Prognostic factors and treatment outcomes for patients with Fournier's gangrene: a retrospective study

Abstract

Fournier's gangrene is a gas‐forming, necrotising soft tissue infection affecting the perineum. It spreads rapidly along the deep fascial planes and is associated with a high mortality rate. With a growing elderly population with comorbidities, the frequency of severe cases of Fournier's gangrene is expected to increase. We retrospectively reviewed 20 patients diagnosed with Fournier's gangrene at our institution from 2003 to 2014 and analysed data. Thirteen patients had diabetes mellitus, two had been diagnosed with liver cirrhosis, and four were chronic alcoholics. Of 15 patients admitted to an intensive care unit, 11 underwent colostomy, and 4 required skin grafts for wound healing. The wide wounds of two patients were healed using vacuum‐assisted closure (VAC ^®) dressing without additional surgery. The mortality rate was 25%, and the patients whose Fournier's gangrene severity index (FGSI) score was higher than 9 points or whose blood urea nitrogen (BUN) level was higher than 50 mg/dl had a poor prognosis. In order to treat Fournier's gangrene, aggressive surgical treatment, including wide debridement and stoma creation, should be considered as soon as possible to improve survival rates. Additionally, VAC dressing is helpful in healing the wide debridement wound without additional reconstructive surgery.

Introduction

Fourier's gangrene is a polymicrobial necrotising fasciitis that results in endarteritis of the perineal, urogenital or perianal subcutaneous layers and causes gangrenous changes in the skin and subcutaneous layers (Fig. 1) 1, 2. When first described by Jean Alfred Fournier in 1883, it was considered a disease of young men, with rapid progression and without definite cause 3. However, many studies have since revealed that there are almost always predisposing causes that induce necrotising infection and that this disease also affects women, infants and the elderly 4. Early detection and diagnosis and active treatment, including aggressive debridement and drainage and administration of broad‐spectrum antibiotics, are key in disease management.

Figure 1.

A case of a patient with Fournier's gangrene. (A) Fournier's gangrene. Necrotic fasciitis spread to the perianal area and scrotum. (B) After debridement. Spontaneous orchiectomy was performed. (C) Reconstruction with skin graft. Skin defect was repaired with a skin graft from the left thigh.

As Fournier's gangrene is uncommon and doctors' experience of it is limited, it is difficult to diagnose this disease before necrosis or gangrene sets in, and disease progression, even to mortality, is rapid 5. However, late detection and inappropriate treatment lead to high mortality; thus, the initial treatment approach, based on an accurate assessment of the disease, is very important.

The mortality rate of this disease, at 0–67%, varies widely between studies 2 and has not improved over time in spite of medical advances. In particular, in developed countries with aging populations and high levels of morbidity due to metabolic diseases, such as diabetes mellitus, the incidence of Fournier's gangrene has been increasing. Therefore, it is important to understand the risk factors, pathophysiology and clinical course of this disease.

Here, we investigated the characteristics and clinical course of patients diagnosed with Fournier's gangrene at our institution over a period of 10 years with the aim of determining the risks and prognostic factors associated with this disease.

Material and methods

This study was approved by the institutional review board (IRB) of our institution (IRB number: EUMC 2016‐11‐029).

The medical records of 20 patients diagnosed with Fournier's gangrene at our institution from January 2003 to July 2014 were reviewed retrospectively. There were no specific exclusion criteria.

We investigated the following: gender, age, body mass index (BMI), previous operative history and predisposing factors of patients, the number of debridements, details regarding the formation of diverting stoma, pathology and bacteriology results and laboratory results [e.g. white blood cell count, haemoglobin, haematocrit, sodium, potassium, bicarbonate, glucose, blood urea nitrogen (BUN), creatinine, aspartate aminotransferase, alanine aminotransferase, cholesterol, triglyceride, protein, albumin, C‐reactive protein and haemoglobin A1c]. In addition, predisposing factors, aetiological factors and Fournier's Gangrene Severity Index (FGSI) scores were examined to determine risk factors and prognostic factors. The FGSI score was devised by Laor et al. 6, and this index included the parameters of temperature, heart rate, respiratory rate, serum sodium, serum potassium, serum creatinine, serum bicarbonate, haematocrit and white blood cell count to evaluate the severity of the disease.

