Is there a need for the Fournier's gangrene severity index? Comparison of scoring systems for outcome prediction in patients with Fournier's gangrene
Florian Roghmann, Department of Urology, Ruhr-University Bochum, Marienhospital, Widumer Strasse 8, D-44627 Herne, Germany. e-mail: firstname.lastname@example.org
Study Type – Prognosis (prospective cohort)
Level of Evidence 2a
What's known on the subject? and What does the study add?
Fournier's gangrene (FG) is a rare but life-threatening disease challenging the treating medical staff. Despite the fact that antibiotic therapy combined with surgery and intensive care surveillance are performed as standard treatment, mortality rates remain high. There have been efforts to develop a reliable tool to predict severity of the disease, not only to identify patients at highest risk of major complications or death but also to provide a target for medical teams and researchers aiming to improve outcome and to gather information for counselling patients. Laor et al. published the FG severity index (FGSI) in 1995 presenting a complex prediction score solely for patients with FG. Fifteen years later, Yilmazlar et al. suggested a new and supposedly more powerful scoring system, the Uludag FGSI (UFGSI), adding an age score and an extent of disease score to the FGSI.
In the present study population we applied two scoring systems for outcome prediction that are solitarily applicable in patients with FG (FGSI, UFGSI), as well as two general scoring systems such as the established age-adjusted Charlson Comorbidity Index (ACCI) and the recently introduced surgical Apgar Score (sAPGAR) to compare them and to test whether one system might be superior to the other. In addition, we identified potential prognostic factors in the study population. By contrast to many earlier studies, we performed a combined prospective and retrospective analysis and provided a 30-day follow up. In the cohort of the present study, older patients with comorbidities as well as a need for mechanical ventilation and blood transfusion are at higher risk of lethal outcome. All scores are useful to predict mortality. Despite including more variables, the UFGSI does not seem to be more powerful than the FGSI. In daily routine we suggest applying ACCI and sAPGAR, as they are more easily calculated, generally applicable and well validated.
- • To compare four published scoring systems for outcome prediction (Fournier's gangrene severity index [FGSI], Uludag FGSI [UFGSI], age-adjusted Charlson Comorbidity Index [ACCI] and surgical Apgar Score [sAPGAR]) and evaluate risk factors in patients with Fournier's gangrene (FG).
PATIENTS AND METHODS
- • In all, 44 patients were analysed. The scores were applied.
- • A Mann–Whitney U-test, Fisher's exact test, receiver operator characteristic (ROC) analysis and Pearson correlation analysis were performed.
- • The results of the present study show a significant association among FGSI (P= 0.002), UFGSI (P= 0.002), ACCI (P= 0.004), sAPGAR (P= 0.018) and death.
- • The differences between the area under the receiver operating characteristic curve of the scores were not significant.
- • Non-survivors were older (P= 0.046), had a greater incidence of acute renal failure (P < 0.001) and coagulopathy (P= 0.041), were treated more often with mechanical ventilation (P= 0.001) and received more packed red blood cells (RBCs; P= 0.001).
- • Older patients with comorbidities and need for mechanical ventilation and RBCs are at higher risk for death.
- • In the present cohort, scores calculated easily at the bedside, such as ACCI and sAPGAR, seemed to be as good at predicting outcome in patients with FG as FGSI and UFGSI.
age-adjusted Charlson Comorbidity Index
Fournier's gangrene severity index
red blood cells
receiver operating characteristic
surgical Apgar Score
total body surface area
Fournier's gangrene (FG) is a fulminant necrotizing fasciitis of the genital, perineal and perianal regions. Although first described by Baurienne in 1764, Jean Alfred Fournier gave the disease its eponymous name in 1883 [1,2]. Fournier defined the disease as an idiopathic, fulminant genital gangrene of acute onset in otherwise healthy young men . Since then, its epidemiology and clinical features have changed substantially as patient age has increased and comorbidities have become more frequent. Nevertheless, it remains a rare but life-threatening disease challenging the treating medical staff .
