Surgical Site Infection in Surgery for Benign Prostatic Hyperplasia: Comparison of Two Skin Antiseptics and Risk Factors

Patients and Methods

The randomized single-center study enrolled 70 patients operated on for BPH in a tertiary general hospital over a period of six months (February–August 2009). The study protocol was approved by our hospital's Institutional Review Board, and informed consent was obtained from all patients.

Patients were operated on by either TUR or open surgery according to prostate size, and all patients had a minimum postoperative followup of three years. Data collection was performed by the surgical team in the perioperative period and stored in a Microsoft Excel^® spreadsheet (Redmond, WA). Several factors were considered as potential risk factors for SSI: Patient age, prostatic size (estimated by digital rectal examination and transabdominal ultrasonography), weight of the enucleated prostatic tissue, presence of an indwelling urinary catheter preoperatively, findings at preoperative urine culture, surgical technique, and operative time. Additionally, we considered the presence of renal dysfunction, bladder stones, and vesicocutaneous fistula as potential risk factors for SSI.

We randomly used as skin antiseptics either 0.5% povidone–iodine (n=24 [43%]) or Chemisol^® (0.5% chlorhexidine in an alcohol base) (n=32 cases [57%]). The antibiotic, administered during the induction of anesthesia, was ciprofloxacin 400 mg intravenously, which was repeated every 12 h until withdrawal of the catheter postoperatively. We used the CDC criteria to define SSI. Preoperative urine cultures were obtained as much as 15 d prior to surgery by spontaneous voiding or during exchange of urinary catheters.

Randomization was performed with the software GraphPad^® (GraphPad Software, La Jolla, CA). Data analysis was done with Statistical Package for Social Sciences (SPSS)^® (IBM Corp., Armonk, NY). Comparison of continuous variables was performed with the Student t-test. The relation between two categorical variables was studied using the χ² and Fisher exact tests. We considered as statistically significant difference of p<0.05.

Results

The mean age of the study population was 72 years (range 57–87 years), and prostate size by digital rectal examination corresponded to grade 1 in 11 patients (16%), grade 2 in 23 patients (33%), and grade 3 in 36 patients (51%). Mean prostate weight by transabdominal ultrasonography was 93 g (range 16–434 g), and the mean weight of the enucleated tissue was 56.4 g (range 15–199 g). Forty-one patients (59%) had a preoperative urinary catheter (preUC), with an average time of 10 mo between catheter placement and surgery. Positive preoperative urine culture (posUC) occurred in 31 cases (45%), of which 29 (94%) had a urinary catheter. Of the 31 posUC, 20 (65%) were of polymicrobial flora. Of the single microorganisms, nine (82%) were gram-negative bacteria (Escherichia coli in five and Enterobacter spp. in four), and two were gram-positive bacteria: One Enterococcus spp. and one Staphylococcus aureus (Table 1). The rate of resistance to ciprofloxacin was 64%, to ampicillin-sulbactam 40%, to cotrimoxazole 80%, to third-generation cephalosporins 33%, to gentamicin 30%, to amikacin 20%, and to carbapenems, zero (Table 2).

Of the 70 patients, 16 were operated on by TUR, including one who required conversion to open cystolithotomy and another to open adenomectomy. Therefore, 56 patients (80%) underwent open surgery, 29 underwent transvesical adenomectomy, 26 underwent transcapsular adenomectomy, and one had open cystolithotomy.

Surgical site infection occurred in 10 patients (18%); in seven, the causative microorganism was identified, all being gram-negative bacteria (four Escherichia coli, two Klebsiella spp., and one Enterobacter spp.; see Table 1). All organisms were resistant to ciprofloxacin, 83% were resistant to ampicillin/sulbactam, 71% were resistant to cotrimoxazole, 50% were resistant to third-generation cephalosporins, 29% were resistant to gentamicin, and none were resistant to amikacin or carbapenems (Table 2). We did not find a significant relation between the kind of antiseptic and the occurrence of SSI (4/24 SSI with povidone–iodine [16.6%] and 6/32 SSI [18.8%] with chlorhexidine; p=1.00).

Table 3 lists the clinical, pathologic, and technical factors associated with SSI. Significant risk factors were the presence of a urinary catheter preoperatively (p=0.003), vesicocutaneous fistula (p=0.008), older age (p=0.02), and a positive preoperative urine culture (p=0.03). Duration of surgery, technique (transcapsular or transvesical), renal dysfunction, and the presence of bladder stones did not show a statistically significant relation to SSI.

Discussion

In our patient cohort, the prostates generally were large, which may explain in part the large proportion of open prostatectomies. The large prostate volume probably is secondary to the great age of most of the study population, with an average in the eighth decade of life [11]. The large proportion of patients with a permanent indwelling urinary catheter prior to surgery has multiple causes. In recent years, Uruguay has faced a deficit in infrastructure and in human resources needed to cope with the growing demand for healthcare. Additionally, the country's meager public resources for healthcare were directed initially to the care of patients with malignant disease, with less funding for the surgical treatment of benign diseases such as BPH. There is a well-known relation between the time of catheterization and the presence of bacteriuria, which supposedly increases by 5%–10% per day of indwelling catheterization [12–14]. This condition explains the high number of posUC among our patients.