Data were analysed using IBM SPSS version 20.0 (SPSS^® Inc., Chicago, IL). Numerical data, such as age or scale score, were represented as mean ± standard deviation, and the data about patient characteristics and univariate analysis of prognostic factors were analysed with the χ ²‐test and Fisher's exact test. The mean values of laboratory examination in patients who survived and patients who expired were compared using Student's t‐test. Logistic regression analyses were used for the multivariate analysis of prognostic factors. We considered P‐values less than 0·05 to be statistically significant.

Results

Demographic characteristics and the proportion of patients with a medical history of comorbidities were described in Table 1. Seven patients (35%) were bedridden due to the sequelae of a cerebrovascular event, two patients (10%) were diagnosed with colorectal cancer and had adjuvant chemotherapy and radiotherapy after curative surgery, and one patient (5%) underwent hepatectomy due to hepatocellular carcinoma. Four patients (20%) were chronic alcoholics, with a mean alcohol consumption of 1·8 bottles of soju (Korean distilled spirits) per day (Table 1).

Table 1.

Demographic characteristics of the patients

Characteristics		n (%)
Gender	Male	18 (90%)
	Female	2 (10%)
Age (years)	Mean age	61·8 ± 12·7
	<65	10 (50%)
	≥65	10 (50%)
BMI (kg/m2)	Mean BMI	24·6 ± 5·1
	<25	15 (75%)
	≥25	5 (25%)
History of anorectal surgery	Yes	2 (10%)
	No	18 (90%)
Predisposing factors	Diabetes mellitus	13 (65%)
	Hypertension	10 (50%)
	End‐stage renal disease with haemodialysis	2 (10%)
	Liver cirrhosis	2 (10%)
	Hepatitis B virus carrier	1 (5%)
	Cerebrovascular accident	7 (35%)
	Malignancy	3 (15%)
	Alcoholism	4 (20%)
	Anal surgery	2 (10%)

BMI, body mass index.

Common symptoms included perineal pain (60%), perineal oedema (50%) and anal bleeding (55%). Other symptoms included perineal necrosis or discharge (10%), difficulty in defecation (5%), difficulty in voiding (10%), suprapubic pain (5%), fever (10%) and general weakness (25%). In addition, some patients were admitted to the emergency room due to dyspnoea (5%) or a depressed mental status (5%).

The most common cause (12 patients, 60%) of Fournier's gangrene was a perianal or perirectal infection that spread to the perineum, genitourinary organs, buttocks and abdominal wall. Gangrenous infections also developed after anal surgery and trauma and from sores. Two patients (10%) developed this condition following radiotherapy associated with a malignant disease (Table 2).

Table 2.

The aetiology of Fournier's gangrene in this study

Aetiology	n (%)
Perianal or perirectal infection	12 (60%)
Genitourinary infection	2 (10%)
Postoperative complication	2 (10%)
Infection of sore	1 (5%)
Trauma	1 (5%)
Cancer‐related radiotherapy	2 (10%)

The mean number of detectable bacterial species was 1·5 ± 1·1. Among them, Escherichia coli was the most common (35%), followed by Streptococcus species (20%). Enterococcus faecium (15%), Acinetobacter baumannii (15%) and Staphylococcus aureus (10%) were also detected in this study. Of all patients, 6 (30%) demonstrated polymicrobial infections, and there were no detectable organisms in 2 (10%) patients (Table 3).

Table 3.

Bacteriological results

Bacterial organism	n (%)
Escherichia coli	7 (35%)
Streptococcus milleri group	3 (15%)
Streptococcus agalactiae	1 (5%)
Enterococcus faecium	3 (15%)
Acinetobacter baumannii	3 (15%)
Staphylococcus aureus	2 (10%)
Klebsiella pneumoniae	2 (10%)
Proteus vulgaris	1 (5%)
Proteus mirabilis	1 (5%)
Pseudomonas aeruginosa	1 (5%)
No bacterial growth	2 (10%)
Data missing	4 (20%)

The mean FGSI score was 6·8 ± 5·1; eight patients (40%) had a score > 9 points. Fifteen patients required treatment in the intensive care unit (ICU) with a mean duration of ICU stay of 7·4 ± 6·6 days. The mean number of debridements was 1·7 ± 0·9, and 10 patients (50%) underwent >2 debridements. Colostomy was required in 11 patients (55%) due to an anal sphincter impairment or contamination of a debridement wound. Four patients (20%) required reconstructive surgery with a skin flap due to a broad wound defect (Fig. 1), and vacuum‐assisted closure (VAC^®) dressing successfully decreased the size of the debridement wound, without additional surgical requirement, in two patients (10%) (Fig. 2). Five patients (25%) died because of gangrenous infection (Table 4).