Despite the fact that antibiotic therapy combined with aggressive surgical debridement and intensive care surveillance have been broadly accepted as the standard treatment, mortality rates remain high. There have been efforts to develop a reliable tool to predict the severity of the disease. A scoring system might allow the treating medical staff to consistently identify patients at highest risk of major complications or death. It could also be useful in counselling patients and relatives for prognosis. Such a score could provide a target for medical teams and researchers aiming to improve outcome, and moreover a measure for quality monitoring and improvement programmes . An ideal score should be simple to use and effective at providing a clear, graded feedback on the status of the patient . In addition it might allow researchers to compare the morbidity of different study cohorts.
One good example for such an index is the 10-point scoring system for evaluation of the newborn infant introduced by Virginia Apgar in 1953 [4,5]. Being directly calculated in the delivery room, it quickly became an essential tool in achieving remarkable safety in modern obstetrics, as it proved to be predictive of 28-day survival [6,7].
In 1981 the Acute Physiology and Chronic Health Evaluation (APACHE) score and in 1991 the Physiological and Operative Severity Score for the Enumeration of Mortality and Morbidity (POSSUM) score were published and proposed as clinical measures of the patient condition. Both scores have been broadly validated but rely on various data elements and require numerous laboratory results which are not uniformly collected [4,8,9]. Laor et al.  published the FG severity index (FGSI) in 1995, also presenting a complex score for outcome prediction solely for patients with FG. Fifteen years later, Yilmazlar et al.  suggested a new and supposedly more powerful scoring system, the Uludag FGSI (UFGSI), adding an age score and an extent of disease score to the FGSI. All these scores have in common the fact that they are not easily calculated at the bedside and none of them has come into standard use in the emergency room even though the FGSI has become a standard for researchers, being routinely published in FG literature.
The additional time and effort required to apply the FGSI and UFGSI in the daily routine would only be justified if they are more powerful at predicting outcome than other established general scoring systems such as the age-adjusted Charlson Comorbidity Index (ACCI)  and the recently introduced surgical Apgar Score (sAPGAR) ,which are easily calculated at the bedside.
We applied the FGSI, UFGSI, ACCI and sAPGAR to the present study population to compare them and to test whether one system might be superior to another in predicting outcome. In addition, we explored whether other potential prognostic factors could be identified in the present cohort.
PATIENTS AND METHODS
The study was approved by the institutional review board. We identified 44 patients diagnosed with FG and treated at our institution between January 2001 and June 2011. A combined prospective (n= 10) and retrospective (n= 34) review of the medical records was performed. The diagnosis of FG was based upon physical examination at admission. In 48% of the patients, a histopathological examination was performed supplementarily, confirming the diagnosis in 100%; in two cases diagnosis was additionally supported by radiological imaging. Patients without any signs of necrosis or soft tissue extension as well as patients showing only a solitary perianal, scrotal or peri-urethral abscess were excluded. The following criteria were analysed in detail: physical examination, admission signs and symptoms, medical history, comorbidities, demographics, skin involvement, laboratory results, antibiotic treatment, extent and number of surgical interventions, and intensive care treatment.
The extent of skin involvement (total body surface area [TBSA]) was calculated using a nomogram that is routinely deployed to assess the extent of burn injuries and which has been modified for use with FG patients by Palmer et al. . According to the nomogram, the surface of the penis, perineum and scrotum each accounts for 1% of the body surface area, whereas each ischiorectal fossa accounts for 2.5%.
Within 24 h of hospital admission, all patients underwent surgical debridement. All necrotic tissue was resected until viable margins were identified. If necessary, debridement was repeated every 24–48 h. All patients immediately received calculated broad-spectrum parenteral antibiotic treatment until a specific therapy based on a resistogram was available. In addition, all patients received a supportive fluid resuscitation pre- and perioperatively. After surgery, the wound was closely monitored. Wound dressings (wet, dry or negative pressure) were frequently changed.
