Impact of surgical intervention timing on the case fatality rate for Fournier's gangrene: an analysis of 379 cases

Authors


Toru Sugihara, Department of Urology, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. e-mail: ezy04707@nifty.com

Abstract

Study Type – Prognosis (outcome)

Level of Evidence 2b

What's known on the subject? and What does the study add?

Reportedly, Fournier's gangrene has a high mortality rate, ∼7.5–40%, and experts recommend early surgical debridement.

This study examines 379 patients and shows that an early intervention, i.e. within 2 hospital days could halve the mortality rate compared with later intervention.

OBJECTIVE

  • • To examine how early surgical intervention influenced cases of Fournier's gangrene (FG) fatality.

PATIENTS AND METHODS

  • • Patients with FG (defined as an International Classification of Diseases-10 code of M72.6 [necrotizing fasciitis] at the perineum or external genitalia), who received surgical intervention ≤5 days after admission, were identified from the Diagnosis Procedure Combination database for the 6-month period July to December, in the years 2007–2010.
  • • Data included age, sex, comorbidities, ambulance use, operations and debridement ranges.
  • • Multivariate logistic regression analysis of mortality was performed to show whether early (≤2 hospital days) or delayed (3–5 hospital days) surgical treatment affected FG outcomes.

RESULTS

  • • A total of 302 male and 77 female patients with FG were identified for which the overall case fatality rate was 17.1% (65 cases).
  • • There were no significant differences in patient characteristics between the early operation group (n= 327) and the delayed operation group (n= 52), with the exception of ambulance use (33.3% vs 17.3%, P= 0.020).
  • • Cystostomy, colostomy, orchiectomy/penectomy (male patients only), or debridement ≥3000 cm2 were performed on 42 (8.8%), 56 (11.5%), 46 (10.8%) and 17 (4.4%) patients, respectively.
  • • Multivariate analysis showed that there was a significantly lower case fatality rate among the early operation group (odds ratio [OR]= 0.38; P= 0.031).
  • • Older age (OR 1.80, for 10-year increments), Charlson comorbidity index score (OR = 1.33, for 1-point increments), sepsis or disseminated intravascular coagulation at admission (OR 4.01), and debridement range ≥3000 cm2 (OR 5.22, compared with other operations) were significantly associated with a higher case fatality rate.

CONCLUSION

  • • Early (≤2 hospital days) surgical intervention for FG is significantly associated with lower mortality than delayed (3–5 hospital days) action.
Abbreviations
FG

Fournier's gangrene

ICD

International Classification of Diseases

OR

odds ratio

DPC

Diagnosis Procedure Combination

CCI

Charlson comorbidity index

DIC

disseminated intravascular coagulation.

INTRODUCTION

Fournier's gangrene (FG) is a well-known but rare life-threatening disease characterized by rapid progression of necrotizing fasciitis. The name of the disease is credited to Jean Alfred Fournier, who described five young men with genitalia gangrene without apparent causal factors in 1883 [1]. Despite modern advances in intensive care, the case fatality rate of FG reported in several series remains high at 7.5–40% [2,3]. Because it is a rare disease, small case series have been the main sources of information about its clinical characteristics, and reliable statistical analyses of factors related to the disease remain limited [4]. Although several experts recognize prompt and aggressive surgical debridement of necrotic tissue as effective treatment for FG [3,5], to our knowledge there is no quantitative evidence demonstrating a mortality reduction benefit from early surgical intervention.

In the present study, we determined the occurrence of FG among both sexes, and assessed the effectiveness of early intervention for the reduction of mortality using the Diagnosis Procedure Combination (DPC) database in Japan.

PATIENTS AND METHODS

DPC DATABASE

The nature of the DPC database has been described previously [6,7]. Briefly, it is a case-mix inpatient administrative claims database in Japan and contains: (i) a main diagnosis, comorbidities at admission, and complications after admission (up to four items), which are described in Japanese with the corresponding International Classification of Diseases (ICD)-10 codes; (ii) the surgical interventions accompanied by original Japanese K-codes; and (iii) the discharge status (dead or alive). The DPC database has collected inpatient data from July 1 to December 31 every year (6 months per year) since its inception in 2002. The number of cases in the database was 2.99, 2.86, 2.57 and 3.19 million in 2007, 2008, 2009 and 2010, respectively, and represented ∼40% of all acute care inpatient hospitalizations in Japan.

Study approval was obtained from the Institutional Review Board of the Tokyo Medical and Dental University, Japan. Given the anonymous nature of the data collection process, informed consent was not required.

