Pattern and predictors of mortality in necrotizing fasciitis patients in a single tertiary hospital

Background Necrotizing fasciitis (NF) is a fatal aggressive infectious disease. We aimed to assess the major contributing factors of mortality in NF patients. Methods A retrospective study was conducted at a single surgical intensive care unit between 2000 and 2013. Patients were categorized into 2 groups based on their in-hospital outcome (survivors versus non-survivors). Results During a14-year period, 331 NF patients were admitted with a mean age of 50.8 ± 15.4 years and 74 % of them were males Non-survivors (26 %) were 14.5 years older (p = 0.001) and had lower frequency of pain (p = 0.01) and fever (p = 0.001) than survivors (74 %) at hospital presentation. Diabetes mellitus, hypertension, and coronary artery disease were more prevalent among non-survivors (p = 0.001). The 2 groups were comparable for the site of infection; except for sacral region that was more involved in non-survivors (p = 0.005). On admission, non-survivors had lower hemoglobin levels (p = 0.001), platelet count (p = 0.02), blood glucose levels (p = 0.07) and had higher serum creatinine (p = 0.001). Non-survivors had greater median LRINEC (Laboratory Risk Indicator for NECrotizing fasciitis score) and Sequential Organ Failure Assessment (SOFA) scores (p = 0.001). Polybacterial and monobacterial gram negative infections were more evident in non-survivors group. Monobacterial pseudomonas (p = 0.01) and proteus infections (p = 0.005) were reported more among non-survivors. The overall mortality was 26 % and the major causes of death were bacteremia, septic shock and multiorgan failure. Multivariate analysis showed that age and SOFA score were independent predictors of mortality in the entire study population. Conclusion The mortality rate is quite high as one quarter of NF patients died during hospitalization. The present study highlights the clinical and laboratory characteristics and predictors of mortality in NF patients.


Background
Necrotizing fasciitis (NF) is a rare infectious disease which is rapidly progressive and potentially fatal in nature [1]. Despite the advanced medical treatment, the rate of mortality remains as high as 24-34 %; posing a challenge for the diagnosis and management [2]. The mortality in NF patients primarily depends upon the time of the medical and surgical interventions and extent of spread of infection to the primary site (subcutaneous tissue, fasciae, skin or muscles) [3]. NF patients with streptococcal infection are associated with increased risk of complications and mortality (up to 80 %) [4]. There are various predisposing factors for NF such as advanced age, diabetes mellitus, peripheral vascular disease, obesity, chronic renal failure and trauma [5]. Therefore, early recognition of these predisposing factors might help in the early definitive management [6]. Moreover, early surgical debridement is a known contributor to improve outcomes in NF patients [5].
To date, the most appropriate tool for diagnosis and discrimination of NF is the LRINEC (Laboratory Risk Indicator for NECrotizing fasciitis score) scoring system proposed by Wong et al. [7]. It is based on laboratory parameters that are readily available for scoring in most institutions to predict survival and discriminate necrotizing from non-necrotizing infections. In addition, various predictors of mortality based on predisposing factors have been reported by different investigators. Previous studies have identified advanced age (>60 years), aeromonas and vibrio infection, liver cirrhosis, cancer, hypotension, band polymorphonuclear neutrophils (PMN) >10 %, and serum creatinine >2 mg/dL to be independent predictors of mortality in NF cases [8,9]. Despite the fact that mortality depends upon the time of diagnosis, and management, the predisposing factors also play an important role in the outcome. The present study aims to determine the various predisposing and prognostic factors associated with mortality in NF patients.

