Introduction

Patients with spinal cord injury (SCI) and neurogenic bladders are at high risk for recurrent urinary tract infections. Predisposing factors include impaired bladder emptying, repeated instrumentation, vesiculoureteral reflux, and indwelling urinary catheters.1 Patients with SCI are also prone to the development of bladder and renal stones.2, 3 One link between chronic bacteriuria and the formation of stones is infection of the urinary tract with urease-producing organisms.4 Urease splits the urea present in urine to form ammonia and carbon dioxide, thus raising urinary pH and causing normally soluble ions to precipitate and form stones.5 Several bacterial species that commonly infect neurogenic bladders can split urease, including Klebsiella, Providencia, Morganella, Pseudomonas, and Proteus. As the Proteus urease has the fastest rate of urea hydrolysis,4, 6 Proteus is the organism most commonly implicated in catheter encrustation, catheter blockage, and stone formation in the urinary tracts of individuals with indwelling catheters. Antimicrobial therapy often fails to eradicate Proteus and other organisms embedded in the crevices of stones.7

Infectious complications of Proteus bacteriuria have been well documented in the general population of persons with indwelling catheters or abnormal urinary tracts.7 However, the outcomes of Proteus bacteriuria specifically in persons with SCI have not been described. Accordingly, we sought to define the relationship in the SCI population between Proteus bacteriuria and the presence of urinary complications, including stones.

Methods

Study design

A retrospective chart review of the electronic medical record was performed on all veterans with SCI who had received care at the Michael E DeBakey Veterans Affairs Medical Center over the 3 previous years, from 1 January 2003 to 2 January 2006. Patients who had undergone urinary diversion/ileostomy were excluded. Demographic information collected included age, time since injury, ethnicity, level of SCI, complete versus incomplete injury, substance abuse (alcohol, illicit drugs, and tobacco), and type of bladder management. In subjects who had more than one type of bladder management during the study period, all methods used were recorded. Thus, each subject could have several types of bladder management. Medical information collected included whether the subject was hospitalized, treated for a decubitus ulcer, treated for diabetes, developed urosepsis, had a prior history of stones, or required treatment for stones during the study period.

Microbiology

All urine cultures collected from subjects during the study period were reviewed. Any organism identified to the genus level was recorded. Proteus bacteriuria was defined as the growth of Proteus species in any urine culture during the study period. Urinalysis data were collected from all urine specimens corresponding to a documented episode of Proteus bacteriuria.

Definitions of stones

Stones were defined as either renal or bladder calculi during the 3-year time frame of the study. Multiple stones was defined as the presence of >1 stone during the time period, regardless of whether the multiple stones were present at the same time or separated in time. Treatments for stones included nephrostomy tube placement, lithotripsy, and surgical removal.

Statistical analysis

Data were compared using SAS (Release 8.2). The Fisher exact test was used for all pairwise comparisons, and logistic regression was used for multivariate analysis.

Results

Demographics

Our medical center currently serves 596 veterans with SCI. We excluded a total of 95 patients for not being seen at our medical center during the study period (71 patients) or for having had urinary diversion (24 patients). Demographic information on the remaining 501 subjects is shown in Table 1. The median age was 57.5 years (interquartile range 50–63 years) and the median time since injury was 15 years (interquartile range 5–26 years). Only 4 subjects had bladder stones, and 1 of these 4 had both bladder and renal stones. The type of bladder management from the most common to the least common was clean intermittent catheterization (N=230), spontaneous voiding (N=136), indwelling Foley catheter (N=99), external condom catheter (N=98), and suprapubic tube (N=56).

Table 1 Demographic information
Full size table

Medical conditions

The following conditions were associated with having Proteus bacteriuria: complete injury, hospitalization, decubitus ulcer, and a history of stones (Table 2). Veterans with tobacco abuse were less likely to have Proteus in the urine. Complete injury, hospitalization, decubitus ulcer, and tobacco abuse remained significant predictors in multivariate analysis (P<0.02). Only hospitalization, decubitus ulcer, and a history of stones were associated with having stones (Table 2), and in the multivariate analysis only decubitus ulcer and a history of stones remained significant predictors of stones (P<0.03). Not surprisingly, hospitalization was significantly associated with decubitus ulcer (P<0.0001). Time since SCI did not prove to be significantly related to Proteus bacteriuria, decubitus ulcer, stones, or type of bladder management.

Table 2 Association of medical conditions with Proteus bacteriuria or stones
Full size table

Microbiology

Many of the bacterial species cultured from subjects' urine were strongly associated with having Proteus bacteriuria during the study period. These organisms included Staphylococcus aureus, Enterococcus, Escherichia coli, Pseudomonas, Morganella, group B Streptococcus, Providencia, and Candida (P<0.001). Serratia was also significantly associated with Proteus bacteriuria (P<0.04). A statistically insignificant trend was noted between Acinetobacter and Proteus bacteriuria (P=0.06). Organisms not significantly associated with Proteus bacteriuria included Enterobacter, Klebsiella, Citrobacter, and Alcaligenes. Of the 71 subjects diagnosed with Proteus in their urine at some point during the study period, 19 (27%) were also diagnosed with stones (P=0.045). Other organisms associated with the presence of stones included Enterococcus, Klebsiella, Pseudomonas, Morganella, and Providencia (P<0.05).

Bladder management

Subjects using indwelling catheters, either transurethral or suprapubic, were significantly more likely to have Proteus bacteriuria (P<0.001), whereas subjects practising spontaneous voiding and clean intermittent catheterization were significantly less likely to have Proteus bacteriuria (P<0.01) (Table 3). The association between external catheter use and Proteus was not significant in either direction. Of these types of bladder management, only suprapubic catheters were significantly associated with stones (P=0.04). The association of indwelling catheters with stones nearly reached significance (P=0.08), and spontaneous voiding was nearly significantly protective against having stones (P=0.07). Indwelling Foley catheters, external catheters, and suprapubic catheters were also significantly associated with having decubitus ulcer (P<0.02).