SAMPLING STRATEGY

We obtained inpatient data for 2007–2010. To test our hypothesis that early surgical intervention reduces FG case fatality we identified, from these data, patients with FG who underwent early surgical intervention (within 2 days of admission) and delayed surgical intervention (within 3–5 days of admission).

Cases of FG were defined as those with necrotizing fasciitis (ICD-10 code, M72.6) accompanied by the Japanese terms meaning perineum, external genitalia, penis, scrotum contents and vulva. The validity of the Japanese disease descriptions were reviewed by two authors (T.S. and H.Y.).

For covariates, we extracted the following information from the database: (i) patient information including age, sex, comorbidities at admission, ambulance use (yes or no), and discharge status (dead or alive); and (ii) surgical procedure codes and number of days from admission to the first surgical procedure (wound repair: K000x–001x; debridement: K002x; skin graft: K013x–017; colostomy: K725–726, K736x; removal of perirectal abscess: K737, K745; proctectomy: K740x; cystostomy: K805; penectomy or orchiectomy: K825–826x, K830x; perineoplasty: K851x). Patients with FG who did not have a surgical procedure, or underwent a surgical procedure >5 days after admission were excluded. Comorbidities at admission were converted into the Charlson comorbidity index (CCI), a well-known pre-existing risk adjuster widely used to assess the burden of comorbidities, and used especially in administrative claims database analyses [8]. Sepsis and/or disseminated intravascular coagulation (DIC) at admission (ICD-10, A02.1, A20.7, A22.7, A24.1, A26.7, A28.2, A32.7, A39.4, A40x–41x, A42.7, B00.7, B37.7 for sepsis; D65 for DIC) were also identified, as sepsis and DIC are important severe conditions but are not included as components of the CCI. We were able to classify the ranges of debridement into three levels by surgical K-codes, as debridement surgeries have three different codes according to debridement range: K002.1, K013.1–013.2, 013–2.2, K015.1–015.2 for <100 cm2; K002.2, K013.3–013.4, K013–2.4, K015.3, K016 for 100–3000 cm2; and K002.3 for ≥3000 cm2.

STATISTICAL ANALYSIS

We described patients' demographic characteristics and frequencies of surgical interventions, and then performed univariate logistic regression analyses of mortality. Next, we used multivariate logistic regression to identify whether early intervention could reduce FG case fatality rates. Independent variables included in the multivariate model were timing of surgical intervention (early or delayed), age, sex, CCI score, and other factors that reached significance at P < 0.10 in univariate models (Table 2).

In univariate comparisons, categorical data were compared using the chi-squared test and continuous data were compared using the Mann–Whitney U-test. A P value of <0.05 was considered to indicate statistical significance. All statistical analyses were conducted using SPSS version 19.0 (IBM, Chicago, IL, USA).

RESULTS

Among 11.6 million inpatients in the DPC database for the period 2007–2010 we identified 379 patients with FG from 242 hospitals who underwent surgical intervention within 5 days of admission. The patients' demographic characteristics and surgical interventions are shown in Table 1. The male to female ratio was 3.92. The median age was significantly higher in females. (61 years in males and 71 years in females; Mann–Whitney U-test, P < 0.001) Patients in the early operation group were more frequently transferred by ambulance (33.3% in the early vs 17.3% in the delayed group; chi-squared test, P= 0.020) Case fatality was lower in the early group than the delayed group, but the difference was not significant (16.2% vs 23.1%, P= 0.222).

Table 1. Distribution of patients' characteristics and surgical interventions between the early and delayed operation groups
 Surgical intervention interval P
Early: within 1–2 daysDelayed: within 3–5 days
  • *

    Male patients only.

Overall, n (%)327 (100.0)52 (100.0) 
Female, n (%)67 (20.5)10 (19.2)0.834
Median (sd) age, years63 (14.5)61 (18.7)0.428
Median (sd) CCI score1 (2.08)1 (1.84)0.314
Diabetes mellitus163 (49.8)26 (50.0)0.984
Sepsis/DIC at admission78 (23.9)8 (15.4)0.176
Ambulance user109 (33.3)9 (17.3)0.020
Surgical interventions, n (%)   
 Cystostomy37 (11.3)5 (9.6)0.717
 Colostomy51 (15.6)5 (9.6)0.259
 Orchiectomy/penectomy*40 (15.4)6 (14.3)0.854
 Debridement  0.551
  Others211 (64.5)31 (59.6) 
  <100 cm224 (7.3)4 (7.7) 
  100–3000 cm276 (23.2)16 (30.8) 
  ≥3000 cm216 (4.9)1 (1.9) 
Case fatality rate53 (16.2)12 (23.1)0.222

Table 2 shows the results of univariate analyses of mortality among each subgroup. With regard to patients' pre-admission factors, high CCI score, older age, sepsis/DIC at admission and ambulance use were significantly associated with high case fatality. Among surgical interventions, only debridement of a wider range was related to high case fatality.