Methods
All consecutive patients admitted to the surgical intensive care at Hamad General Hospital (HGH) with a diagnosis of necrotizing fasciitis were retrospectively included in this analysis between January 2000 and December 2013. Patients were categorized into 2 groups according to their hospital mortality (survivors versus non-survivors). The present study included patients for which the operative notes and or histopathological findings indicate NF. Data included age, sex, presentation and duration of symptoms, predisposing factors, risk factors, causative microbiological organisms, on-admission laboratory parameters, the Sequential Organ Failure Assessment (SOFA) and LRINEC score, number of operative debridement, length of intensive care and hospital stay, recurrence and in-hospital mortality. The anatomic site of infection was classified as extremity (upper and lower limbs), abdominal/groin, chest/breast, neck/facial, sacral and perineum. Exclusion criteria included patients not admitted in the ICU (uncomplicated mild forms of NF and were managed in the regular ward), patients with cellulitis or superficial infection not requiring aggressive debridement or antibiotics), and non-surgical patients.
The ratio of partial pressure of arterial oxygen and fraction of inspired oxygen (PaO2/FiO2), platelets count, bilirubin level, Glasgow coma score, mean arterial pressure (MAP), vasopressor use, creatinine level and urine output were used to calculate SOFA score [10]. The laboratory parameters such as C-reactive protein (CRP), WBC, hemoglobin, sodium level, creatinine concentration and glucose level were used to calculate the LRI-NEC score [7]. The various recorded parameters were analyzed according to the final outcome. This study was approved by the medical research center at HMC, Qatar with IRB#14066/14.

Statistical analysis
Data were presented as proportions, median (range) or mean (± standard deviation), as appropriate. Baseline demographic characteristics, laboratory findings, clinical presentation, bacteriology and predisposing factors were compared between non-survivors and survivors. Analyses were conducted using Student-t test for continuous variables and Pearson chi-square (χ 2 ) test for categorical variables; Fisher's exact test was used, if the expected cell frequencies were below 5. A 2-tailed p < 0.05 was considered significant. Multivariate logistic regression analysis was performed to look for the predictors of mortality in the overall NF cohort along with the odd ratio and 95 % confidence interval. Data analysis was carried out using the Statistical Package for Social Sciences version 18 (SPSS Inc, Chicago, Illinois).

Management and outcomes
The median number of surgical debridements performed was 2 (ranged 1-8) and the hospital length of stay was 16 (2-295) days. The number of debridements was comparable in the 2 groups. The median ICU stay [9 (1-75) vs. 5 days (1-43); p = 0.002], overall hospital stay [20.5 (2-273) vs. 15 (2-295) days; p = 0.02] and the frequency of septic shock (48 % vs. 20 %; p = 0.001) were significantly higher in non-survivors than the survivor group. Recurrent admissions for NF were required for 13 (4 %) patients; of whom 11 patients were admitted twice and two patients required three admissions. A total of 85 patients died in the present study with an overall mortality rate of 26 %. Table 3 shows the major causes of mortality which mainly involved septic shock alone and a combination of bacteremia and multiorgan failure. Table 4 shows multivariate analysis for the major predictors of mortality. SOFA scoring followed by age were the independent predictors of mortality in the present study cohort. The proportion of mortality based on the bacteriology results is given in Table 5.
None of the co-morbidities showed significant association with types of microorganisms and combination of antibiotics used except coronary heart disease (CHD). Significantly higher frequency of CHD patients were prescribed more than two antibiotic combinations (p = 0.009). Table 6 compares the co-morbidities with microbiological data and antibiotics used.