Figure 2.

Vacuum‐assisted closure (VAC^®) dressing. (A) Necrotising fasciitis had spread to the suprapubic abdominal wall. (B) A vacuum‐assisted closure dressing was applied.

Table 4.

The clinical course of patients

Characteristics		n (%)
FGSI (points)	Mean score	6·8 ± 5·1
	≤9	12 (60%)
	>9	8 (40%)
Septicaemia	Yes	3 (15%)
	No	17 (85%)
Mean hospitalisation (days)		36·9 ± 41·3
ICU care	Mean duration (days)	7·4 ± 6·6
	Yes	15 (75%)
	No	5 (25%)
Timing of operation	Mean hospital day of operation (HD)	2·7 ± 1·9
	HD #1 or #2	11 (55%)
	≥ HD #3	9 (45%)
Number of debridements	1	10 (50%)
	2	9 (45%)
	≥3	1 (5%)
Colostomy	Yes	11 (55%)
	No	9 (45%)
Reconstruction with skin flap	Yes	4 (20%)
	No	16 (80%)
Vacuum‐assisted closure	Yes	2 (10%)
	No	18 (90%)
Mortality	Yes	5 (25%)
	No	15 (75%)

FGSI, Fournier's gangrene severity index; HD, hospital day; ICU, intensive care unit.

We investigated the correlation between prognosis and old age (≥65 years), high BMI [≥25 kg/m²], diabetes mellitus, liver cirrhosis, renal failure, malignancy and FGSI score. There was no significant association between diabetes mellitus and mortality in both univariate (P = 0·417) and multivariate analysis [P = 0·781, odds ratio (OR) 2·234, 95% confidence interval (CI) 0·008–643·993]. Nor was there a significant association for hypertension, liver cirrhosis, renal failure, malignancy, cerebrovascular events and alcohol abuse (P < 0·05). Patients with an FGSI score greater than 9 points had a higher mortality rate in univariate analysis (P = 0·035); however, this was not significant in the multivariate analysis (P = 0·526, OR 3·375, 95% CI 0·079–144·394).

In terms of laboratory analysis, the mean level of BUN was higher and that of albumin was lower in patients who died (79·0 ± 54·5 and 2·2 ± 0·3 mg/dl, respectively) than in those who survived (25·1 ± 13·8 and 2·9 ± 0·6 mg/dl, respectively). However, these findings did not show a statistical association with prognosis (P = 0·091 and 0·052, respectively). On the other hand, when prognosis was examined according to BUN level, those with a level > 50 mg/dl had a higher mortality rate than those with levels < 50 mg/dl by both univariate (P = 0·001) and multivariate analysis (P = 0·033, OR 31·090, 95% CI 1·308–738·865) (Table 5). On the other hand, there was no significant difference in the prognosis between patients with and without hypoalbuminemia (P = 0·160) (Table 5).

Table 5.

Risk factors for a poor prognosis in Fournier's gangrene

	Total (n = 20)	Expired (n = 5)	Univariate analysis		Multivariate analysis
			P‐value	RR	P‐value	OR (95% CI)
Age (years)
<65	10	1 (10%)	0·121	6·000	0·807	1·595 (0·038–67·590)
≥65	10	4 (40%)
Gender
Male	18	4 (22·2%)	0·389	3·500	0·987	0·963 (0·008–110·394)
Female	2	1 (50%)
BMI (kg/m2)
<25	15	3 (20%)	0·371	2·667		—
≥25	5	2 (40%)
FGSI (points)
≤9	12	1 (8·3%)	0·035*	11·000	0·526	3·375 (0·079–144·394)
>9	8	4 (50%)
Diabetes mellitus
Yes	13	4 (30·7%)	0·417	2·667	0·781	2·234 (0·008–643·993)
No	7	1 (14·3%)
Liver cirrhosis
Yes	2	1 (50%)	0·389	3·500		‐
No	18	4 (22·2%)
Renal failure
Yes	2	1 (50%)	0·389	3·500		‐
No	18	4 (22·2%)
Malignancy
Yes	3	1 (33·3%)	0·718	1·625		‐
No	17	4 (23·5%)
Timing of operation
HD #1 or #2	11	2 (18·2%)	0·617	2·250		‐
≥HD #3	9	3 (33·3%)
Number of debridements
≥2	10	2 (20%)	0·606	0·583		‐
1	10	3 (30%)
Diversion stoma
Yes	11	2 (18·2%)	0·436	0·444		‐
No	9	3 (33·3%)
ICU care
Yes	15	5 (33·3%)	0·136	0·667		‐
No	5	0 (0%)
BUN level (mg/dl)
<50	15	1 (6·7%)	0·001*	56·000	0·033*	31·090 (1·308 − 738·865)
≥50	5	4 (80%)
Albumin level (mg/dl)
<3·0	14	5 (35·75)	0·160	0·692		‐
≥3·0	4	0 (0%)