Table 2. Univariate analysis of case fatality rates among 379 patients with FG, who underwent surgical interventions within 5 hospital days
 Case fatality rate, %OR (95% CI) P
  • *

    Male patients only.

Overall17.1  
Female (reference male)23.41.66 (0.90–3.05)0.107
Age (by 10-year increments) 1.63 (1.33–2.01)<0.001
CCI score (by 1-point increase) 1.34 (1.19–1.50)<0.001
Diabetes mellitus15.30.78 (0.45–1.33)0.353
Sepsis/DIC at admission34.83.94 (2.24–6.96)<0.001
Ambulance user23.71.88 (1.09–3.26)0.024
Surgical interventions   
 Cystostomy19.01.16 (0.51–2.63)0.730
 Colostomy16.10.91 (0.42–1.97)0.817
 Orchiectomy/penectomy*21.71.64 (0.75–3.60)0.213
 Debridement   
  <100 cm210.7reference 
  100–3000 cm27.60.68 (0.17–2.85)0.604
  ≥3000 cm241.25.83 (1.25–27.1)0.025

Figure 1 shows the relationship between case fatality and the number of days from admission to the first surgical procedure. Case fatality rates were 16.0, 16.9, 19.0, 26.7 or 25.0% among patients who underwent surgery 1, 2, 3, 4 or 5 days after admission, respectively.

Figure 1.

Relationship between case fatality and the number of days from admission to the first surgical procedure.

Table 3 shows the multivariate logistic regression analysis of case fatality. Early surgical intervention (within 2 days of admission) was significantly correlated with a lower case fatality rate (dds ratio [OR]= 0.38, P= 0.031). Higher case fatality was significantly associated with older age, higher CCI score, sepsis/DIC at admission and a larger debridement range.

Table 3. Multivariate logistic regression analysis of case fatality among 379 patients with FG undergoing early or delayed surgical intervention
 OR95% CI P
Female (reference male)1.140.55–2.380.730
Age (by 10-year increase)1.801.39–2.33<0.001
CCI score (by 1-point increase)1.331.17–1.52<0.001
Sepsis/DIC at admission4.012.03–7.95<0.001
Ambulance use1.460.76–2.820.262
Debridement range (reference, others) (unspecified ranges)   
 <100 cm20.290.08–1.050.060
 100 – 3000 cm20.450.19–1.040.062
 ≥3000 cm25.221.49–18.290.010
Intervention within 2 hospital days (reference 3–5 hospital days)0.380.16–0.920.031

DISCUSSION

Early surgical intervention (within 2 days of admission) significantly reduced FG case fatality compared with delayed intervention (within 3–5 days of admission). Historically, several experts' have supported early aggressive debridement with broad-spectrum antibiotics as treatment for FG [5,9]. Korkut et al. [10] retrospectively analysed nine fatal cases among 45 patients between 1990 and 2001 and found that all fatal cases had a long interval from the onset of clinical symptoms to the initial surgical intervention (>6 days). They suggested that the interval could have a significant impact on mortality; however, another study of 68 cases failed to demonstrate this tendency [11]. Theoretically, it is quite reasonable to remove necrotic tissues quickly, with the intention of halting the progress of FG [3]; however, because of the rarity of FG, previous case series studies have mainly been provided from tertiary medical care centres, and the number of patients in these studies has been limited to a maximum of 80 cases [2,12]. For that reason, quantitative assessments of the potential benefit of prompt intervention have never been undertaken. We were able to overcome these limitations by taking advantage of the DPC database because it contains information on comorbidities, operation timing and debridement ranges, and has a high representativeness; cases represent ∼ 40% of Japanese acute hospitalizations.

According to the present data, 52 patients (13.7%) received delayed surgeries. We hope the present results will promote early treatment actions for FG and contribute to a further decline in FG mortality rates.

