Fournier’s gangrene at a tertiary health facility in northwestern Tanzania: a single centre experiences with 84 patients

A total of 88 patients were seen over a period of study with four patients excluded
from the study on account of incomplete data, leaving 84 patients for the final analysis.
There were 82 (97.6 %) males and 2 (2.4 %) females with a male to female ratio of
41: 1. The age of patients at presentation ranged from 15 to 76 years with a median
of 34 years (+IQR of 32 to 41 years). The modal age group was 31–40 years accounting
for 58.3 % of cases (Fig. 1). Most of patients, 68 (81.0 %) had either primary or no formal education and more
than 80 % of them were unemployed. The majority of patients, 72 (85.7 %) came from
the rural areas located a considerable distance from the study area and more than
ninety percent of them had no identifiable health insurance.

Fig. 1. Distribution of patients according to the age group

Predisposing factors for Fournier’s gangrene were identified in 69 (82.1 %) patients
of which diabetes mellitus was the leading predisposing factor associated with Fournier’s
gangrene and accounted for 14 (16.7 %) patients. The predisposing factors for Fournier’s
gangrene were not identified in 17.9 % of cases (Table 1).

Table 1. Distribution of patients according to predisposing/etiological factors

The scrotum was the most frequent anatomical location involved accounting for 78.6 %
of cases. Other anatomical location and extent are shown in Table 2.

Table 2. Anatomical location and extent of Fournier’s gangrene

The duration of symptoms before admission ranged from 1 to 18 days with a median of
4 days (+IQR of 2–8 days). The majority of patients, 64 (76.2 %) sought medical attention
after more than 48 h of the onset of symptoms. The onset of symptoms was insidious
in 48 (57.1 %) patients and abrupt in 36 (42.9 %) patients. Of the 84 patients, 62
(73.8 %) presented with established ulcers of the involved area, 12 (14.3 %) had established
gangrene while the remaining 10 (11.9 %) presented with cellulitis, which later progressed
to scrotal gangrene or resolved with treatment. All patients presented with fever,
pain and discharge of sero-purulent material from the site. They also had swelling,
redness, tenderness, and black dermal necrosis of the skin.

Hematological investigations revealed leukocytosis in all the patients (100 %), elevated
ESR in 76 (90.5 %) and anemia (hemoglobin 10 g/dL) in 80 (95.2 %) patients. Out of
80 patients tested for HIV infection, nine (11.3 %) were HIV positive. Of the HIV
positive patients, three (33.3 %) patients were known cases on ant-retroviral therapy
(ARV) and the remaining 6 (66.7 %) patients were, CD 4+ count among HIV positive patients
was available in only 5 patients and ranged from 113 to 678 cells/?l with a median
of 236 cells/?l (+IQR of 232–240 cells/?l).

In this study, aerobic bacterial culture results were obtained in only 46 (54.8 %)
patients. Of these, 38(82.6 %) had polymicrobial bacterial growth while 8 (17.4 %)
had monomicrobial bacterial growth. Escherichia coli (28.3 %) and Staphylococcus aureus (17.4 %) were the most frequent bacterial organisms isolated (Table 3). Anaerobic cultures were not done due to lack of facility to perform this test.
Antibacterial susceptibility testing revealed that most of pathogens isolates had
multiple resistant to almost all tested antibiotics (such as ampicillin, augumentin,
cotrimoxazole, tetracycline, penicillin, gentamicin, erythromycin, oxacillin etc.)
except for Meropenem and imipenem, which were 100 % sensitive each respectively.

Table 3. Distribution of patients according to the type of micro-organisms isolated

Before surgery, all patients underwent aggressive fluid resuscitation and were treated
mostly with parenteral broad-spectrum triple antimicrobial agents, using a third-generation
cephalosporin, an aminoglycoside and metronidazole and received hemodynamic support
when required. Meropenem and imipenem, which were 100 % sensitive were given to only
few patients who managed to purchase. Mechanical ventilation, continuous monitoring,
and inotropic support were applied when necessary in patients with cardiopulmonary
failure due to sepsis. All patients underwent radical surgical debridement, ranging
from 1 to 8 procedures, with a median of 3 (+IQR of 1 to 5). Debridement consisted
of excision of all necrotic tissue, cleansing with hydrogen peroxide, then saline
and packed with dressings soaked in povidone-iodine. Orchidectomy was carried out
unilaterally for gangrenous testes in 3 (3.6 %) patients. There were no adjunct surgeries
such as suprapubic cystostomy, colostomy or penectomy. After the initial surgery,
the wound was closely monitored, adequate nutrition was ensured to support wound healing
and early enteral feeding was considered. The patients underwent repeated debridements
as necessary and further necrotic tissues were debrided when needed under local or
no anesthesia. Closure of wounds was commenced as soon as healthy, viable tissue allowed
re-approximation. The majority of the patients, 65 (77.4 %) had wound closure by secondary
closure. Other patients had their wound closed by skin grafting and flap rotation
in 14 (16.7 %) and 5 (5.9 %) patients respectively.

The overall length of hospital stay (LOS) ranged from 8 to 56 days with a median of
28 days (+IQR of 26 to 32 days). The median LOS for non-survivors was 7 days (range
1–16 days). Factors predicting the length of hospital stay in univariate analysis
were systemic inflammatory response syndrome on admission (OR = 2.4, 95 %CI (1.1–6.3),
p = 0.034), late presentation 48 h (OR = 4.8, 95 % CI (3.2–6.9), p = 0.016), HIV
positivity (OR = 5.2, 95 % CI (2.8–9.0), p = 0.003) and diabetes mellitus (OR = 2.1,
95 % CI (1.0–6.7), p = 0.011). According to multivariate logistic regression analysis,
systemic inflammatory response syndrome [odds ratio (OR) = 3.9, 95 % confidence interval
(CI): 1.5–8.5, p = 0.002) and diabetes mellitus (OR = 3.0, 95 % CI: 2.1–5.9, p = 0.044)
were the main predictors of the length of hospital stay.

In this study, twenty-four patients died giving a mortality rate of 28.6 %. Univariate
analysis revealed high mortality rates in the following conditions; advancing age
60 years (p = 0.003), hospital admission later than 48 h of onset of symptoms (p = 0.012),
systemic inflammatory response syndrome on admission (p = 0.001), diabetes mellitus
(p = 0.002), HIV positivity (p = 0.005), low CD 4 count (200 ?l/cells) and FGSI 9
(p = 0.003). According to multivariate logistic regression analysis; advancing age
60 years (Odds ratio = 2.4, 95 % CI (1.7–4.6), p = 0.002), late presentation 48 h
(Odds ratio = 3.9, 95 % CI (2.2–9.4), p = 0.000), systemic inflammatory response syndrome
on admission (Odds ratio = 7.3, 95 %CI (2.5–9.7), p = 0.016), diabetes mellitus (Odds
ratio = 2.2, 95 % CI (1.7–5.9), p = 0.010), extension of infection to the abdominal
wall (Odds ratio = 6.7, 95 %CI (3.8–9.0), low CD 4 count (200 ?l/cells) (Odds ratio = 4.9,
95 % (2.1–8.7), p = 0.021) and FGSI 9 (Odds ratio = 3.7, 95 % CI (2.8–5.6), p = 0.004).