant ethnic differences in exposure to this precipitating event (Table 3). Two people (0.8%) were noted to have acupuncture in the affected area before developing NF.

Out of 247 NF patients’ charts reviewed, either debrided tissue, blood, aspirated fluid or wound swab culture results (in the form of hard copy or electronic laboratory reports) were available for 219. Another 19 patients had their culture results documented in their clinical notes but no laboratory reports were filed in their charts. Specimens were taken at operation (debrided tissue or swab) or preoperatively (blood, blister fluid or wound swab). In total, there were culture positive results from one or more specimens in 96% (228/238) patients. The extent of microbiology investigations and their documentation in different hospitals were not uniform. For example, the number of antibiotics against which a cultured organism was tested for susceptibility differed substantially from hospital to hospital and also within the hospital from case to case. We are not sure whether specimens were cultured for all possible organisms. For example, for many cases anaerobic culture reports were not available. Some reported specific anaerobic organisms isolated while some indicated that there was growth of mixed anaerobes. It is also important to note that a third of all patients received antibiotics before hospital admission and specimens for culture were taken. Table 4 is a list of organisms isolated either solely or in combinations, and their frequency of occurrence.

The most common organism isolated was Streptococcus pyogenes (41.2%) either solely (20.6%) or in combination with other organisms (20.6%). Streptococci other than Streptococcus pyogenes, usually in combination with other organisms, were isolated in 17.6% cases. For most of the Streptococcus pyogenes isolates antibiotic susceptibility testing was not done. The culture reports mentioned that Streptococcus pyogenes was universally susceptible to Penicillin.

The second most common organism isolated was Staphylococcus aureus (31.9%). Most (88.0%) Staphylococcus aureus isolates were resistant to penicillins. Susceptibility to erythromycin and flucloxacillin was 89.8% and 88.6% respectively. Methicillin-resistant Staphylococcus aureus (MRSA) was cultured in 8 (3.4%) cases.

The ethnic distribution of patients affected by these two organisms and the antibiotic susceptibility results of Staphylococcus aureus to three common antibiotics (penicillin, erythromycin and flucloxacillin) are presented in Table 5. Although a higher percentage of Pacific people cultured Streptococcus pyogenes and Staphylococcus aureus, these ethnic differences were not statistically significant.

Many NF patients developed several complications and underwent specific interventions during their hospital stay. Table 6 shows the frequency and ethnic distribution of some complications (renal failure, shock, multi-organ dysfunction, coagulation abnormalities and myocardial infarction), specific interventions (surgical debridement, skin grafting, blood product transfusion, ICU admission and dialysis) and two specific outcomes (amputation and death). The ethnic differences in the extent of complications, interventions and outcomes were not statistically significant.

This is the largest NF case-series so far reported and shows several important features of the risk factors, microbiology, and outcomes for this infection. Our analysis confirms the earlier finding of a positive association of NF with increasing age and with male sex [3]. More than 80% of patients had one or more recognized predisposing conditions. For Europeans and people of Other ethnicity, two common conditions were diabetes and NSAID use, while those for Maori and Pacific people were diabetes and obesity. Increasing rates of diabetes and obesity in Pacific people and Maori are well documented [28], and these trends could partly explain the increasing incidence of NF in recent years [3]. An association has been documented between NF and NSAIDs and there are plausible mechanisms by which NSAIDs might predispose to severe bacterial infection [14]. We found that nearly a quarter of patients had a history of NSAID use within the seven days preceding onset of symptoms. This finding is consistent with the results of two studies from Middlemore Hospital in Auckland [9, 12]. However, the definitive answer regarding the question of causation has not been resolved [14, 29]. As in other studies [16], we found that accidental trauma, surgery, ulcer, burn and intra venous drug use (IVDU) had triggered the process in some cases. A third of patients had an antecedent history of accidental trauma (including blunt trauma without a breach in the skin).

Traditional Samoan tattooing was documented as the precipitating event for NF in four Samoan men, one of whom died. There were no further details of the process of tattooing or infection control procedures recorded in the patient notes, so we cannot implicate any particular aspect of the procedure as contributing to the risk of developing the condition. It was reported in two charts that NF developed at the site of acupuncture. No further details were available. Although there was considerable media attention and publicity about the risks from traditional Samoan tattooing, this review found that the contribution from this exposure was fortunately small in relation to the overall burden of NF. This finding is consistent with the international literature which has reported only a few sporadic cases of NF following traditional tattooing [17, 30]. It is probable that some people developed less severe forms of infection (e.g. cellulitis) following traditional tattooing [20]. However, exploration of those outcomes was not part of this study.

We found that Streptococcus pyogenes, either singly or in combination with a staphylococcus or other organism, was the most frequent organism isolated in NF cases. Fifty-eight percent of culture positive cases in this group grew Streptococcus pyogenes or a non-Group A streptococcus. This finding is consistent with the microbiology findings of a case series of 198 necrotizing soft tissue infections (NSTI) in the US, where a streptococcus was isolated in 56% of cases [22]. Staphylococcus aureus including MRSA was cultured in more than a third of our reviewed cases for whom a microbiology result was available. This proportion is higher than that reported for the other series [9, 10, 12, 16, 22]. Only about 10% of staphylococci were MRSA (8 out of 76). A wide variation of the proportion of MRSA from NF patients has been reported [22, 31‐33]. Our result is near the lower end of this range [22, 33].

In addition to the above two major organisms, many others were cultured from reviewed NF patients. This result is consistent with the findings of the NSTI case series already mentioned, though the frequencies were different [22].

Proportions of Pacific people with Streptococcus pyogenes and Staphylococcus aureus positive culture are higher (though not significantly) than in all other ethnic groups. From this limited information, it appears that staphylococcus may be playing a bigger role in the pathogenesis of NF in NZ, particularly affecting Pacific people. This observation might partly explain the higher rates observed in Pacific people but needs further exploration to assess whether it is an important finding.

Our study documents the severe consequences of NF in terms of the high proportion with serious complications and death. We could locate only one other study reporting the extent of health care resource used to manage NF [23]. Although some studies reported the proportion of NF cases resulting in death, limb loss and ICU care, we could not locate any large study which reported complication rates and specific interventions to manage these as reported in this current paper.

Our study has important limitations. The NF patient records reviewed were confined to those admitted to a sample of NZ hospitals from 2000 to 2006, though it did include the majority of NF cases discharged over that seven year period. Consequently, it may not be possible to generalize our findings to all hospitals in NZ and to other time periods. This study was also limited to reviewing data on NF cases and did not include a comparison population. As a result, it is difficult to evaluate the significance of the risk factors which appeared to be associated with NF, such as diabetes and NSAID use. Our review of cases was limited to information recorded in patient notes, so is likely to underestimate the proportions of some characteristics, particularly predisposing conditions and precipitating events.