CI, confidence interval; BUN, blood urea nitrogen; BMI, body mass index; FGSI, Fournier's gangrene severity index; HD, hospital day; ICU, intensive care unit; RR, relative ratio; OR, odds ratio.

Discussion

Fournier's gangrene is not common and can still be fatal despite medical development. In particular, the prevalence of this disease has recently begun to increase due to various predisposing factors, such as diabetes mellitus, immune‐related disease and malignancy 7. According to previous studies, the incidence is 1:7500–1:750 000 people 2 or 1:25 000–1:250 000 people 8, and the mortality rate is 3–67% 2, 9. Kim et al. 7 reported a mortality rate of 14·8%, and another study in Korea 10 reported a rate of 27·2%. In this study, the mortality rate was 25%.

This disease can occur in both men and women; however, men are known to be more susceptible, with a ratio of 10:1 2, 11. In this study, there were two female patients, and the male‐to‐female ratio was 9:1. Bilton et al. 12 found that Fournier's gangrene occurred predominantly in individuals aged between 30 and 60 years, and another recent study also reported that, although this disease can occur at all ages, it is most common in individuals in their fifties 2. However, the mean age of patients in this study was 61·8 ± 12·7 years, and 50% of patients were aged over 65 years. This implies that the age of those affected by Fournier's gangrene is increasing. In terms of BMI, 25% of the patients in this study were obese according to World Health Organization (WHO) criteria (Table 1). Norton et al. 13 reported that obesity was a predisposing factor in Fournier's gangrene; however, our results did not show a significant association with prognosis.

Various underlying diseases have been considered to contribute to the occurrence and aggravation of Fournier's gangrene. Diabetes mellitus, alcohol abuse, immune suppression, malignancy, liver disease and renal disease are reported to be positively associated with mortality 2, 14, 15; in particular, diabetes mellitus is the most common and important predisposing factor 2, 16. However, the relationship between diabetes mellitus and mortality is controversial, with some studies showing an association with both incidence and mortality 17, 18, while others show an association with incidence but not with mortality 6, 16. In this study, 65% of patients had diabetes mellitus; however, there was no significant association with mortality.

Reports on the timing of hospital presentation vary, with the period from symptom manifestation to hospital attendance ranging from approximately 1 to 30 days 19, 20. Generally, patients attend an emergency room within 2–7 days of complaining of symptoms such as perineal oedema or discomfort, purulent exudates, necrosis or shock 2. In this study, patients already had advanced perineal inflammation upon presentation.

Aetiological factors vary; in most cases, the disease is related to sanitary conditions of the perineum 21. Infectious sources are mostly anorectal diseases, urological diseases, intraperitoneal diseases or traumatic injuries 9, and in some cases, the cause cannot be deciphered. In this study, perianal or perirectal infections were the most common causes, while one patient had an infected sore and another had an infection in a traumatic wound. Notably, two patients had severe infections after rectal cancer radiotherapy, in keeping with a previous study, which reported that malignant disease and immunosuppression as a result of treatment could be risk factors for Fournier's gangrene 15. The proportion of rectal cancer patients has recently been increasing, and radiotherapy is a major and important part of treatment; therefore, surgeons should be aware of the relationship between malignancy or radiotherapy and Fournier's gangrene.