To calculate Laor's FGSI the following variables at admission were evaluated: temperature, heart rate, respiratory rate, serum sodium, serum potassium, serum creatinine and serum bicarbonate concentrations, haematocrit and leucocyte counts. A deviation from normal values was graded from 0 to 4 (Table 1) . According to Yilmazlar et al. , age and dissemination scores were added to the FGSI, resulting in the UFGSI.
Table 1. The FGSI* and the UFGSI**
|Heart rate, beats/min||>180||140–179||110–139||–||70–109||–||55–69||40–54||<39|
|Respiratory rate, breaths/min||>50||35–49||–||25–34||12–24||10–11||6–9||–||<5|
|Serum sodium, mmol/L||>180||160–179||155–159||150–154||130–149||–||120–129||111–119||<110|
|Serum potassium, mmol/L||>7||6–6.9||–||5.5–5.9||3.5–5.4||3–3.4||2.5–2.9||–||<2.5|
|Serum creatinine, mg/100 mL (×2 for acute renal failure)||>3.5||2–3.4||1.5–1.9||–||0.6–1.4||–||<0.6||–||–|
|White blood cell count, total/mm3× 1000||>40||–||20–39.9||15–19.9||3–14.9||–||1–2.9||–||<1|
|Serum bicarbonate (venous), mmol/L||>52||41–51.9||–||32–40.9||22–31.9||–||18–21.9||15–17.9||<15|
The CCI was calculated by forming a sum of 19 medical conditions weighted 1–6, with total scores ranging from 0 to 37. Age has also been found to be an independent risk factor for death from a comorbid condition. For each one-point increase, another point can be added to the CCI score for each decade of life over the age of 50, resulting in the ACCI . We used the MS Excel CCI Macro for rapidly calculating the score .
The sAPGAR score is based on three intraoperative variables – lowest heart rate, estimated blood loss and lowest mean arterial pressure – ranging from 0 to 10. Zero points corresponded to the highest risk for death (Table 2) .
Table 2. The sAPGAR 
|Estimated blood loss, mL||>1000||601–1000||101–600||≤100||–|
|Lowest mean arterial pressure, mmHg||<40||40–54||55–69||≥70||–|
|Lowest heart rate, beats/min||>85||76–85||66–75||56–65||≤55*|
Mortality was defined as disease-related death during the first 30 days after diagnosis. Differences in clinical variables between survivors and non-survivors were analysed using commercial statistical software (MEDCALC Software, Belgium, 2010; and SPSS Statistics, version 19.0, IBM, Chicago, IL, USA). The applied tests were as follows: Mann–Whitney U-test, Fisher's exact test, receiver operator characteristic (ROC) analysis and Pearson correlation analysis. P values are reported as the result of two-tailed testing. P < 0.05 was considered to indicate statistical significance.
In all, 44 patients with a median (interquartile range [IQR]) age of 59 (48–65) years were included. Despite all efforts, 13/44 patients (30%) died. The surviving patients were significantly younger than those who died (P= 0.046; Table 3).
Table 3. Comparison of baseline characteristics describing survivors and non-survivors. Values are medians (IQRs)
|Age, years||59 (48–65)||62 (52–71)||52 (43–64)||0.046†|
|Amount of packed RBCs||0 (0–3)||2 (2–9)||0 (0–2)||0.001†|
|Serum creatinine, mg/100 mL||1 (0.8–1.8)||1.5 (0.9–3.1)||1 (0.7–1.6)||0.049†|
|Hematocrit, %||37 (32–42)||35 (30–37)||40 (33–43)||0.014†|
|Lowest mean arterial pressure during surgery, mmHg||67 (62–85)||60 (55–70)||73 (65–85)||0.010†|
|Body mass index||28 (25–32)||26.3 (9.6–31.5)||27 (25–32)||0.685|
|Total body surface area involved TBSA, %||3 (2–6)||5 (2–6)||3 (2–5)||0.549|
Regression analysis revealed a strong correlation between FGSI, ACCI, sAPGAR and UFGSI and death. All scores significantly stratified survivors from non-survivors (Table 4). The recommended threshold values are listed in Table 5. There was no significant difference between area under the ROC curve for FGSI and those for UFGSI, ACCI and sAPGAR (Table 6). The ROC analysis is shown in Figs 1–4.