Discussion
The association of high morbidity and mortality in necrotizing fasciitis (NF) patients urges the need for early diagnosis and identification of potential risk factors of worse outcomes. The present study is interestingly large series of NF cases from a Middle Eastern small population country that assesses various contributing factors to mortality. NF is a fulminant life-threatening infection of the musculoskeletal soft tissues characterized with rapid progression that typically requires urgent surgical interventions [11,12]. The classic and frequent manifestations associated with NF usually include a triad of pain, tender local swelling, and fever [13,14]. Consistent with earlier reports, this triad was more frequently observed among survivors in the current series. Moreover, out of proportion pain on physical examination and unresolved cellulitis are major diagnostic clues for NF, however, these clinical features often appear later in the disease course [9]. Therefore, delayed diagnosis is usually associated with high mortality (up to 25 %) among young adults which could even reach 44 % in elderly population [9,15]. The number of in-hospital deaths in the present study is 26 % which is consistent with earlier studies. The current literature suggests that NF could occur at any age but is mostly reported within the age range of 32 to 57 years [16,17]. In the present study, the mean age was 51 years and the non-survivors were 14.5 years older than the survivors at the time of presentation. The reason of the frequently observed association of NF with advanced age could be explained in part by the preexisting co-morbidities and immunosuppression. In this context, Golger et al. [18] reported advanced age, streptococcal toxic shock syndrome and immunocompromised status to be independent predictors of mortality in NF patients.
The frequently associated co-morbidities in NF are diabetes mellitus, malignancy, chronic cardiac disease, peripheral vascular disease, chronic renal disease, and immune-suppression [19]. Other predisposing factors include traumatic injuries, smoking, history of muscular injection and paraplegia. Diabetes mellitus, hypertension, and renal impairment were the most frequent comorbidities associated with mortality in the current series. Diabetes mellitus remains the main co-morbidity in NF patients which is associated with prolonged hospitalization and increased mortality [13,20]. In this study, patients with a history of diabetes mellitus showed considerably rapid progress of the severity of NF and mortality. This finding could be attributed in part to the hyperglycemic status that compromises the immunity status and fosters bacterial growth. However, the initial readings of serum sugar in the study cohort were nonsignificantly lower in non-survivors. Unfortunately, the current database did not include HbA1c to explain inpart this finding. The other common comorbidity in the   present study was hypertension, which might cause disruption of the microvascular supply and reduction of tissue oxygenation and antimicrobial delivery. The frequency of hypertension was significantly higher in the non-survivors group. Consistently, Huang et al. [8] demonstrated a high association of hypertension among NF non-survivors. Earlier studies have also outlined the increased risk of NF in the presence of the abovementioned pathologies [21,22]. Furthermore, elderly patients with such co-morbidities who are suspected to have NF should be evaluated thoroughly to rule out NF, even in the absence of the usual hard manifestations. Although, NF might affect any part of the body, earlier studies have reported frequent involvement of the extremities, perineum, head & neck and truncal regions [23]. In the current series, the most frequent sites of infection included lower limbs, perineum, abdominal/ groin and neck/facial regions. The site of infection and its expansion also affect mortality. It has been suggested that affection of the head and neck region is associated with higher mortality as accounted for the proximity with various vital anatomical structures [24]. Mao et al. [25] analyzed the craniocervical NF cases with and without thoracic extension and observed a poor survival with thoracic extension as compared to non-thoracic extension. An earlier study reported a lower rate of mortality in extremity infection in comparison to abdominal and perineal infections [26]. Urschel [27] suggested that NF infection extending proximally to pelvis or trunk might have worse prognosis. Therefore, early and aggressive treatment aimed at restriction of the infection with repeated surgical debridement could be useful in achieving better survival rates.
Unfortunately, the first stage of NF disease is frequently masked by non-specific manifestations, which prevents effective and timely specific therapy [28]. Therefore, early identification and diagnosis is mandatory and should not rely only on the clinical signs alone [6]. Consequently, prognostic indicators such as laboratory markers and specific patient characteristics obtained from the medical history would assist in the early diagnosis, risk stratification and decision making [29]. Earlier studies identified some laboratory findings such as anemia, elevated creatinine, and increased white cell count to be non-specifically associated with NF which might affect prognosis. In the present series, non-survivors had significantly lower levels of hemoglobin and platelet count and had higher serum creatinine as compared to survivors. Similarly, an earlier study observed that non-survivors had significantly lower levels of hemoglobin and platelet and presented with higher levels of serum glucose and creatinine than the survivors [8]. An earlier study reported that aeromonas infection, advanced age, band PMNs >10 %, serum creatinine (>2.0 mg/dL), and an activated