The causative organisms in Fournier's gangrene are mainly aerobic Gram‐negative bacilli or Gram‐positive cocci, and most cases include a mixed infection of more than three organisms 14, 19, such as Escherichia coli or Proteus, Enterococcus, Pseudomonas or Klebsiella species 15, 22. The most commonly detected bacteria is Bacteroides fragilis 14, 23, while Staphylococcus aureus or Streptococci are frequently identified in patients with diabetes mellitus 24. In this study, E. coli was the most commonly detected organism in 35% of patients, followed by Streptococcus species at 20%. Enterococcus, Acinetobacter, Staphylococcus, Klebsiella, Proteus and Pseudomonas species were also cultured, and for two patients, no bacterial organisms were identified.

The management of Fournier's gangrene can be summarised into three parts: early detection, immediate and aggressive debridement and adequate administration of broad‐spectrum antibiotics 2, 6, 19. Norton et al. 13 also suggests patient stabilisation, broad‐spectrum antibiotic administration and early invasive surgical management, with the most important aspect being prompt surgical management. In this study, all patients underwent aggressive debridement and broad‐spectrum antibiotic therapy. If additional necrotic lesions were identified after the first debridement, debridement was repeated, with a mean number of 1·7 debridements conducted. In this study, 55% of patients required ostomy, and although there was no significant association between ostomy and prognosis, other studies have reported that patients with ostomy had a poor clinical course 25. In fact, ostomy can help to heal the wounds caused by an anorectal infection in Fournier's gangrene 2. Four patients (20%) required skin grafts, and in two, the wide wounds were healed by a VAC dressing without additional surgical requirements. The VAC device usually consists of a sterile, open‐cell foam sponge that is placed in the wound and then covered with a transparent adhesive drape and non‐collapsible tubing that is connected to a suction pump that provides continuous negative pressure. This device promotes debridement and increases perfusion, fibroblast migration and cell proliferation in the wound, thus facilitating rapid wound closure, drawing the wound edges together and reducing the would surface area 2. Thus, the adequate application of VAC dressings may be helpful in wound healing, thereby reducing the requirement for additional skin graft surgery.

The mortality rate for Fournier's gangrene is 3–67% 2, 9. The known prognostic factors are age; a history of hypertension, heart failure, renal failure or coagulopathy; procedures performed during admission (e.g. ostomy, ventilator care with intubation and haemodialysis); hospitalisation 2, 25; the size of the necrotic area 26, 27; FGSI score 2, 6; and sepsis or shock at admission 28. In this study, 65% of patients had diabetes mellitus; however, there was no significant association between diabetes mellitus and mortality. Other underlying diseases, such as hypertension, liver cirrhosis, renal failure, malignancy, cerebrovascular events and alcohol abuse, were not significantly associated with prognosis either. Only patients with a FGSI score higher than 9 points had a higher mortality rate. FGSI describes severity by scoring vital signs and metabolic parameters 6 and may be helpful in making initial therapeutic decisions, such as the performance of more active and broad treatments in patients with a score >9. In terms of laboratory analysis, cases with higher BUN levels (>50 mg/dl) 29 or with hypoalbuminemia 30 have been shown to have a poorer prognosis. In this study, if the initial BUN level was >50 mg/dl, patients were predicted to have a poorer prognosis; however, there was no significant difference in prognosis for patients with and without hypoalbuminemia. Although these prognostic factors can help to determine the principles of treatment, the most important factor is adequate and active treatment according to disease manifestations, vital signs and laboratory results, which may vary depending upon the timing of hospital presentation.

The main limitation of this study was the small sample size. However, as the incidence of Fournier's gangrene is low, the 20 patients presenting at this one institution is representative.

Conclusions

The incidence of Fournier's gangrene is predicted to increase due to the aging demographics and increases in metabolic or immune‐related diseases in developed countries. Despite medical developments, the mortality rate of this disease is high as ever. The most effective form of disease management is aggressive debridement and combined antibiotic therapy. Moreover, a VAC dressing can help to promote wound healing and to reduce the chance of reconstructive surgery with skin grafting. In addition, an adequate understanding of Fournier's gangrene by treating physicians is very helpful. Therefore, surgeons should be mindful that prognosis can be improved through a thorough understanding of predisposing diseases or prognostic factors at initial presentation and that the performance of procedures, such as diverting stoma or haemodialysis, can improve prognosis.