Table 4. Comparison of prediction scores in survivors and non-survivors. Values are medians (IQRs)
|FGSI||3 (1–8)||6 (4–11)||2 (1–4)||0.002*|
|UFGSI||4 (3–10)||7 (6–12)||3 (2–7)||0.002*|
|ACCI||3 (0–7)||4 (4–7)||2 (0–4)||0.004*|
|sAPGAR||6 (5–7)||4 (4–6)||6 (6–8)||0.018*|
Table 5. Threshold values for the applied scores
Table 6. Comparison between the ROC curve of the FGSI with those of the UFGSI, ACCI and sAPGAR
There was no significant difference in TBSA involvement between survivors and non-survivors (Table 3). Moreover, there was no significant association between extent of TBSA involvement and time elapsed until death. Comparing survivors and non-survivors, diverting colostomy (16% of all patients, n= 7) and suprapubic cystostomy (80% of all patients, n= 35) were not associated with mortality. In the present study, the need for mechanical ventilation was associated with a higher mortality rate (P= 0.001). The amount of packed red blood cells (RBCs) given was also accompanied by a lethal outcome (Table 3). A total of 25% of the wounds (n= 11) were initially treated with negative pressure dressings (vacuum wound closure) as opposed wet and dry dressings. However, there was no correlation between the utilization of vacuum wound closure and mortality.
In all, 75% of the patients had at least one of the following comorbidities: renal failure, coagulopathy, peripheral artery occlusive disease, diabetes, alcoholism, smoking and cardiac insufficiency (Table 7). Acute renal failure and coagulopathy rates as well as serum creatinine were significantly lower among survivors than among non-survivors (Table 3). By contrast, hematocrit and lowest mean arterial pressure were significantly lower among non-survivors than among survivors (Table 3). There were no significant differences between survivors and non-survivors in terms of antibiotic and surgical therapy, body mass index (Table 3), heart rate and body temperature at admission. Fourteen patients (32%) were transferred to our institution from other hospitals or departments. The outcomes for these patients were not statistically different from those of patients initially treated in our department. During the initial operative debridement, aerobic and anaerobic cultures were obtained in all patients. The most frequent bacterial organisms detected were Streptococcus spp., Bacteroides spp., Escherichia coli, Staphylococcus spp., Enterococcus spp. and Klebsiella pneumonia. In seven of 44 cases (16%), wound swabs were negative upon microbiological evaluation. Positive cultures were polymicrobial in 21/37 (57%). We did not find a significant correlation between bacterial species and patient outcome.
Table 7. Difference in comorbidities between non-survivors and survivors (75% of the patients had at least one of the comorbidities in the table)
|Acute renal failure||27||67||4||<0.001†|
|Peripheral artery occlusive disease||14||11||18||1.000|
Fournier's gangrene is a rare but life-threatening fulminant disease. Owing to small study cohorts, the published literature is limited, especially with regard to the validation of prediction tools for clinical outcome. In the present study, we report a combined prospective and retrospective analysis of 44 patients treated at our institution during the last decade. We applied the prediction indices developed by Laor et al. (FGSI), Yilmazlar et al. (UFGSI), Charlson et al. (ACCI) and Gawande et al. (sAPGAR) to compare these individual scores [4,10–12].
In all, 30% of the patients in the present study died from the disease, corresponding well with the published literature (7.5–45%) [11,15–18]. In the present patient cohort, application of all scores revealed a significant difference between non-survivors and survivors (Table 4). Therefore, these findings support other retrospective series that appraise the FGSI as a reliable tool to predict outcome (30 patients ; 25 patients ; 19 patients ; 69 patients ; 18 patients ). Only three studies were performed to investigate an association between FGSI and mortality in a prospective setting, but these had contradictory results (20 patients ; 80 patients ; 51 patients ). In 2010, Erol et al.  described an association between the Charlson Comorbidity Index and death in a small cohort. The data of the present study support these findings. To our knowledge this is the first study to evaluate outcome prediction with the sAPGAR in patients with FG.