Most of the posUC had polymicrobial flora. This probably is attributable to the prolonged preoperative catheterization, to consecutive catheter changes, and to colonization of the urine by nosocomial organisms. Moreover, inadequate use of antibiotics and problems in urine collection may have added to the tally. We note that urine cultures were performed at the change of an indwelling bladder catheter in most patients. Identified organisms were mostly gram-negative bacteria, which had a high rate of non-susceptibility to ciprofloxacin (64%). Indiscriminate use of fluoroquinolones in recent years for the treatment of pyuria or asymptomatic bacteriuria in patients with indwelling catheters may have accounted for the elevated degree of drug resistance found in our series.

There are few publications reporting SSI in open prostatic surgery for BPH. Spanish articles report infection rates ranging from 4% to 16% in open prostatectomies, which is similar to the 18% rate encountered in our series. However, when we compare the other European studies with ours, we find a much smaller percentage of men with a preoperative urinary catheter in foreign publications [15–17].

Our main objective was to determine whether there were differences in the rate of SSI with two skin antiseptics commonly used in open prostatic surgery. The fact that there was no statistically significant difference between the infection rates using the two antiseptics is in accordance with most published studies on the subject [18,19]. However, if we take into account the reported incidence of SSI in prostatic surgery as being 2.6% in Uruguay and as much as 16% in a large Spanish study, in order to detect statistically significant differences in the rates of SSI between antiseptics, the number of patients required would be greater than 100. Thus, we cannot positively conclude that there is no difference in the efficacy of the antiseptics we studied.

An interesting finding was that the microorganisms cultured from the surgical incisions affected by SSI generally were not the normal flora of the skin, but instead bacteria already present in the urinary tract. The mechanism of infection, then, seems to be “from inside out” and not “from outside in.”

Regarding bacterial sensitivity, antibiograms showed a high sensitivity to aminoglycosides and carbapenems, but an important extent of resistance to ciprofloxacin. This antibiotic, used widely in the prophylaxis of various urologic procedures, has good prostate penetration; however, our resistance rates demonstrate clearly that we should rethink whether fluoroquinolones should still be considered first-line antibiotics for prophylaxis in our environment, at least for this type of surgery. In our series, ciprofloxacin was used until catheter withdrawal and not only as prophylaxis; even so, the SSI rate was higher than we had expected beforehand.

Analyzing potential risk factors for SSI, the presence of an indwelling bladder catheter preoperatively was the most important (p=0.003). Because the organisms cultured from the surgical incision belonged to the urinary flora, we conclude that any factor that increases the likelihood of bacteriuria will increase the risk of SSI. Indeed, perhaps we should include patients with indwelling catheters in the category of contaminated or infected operations, as 74% of the urine cultures were positive in this setting. Interestingly, current guidelines for perioperative infections in urology do not specify measures for open surgery for BPH in patients with urinary catheters, suggesting only prophylactic antibiotics [20–22]. Observing our data, we believe that in this setting, we should use therapeutic rather than prophylactic antibiotics because of the significant likelihood of postoperative SSI.

Not surprisingly, a bladder fistula also was associated with a higher risk of SSI. The drainage of contaminated urine through the incision secondary to defective closure of the bladder wall predisposes to infection. Greater age has a statistically significant effect on SSI rates, probably because of concomitant comorbidities and impairments of defense mechanisms against infections that are seen with increasing age.

The value of preoperative urine cultures has long been debated. In this respect, the preoperative treatment of patients with positive cultures can substantially reduce infectious complications, mainly in endourologic procedures [23,24]. In this series, a higher risk of SSI was shown in males with positive preoperative cultures, despite the absence of a significant correlation between the bacteria isolated from the preoperative cultures and in the cultures obtained from the infected surgical site. This fact, coupled with the large percentage of urine cultures with polymicrobial flora as a result of the high number of patients with preoperative urinary catheters, leads us to question whether, in this group of patients, there is a real benefit in carrying out routine preoperative urine cultures, and if this is done, whether the prescribed antibiotic should be chosen according to the bacteria identified initially.

Millin's transcapsular operative technique has demonstrated a benefit in terms of less postoperative bleeding and less time for a bladder catheter but showed no benefit in terms of SSI compared with the transvesical approach. Kidney dysfunction is a well-recognized contributor to immunosuppression that predisposes to infection, similar to bladder stones. None of these factors was associated with a higher risk of SSI in our series, probably because of the small number of patients with this condition.

Conclusions

Surgical site infection in open surgery for BPH is unrelated to the type of antiseptic. The most important risk factor for an SSI is the preoperative presence of a bladder catheter. We believe that timely treatment of obstructive uropathy, avoiding bladder catheterization, is an essential prophylactic measure against SSI in open BPH surgery. As isolated cultures from the surgical incision yield urinary flora, prophylactic antibiotics should be directed against them according to their presumed sensitivity pattern. In patients with a urinary catheter preoperatively, we should consider a therapeutic antibiotic, given the higher risk of SSI.