prothrombin time (>60 s) were found to be the independent predictors of mortality in NF patients [8]. In the present study the major predictors of mortality were age and SOFA scoring. SOFA score is a useful tool to assess the severity of NF based on the involvement of major organ systems. In the present series, non-survivors had significantly higher median SOFA scores which are in accordance with the current literature. The initial increase in SOFA score during the first two days of ICU admission successfully predicts high rates of mortality (50-95 %) [30]. This finding could be used as an alarming indictor and encourages physicians to refer those patients as early as possible to the tertiary care centers for the appropriate intensive care. Therefore, the use of validated prognostic factors in daily clinical practice, especially for initial diagnosis in emergency departments, would help physicians for timely management and obtaining better outcomes.
It has been suggested that bacteremia is one of the frequent complications of NF which has been associated with higher risk of mortality [31]. In the current study, 21 % of the patients had positive blood and tissue cultures and subsequently had higher mortality rate in comparison to those who had negative blood culture. Similarly, Huang et al. [8] observed four-fold increased rate of mortality in patients with positive blood cultures than those who had negative cultures.
Consistent with previous reports [13], gram positive microorganisms, mainly streptococcus and staphylococcus organism, were frequently identified in the present study cohort. On the other side, bacteriodes and E-coli were the predominant gram negative organisms. In the present study, monobacterial infections with pseudomonas and proteus were the most commonly associated microorganisms among non-survivors. However, earlier  studies reported clostridial [32], beta-streptococci [33], aeromonas and vibrio [9] infections to be associated with poor outcomes. Prompt and aggressive debridement is important for the management of NF. The debridement aims to remove all necrotic tissue until the local infectious process is treated. There is a positive correlation between the survival rate and early diagnosis with appropriate surgical debridement in NF patients [34]. In the present study, the median number of debridement procedures performed per patient was two interventions, and these were comparable for both non-survivors as well as survivors. Data suggested that early surgical intervention is crucial in reducing morbidity and mortality in NF patients [35]. However, there is still a lack of clear definition on 'How early should we be'. Kobayashi et al. showed significantly lower mortality in the early intervention group (within 12 h after diagnosis) [36]. Delay of surgical treatment of >12 h was associated with an increased number of surgical debridement, septic shock and acute renal failure [36]. Hadeed et al. [35] reported outcomes of earlier surgical treatment (within the first 6 h) and found that although there was no statistically significantly difference in mortality between the study groups, higher mortality among late intervention group was clinically significant. Moreover, the outcomes in terms of the duration of hospital and intensive care unit stay were in favor of early intervention [35].
The appropriate and early antibiotic use and intensive care measures significantly appear to affect patients' outcomes. In the past decades, patients with NF have higher mortality rates (up to 70 %), however, currently with improved surgical and intensive care treatment, mortality rates have declined to less than 30 % [37,38]. Not only delayed diagnosis and surgical intervention influences in-hospital mortality, but also, the development of secondary complications has unfavorable impact [39]. The major complications that significantly related to mortality in the present series were bacteremia, septic shock and multiorgan failure. Therefore, appropriate prevention and management of such complications are vital for improving the outcome in these vulnerable patients [39]. Table 7 summarizes the published studies of mortality in NF and NSTI worldwide between 2000 and 2016.
There is no consensus for specific predictors of mortality between these 20 studies including the current study. The design and objectives of studies as well as the availability of clinical and laboratory data are the main reason of the diversity of predictors of mortality among these studies. The present study has several limitations. It is retrospective in nature. It lacks information regarding the exact time of commencing antibiotics, delay in diagnosis, the timing of surgical debridement and the type of surgery performed post diagnosis. Also, the sensitivity and minimum inhibitory concentration of the bacteria and percentage of multi drug resistant strains is not available . Data describing the empiric antibiotic treatment in the emergency room is also not available for analysis. Moreover, procalcitonin has been introduced recently at HGH; therefore not all the NF in the past underwent procalcitonin assessment at admission. Further prospective studies are required to determine the time interval between the diagnosis and treatment which could possibly influence the mortality among NF patients.

Conclusion
The mortality rate is quite high as one quarter of NF patients died in the hospital. The present study highlights the clinical characteristics and predictors of mortality in NF patients. It is important to have a high index of suspicion at initial presentation. Use of prognostic tools in the daily clinical practice will help physicians for the proper on-time management. The present study provides useful information on the severity and outcome of NF patients that will inform institutional guidelines for the on-time treatment of NF.