The data of the present study do not support a more precise outcome prediction using the UFGSI, ACCI and sAPGAR than is obtained with the FGSI. The differences between areas under the ROC curve were not significant (Table 6). Interestingly, the ACCI showed the largest AUC (Table 6, Fig. 3). However, the results of the present study show that all scores were able to identify high-risk patients in the present cohort. Hence, it might be considered appropriate to calculate the ACCI in the emergency room or the sAPGAR directly after the initial operation, as they are not just useful for a small group of patients with a rare disease. Simplifying the use of scoring systems might improve their acceptance and prospective application in a clinical setting at the bedside [4,14].
Compared with the initial description of FG in 1883 by Fournier, patient age has increased over the last century [10,15,16,18,22,23,27]. With a median age of 59 years, the data of the present study confirmed this finding. In the present study, patients who survived were significantly younger than non-survivors (Table 3). This result is consistent with several other reports [10,11,24]. According to these findings, Yilmazlar et al.  developed an age score as one of three modules for calculating the UFGSI (Table 1). In our cohort, however, there was no additional clinical value in adding the age score to the regular FGSI.
Comorbidity is frequent in patients who develop FG today and a comorbidity score such as the ACCI has shown utility at outcome prediction (Table 4). Thus the data of the present study contradict Fournier's initial definition . A total of 75% of our patients had at least one of the following conditions: diabetes, renal failure, cardiac insufficiency, coagulopathy (any disturbance of coagulation including those due to iatrogenic intervention), peripheral artery occlusive disease, alcoholism and smoking. Prior studies have suggested that comorbidities such as diabetes might predict poor outcome . In the present series, there was no significant difference in mortality between diabetic and non-diabetic patients, but mortality was associated with renal failure (Table 7), which is one of the FGSI variables. Moreover, there was a significant correlation between coagulopathy and mortality (Table 7), suggesting that sufficient perfusion is important for disease control.
There is no consensus as to whether the extent of TBSA involved in the necrotizing process is associated with outcome. Spirnak et al.  reported a higher mortality rate in patients with a greater extent of TBSA involvement. These patients also had to undergo surgical treatment at higher frequencies. Spirnak et al.'s findings were consistent with reports by Tuncel et al.  and Yeniyol et al. , who reported that the TBSA involved in FG was one of the most important risk factors. Yilmazlar et al.  also showed a significant association between the extent of FG and death. As a consequence, Yilmazlar et al. included a dissemination score in the UFGSI (Table 1). However, this issue remains controversial: consistent with the results of other series, in the present study we failed to show an association between TBSA and outcome (Table 3) [10,17,24].
Both aerobic and anaerobic bacteria seem to be causally involved in the development of FG. Bacterial species cultured from our patients are concordant with most studies, which reported common isolates such as E. coli, Bacteroides spp., Enterococcus spp., Streptococcus spp. and Staphylococcus spp [17,26,27]. Results from our series revealed predominantly polymicrobial (57%) infections. Based upon this finding, we could not identify one predominant species causing FG. Other studies suggest that a primary colorectal source [27,28] and colostomy  were risk factors for increased mortality. We could not confirm this association in our dataset. We believe that FG is a multifactorial disease caused by different bacterial species primarily originating from the genitourinary or colorectal regions, and is more likely in older patients with comorbidities.
The present study is limited by the number of patients and the partially retrospective approach. We are aware of the fact that a multi-institutional, prospective study is needed to acquire the number of cases necessary to further validate the findings of the present report in an adequately powered analysis. However, in contrast to many other studies, we report a mortality rate based on a follow-up. In addition, we provide histopathological proof of diagnosis in 48% of the study cohort.
In summary, older patients with comorbidities are at higher risk of experiencing serious complications and lethal outcome. FGSI, UFGSI, ACCI and sAPGAR are useful tools for predicting mortality. Despite including more variables, the UFGSI does not seem to be more powerful than the FGSI. We suggest applying the ACCI and sAPGAR, as they are more easily calculated, generally applicable and well validated.
CONFLICT OF INTEREST