Despite today's advances in acute care management, FG still had a high case fatality rate. The case fatality rate in the present study was 17.1%. Previously reported FG mortality rates vary widely because mortality depends on the study population, available hospital facilities, staffing levels, and time periods [3,5]. Eke [3] summarized 1726 FG cases among studies published from the 1950s to 1990, and reported an FG case fatality rate of 16%. According to 27 contemporary case series listed by Sorensen et al. [2], the case fatality rates range between 25 and 88% in the 1980s and 8.8 and 24% in the 2000s, which implies a decrease in mortality over time; however, those studies included data from 1641 males with FG (the largest single study), from the State Inpatient Database in the USA and estimated a population mortality rate of 7.5%, which is less than half the rate found using Eke's data and the present data. Sorensen et al. acknowledge that their estimated FG mortality was substantially lower than previously reported in the literature, but we believe that their methodology could have led to contamination and underestimation of mortality. To identify FG patients, they used the ICD-9-CM system and defined patients as those diagnosed with a code of 608.83 (vascular disorders of male genital organs) who underwent some kind of excision and debridement or died. The code 608.83 is not limited to necrotizing fasciitis; it is also applicable for haematoma of the seminal vesicle, spermatic cord, testis, and scrotum. We therefore conclude that the cases included in the study by Sorensen et al. were contaminated with non-necrotizing cases such as ‘scrotum haematoma undergoing haematoma evacuation’ and this led to a lower mortality rate.

In terms of surgical intervention, cystostomy, colostomy and orchiectomy did not significantly influence mortality. Whether stoma formation reduces the FG case fatality is controversial. Some publications describe an increased mortality among patients with FG who require stomas [9,10]; however, it is conceivable that surgeons are more likely to select stoma formation for the treatment of patients with advanced disease. Other recent studies do not find mortality differences between patients with and without stoma formation, and argue that stoma formation should be limited to selected patients to protect wounds from faecal contamination [4,13]. In the present analysis, we found that patients who needed stoma formations, and those who did not, had an equal survival rate.

Other factors affecting case fatality were age (OR = 1.80, for 10 years increments), CCI score (OR = 1.33, for 1 point increments) and sepsis/DIC at admission (OR = 4.01). Diabetes mellitus on its own (P= 0.353), and ambulance use (P= 0.076) were not significantly associated with FG case fatality. These results are consistent with those of the study by Sorensen et al. [14] with the exception of ambulance use, which was as a significant factor in that study. Mills et al. [15] analysed 688 cases of necrotizing soft-tissue infections (not limited to FG) from the National Surgical Quality Improvement Program Registry and found that age and several comorbidities correlated with mortality, but sex and diabetes mellitus did not. We suggest, therefore, that increasing age and high CCI score are obvious risk factors for FG case fatality, although they have not been adopted as indicators of the FG Severity Index, which is calculated from a combination of physiological variables at admission and is widely used to predict FG outcomes [11,16]. Diabetes mellitus is a well-known predisposing factor for FG [5], and observational data have identified a spatial relationship between high prevalence of diabetes mellitus and a high incidence of FG [2]; however, the majority of data (including our own) suggest that diabetes mellitus on its own does not have a strong influence on FG case fatality.

Several caveats must be considered when interpreting the results of the present study. Because the diagnosis of FG is completely subjective (the doctor's judgement), an over- or underestimation of diagnosis may have occurred. Furthermore, because critically ill patients are likely to be transported to large hospitals, and because there is a bias toward large hospitals in the DPC database [17], the calculation of case fatality could be an overestimation. Debridement ranges were available for only 37% of cases. Finally, we were not able to access several important clinical variables, such as the results of laboratory tests, microbiological cultures, and imaging tests. We lacked the data from the FG Severity Index for each patient; however, we believed that sepsis/DIC at admission and ambulance use could be good proxy markers for severity.

Despite these limitations, using a nationwide database with high representativeness enabled us to perform a quantitative study of this rare disease, and to show the necessity of rapid medical intervention to improve FG outcomes.

In conclusion, the FG case fatality rate of 17.1% is still high, and increasing age, CCI score and sepsis/DIC at admission influenced mortality. Surgical intervention within 2 days of admission is significantly associated with low mortality.

ACKNOWLEDGEMENTS

This study was funded by a Grant-in-Aid for Research on Policy Planning and Evaluation from the Ministry of Health, Labour and Welfare, Japan (Grant number: H22-Policy-031) and by the Funding Program for World-Leading Innovative R&D on Science and Technology (FIRST program) from the Council for Science and Technology Policy, Japan (Grant number: 0301002001001).

CONFLICT OF INTEREST

None declared.

Ancillary