Table 3 Association of bladder management strategy with Proteus or stones
Full size table

Analysis of subjects with stones

During the study period, 90 SCI patients were diagnosed with either bladder or renal stones. In these subjects, having Proteus bacteriuria was associated with requiring treatment for stones and having multiple stones (P<0.01). Of the 90 patients with stones, 25 (28%) required treatment, most often with lithotripsy (22/25, 88%). Six of the 90 (7%) patients with stones developed urosepsis.

Urinalysis data

When available, urinalysis data were recorded during each episode of Proteus bacteriuria. These results appear in Table 4. Means differ considerably from medians owing to the presence of outliers in each category. For example, one urine sample had 15 429 wbc/hpf, whereas another urine sample had 5100 rbc/hpf. Of note, the urinary pH was slightly basic, as expected. Also, 66% of such urines had large leukocyte esterase and 43% had a positive nitrite test.

Table 4 Urinalysis data during episodes of Proteus bacteriuria
Full size table

Conclusions

Our study confirms that bacteriuria with Proteus predicts urologic complications in persons with SCI. Proteus bacteriuria was more common in persons who had a greater degree of spinal cord impairment, were hospitalized, had decubitus ulcers, and used indwelling catheters. Proteus bacteriuria was associated with having stones, and subjects with both Proteus and stones were more likely to require treatment for stones or have multiple stones. Thus, the presence of Proteus in the bladder is not a benign condition and may be a marker for poor urinary tract health overall.

The association between Proteus bacteriuria and other nosocomial urinary pathogens was not surprising. Several of the less common nosocomial pathogens were not significantly associated with Proteus, such as Alcaligenes, but the very few infections caused by these organisms during our study were unlikely to achieve statistical significance. The lack of association of Enterobacter, Klebsiella, and Citrobacter with Proteus cannot be attributed to low numbers of infections, as these organisms are relatively common. Perhaps these organisms, unlike Proteus, have an equal propensity to infect healthy and diseased urinary tracts. Of the organisms found to be associated with stones, one of these, Enterococcus, does not make urease. We suspect that the strong association of this nosocomial pathogen with Proteus bacteriuria accounts for its association with stones.

Proteus has previously been reported as a common cause of bacteriuria in chronically catheterized residents of long-term care facilities.8 Furthermore, the relationship between Proteus bacteriuria and urinary catheter encrustation and blockage is well documented.9, 10 Proteus bacteriuria has likewise been documented in persons with SCI,1, 11, 12 but not specifically in association with stones. Two retrospective reports published two decades ago of almost 6000 patients with SCI found that bladder calculi were associated with the complete injuries and with Klebsiella bacteriuria.2, 3 In these studies, Proteus bacteriuria was nearly significantly associated with development of bladder stones within 2 years of hospital discharge for initial injury (P=0.06). Renal stones were likewise associated with neurologically complete lesions, and having an indwelling catheter showed a trend toward an association with stones. Thus, our results confirm and expand upon these earlier studies. Our Veterans Affairs Medical Center serves a large and diverse population of SCI patients, so we suspect that our results would also be applicable to other predominantly male SCI populations. As only 2% of our SCI subjects were female, we cannot presume that our results can be generalized to women with SCI.

Our study has several limitations. Like all retrospective studies, we relied on data reported in the medical record. We looked at all urine culture results on SCI subjects during the study period, but we do not know the reason why urine was cultured at a given time. The computerized medical record did not adequately capture data about catheter blockage; hence we were unable to study that outcome. Dates at which catheter management strategies changed were likewise not recorded; therefore, we cannot report a temporal relationship between bladder management strategies and Proteus bacteriuria or stones. We also suspect that the data on substance abuse were unreliable, both because physicians may not always record this information and because patients may not disclose the extent of their substance use. The apparent protective effect of tobacco use on Proteus bladder infection could be attributed to the fact that subjects with lesser degrees of neurological impairment, and thus less likely to have Proteus bacteriuria, had greater opportunities to acquire and smoke cigarettes. As with all observational studies, we can demonstrate associations of various outcomes with Proteus bacteriuria, but we cannot prove causality.

Proteus is extremely difficult to eradicate from the catheterized urinary tract. In a study that investigated the molecular epidemiology of catheter-associated bladder infections with Proteus, the same strain of Proteus persisted in one patient over 121 days despite eight catheter changes, a 20-day period without catheterization, and an 8-day course of antibiotics.9 In our own study of bladder colonization with benign E. coli in patients with SCI and indwelling urinary catheters, we were unable to eradicate Proteus bacteriuria in four of four subjects despite a week of targeted antibiotics accompanied by catheter change during the course of antibiotics.13 Two of these subjects developed catheter encrustation and blockage within <3 weeks. As we have confirmed that the same SCI patients who are infected with Proteus are likely to experience complications such as stones and catheter encrustation, a more effective approach to clearing this organism from the urinary tract is needed. Urinary catheters inflated with triclosan prevented catheter colonization and biofilm formation by Proteus in vitro.14 Microbicins, or natural antibiotics produced by bacteria, may offer another defensive strategy against Proteus, but as yet we have only used microbicins against E. coli in vitro.15 The data that we have collected on the conditions associated with Proteus bacteriuria in SCI patients indicate that bladder colonization with Proteus is harmful, and we anticipate using this information to design future clinical studies targeting this pathogen.