Urinary Tract Infection (UTI) is a well recognized cause of morbidity in spinal cord injury (SCI) patients with an approximate 20% prevalence of recurrent UTI (rUTI) in the SCI population [1,2]. Despite the significant effect of rUTI, the causative physiologic and patient characteristics are not well described. Demographic variables such as gender, age, and race and urodynamic study (UDS) findings such as detrusor overactivity (DO, bladder compliance and capacity have inconsistently been associated with rUTI. We therefore sought to identify patient demographic and urodynamic characteristics associated with rUTI in SCI patients who manage their bladders with clean intermittent catheterization (CIC).

The records of 136 SCI patients who perform CIC for bladder management were retrospectively reviewed. We excluded non-SCI neurogenic bladder (NGB), patients not using CIC, and patients with incomplete records. Demographic variables, comorbidities, and urine culture information corresponding to the 12 months prior to each patient’s initial urology presentation were recorded. Data from each patient’s initial video urodynamics (VUDS) evaluation, including variables such as cystometric capacity and bladder compliance/storage pressures were recorded. Recurrent UTI was defined as ≥ UTIs in a year. 
Our study consisted of three main steps. We first divided patients into two groups, those with and those without self-reported rUTI independent of urine culture results. Normal distribution was not assumed. We subsequently performed Mann-Whitney U tests and Fisher’s exact tests to assess for differences between these two populations. Second, we separated individuals into cohorts with or without three or more culture-proven UTI and subsequently repeated our analysis. Last, we performed multivariable logistic regression to assess the association between our variables and the risk of having three or more culture-confirmed UTIs. Statistical analysis was performed using a combination of IBM SPSS 24 and Stata 13.0. Significance was considered at alpha = 0.05. Demographic and clinical factors were reported as mean ± standard deviation. Urodynamic variables were found to be more subject to outlier values and were consequently reported as median values with associated ranges.

136 new SCI patients who used CIC and had adequate medical records including VUDS were included in the initial analysis. 124 patients had adequate urine culture information and were included in culture-specific analysis. The initial analysis of 136 patients compared baseline characteristics of those who self-reported rUTI in the year prior to evaluation [n=58 (42.6%)] and those who did not [n=78 (57.4%)]. No significant differences between the two groups were noted, except patients reporting rUTI had a higher mean BMI (26.4 vs 24.8, p=0.048). Only 24/46 patients (52%) who self reported rUTI had supportive urine culture results.
The second analysis compared patients with [n=30 (24.2%)] and without [n=94 (75.8%)] three or more culture-proven infections in the year leading up to urology referral. Race designated as African-American or Other was associated with a higher prevalence of rUTI (p=0.033). In addition, patients with rUTI were closer in time to their SCI compared to those without rUTI (mean 3.1 vs 5.6, p=0.018). While BMI was higher in the rUTI patients, it did not reach statistical significance (26.7 vs 25.3, p=0.336). Similarly, younger age (34.0 vs 38.4 years, p=0.142) and female preponderance (23.3% vs 10.6%, p=0.123) were more common with rUTI but failed to reach statistical significance. There were no significant differences in UDS variables between cohorts.
Lastly, multivariable logistic regression was performed to determine risk factors for rUTI. Notably, female gender (OR 4.96, p=0.011), years since SCI (OR 0.91, p=0.036), and African American race (OR 5.16, p=0.002) were the only statistically significant factors identified.
Other variables included in the regression model, such as age at presentation, BMI, SCI level, history of decubitus ulcers or diabetes, presence of bladder/renal calculi, presence of DO, volume at first DO, maximum detrusor pressure with DO, end-fill detrusor pressure, bladder compliance, and cystometric capacity were not significant.

Patients with self reported rUTI had a significantly higher mean BMI than those without rUTI. While a higher BMI has previously been positively correlated with UTI risk, this was not the case when our culture results were factored into the calculation. Furthermore, the fact that only 24/46 patients (52%) who self reported rUTI in fact had supportive urine culture results emphasizes the importance of obtaining urine cultures and upholding specified definitions/guidelines for UTI in the clinical management of SCI patients.  
In our second analysis involving patients with urine culture results, race was a significant risk factor for rUTI with African Americans and other non-Caucasian races more likely to have had rUTI (OR 5.16 and 3.91, respectively). However, only African American race met our set significance level. 
Female gender was found to be an adjusted determinant of rUTI in our patient population, however the role of gender can be hard to interpret as males disproportionately compose the bulk of the SCI population: in our study, only 20 of 136 (14.7%) patients were female. Gender has been inconsistently associated with UTI in the SCI literature.
We also noted that patients with rUTI were younger (34.0 years vs 38.4 years), though this did not meet statistical significance. Rather than age per se, it may be that experience dealing with NGB issues is the key. Our patient variables were collected as they were first evaluated by our department and as such many of the patients were relatively novice regarding the management of urologic sequelae of their injuries. We found that time since SCI was a significant protective factor against rUTI and thus it is possible that the experience individuals accumulate when managing their condition may provide benefit. However, the exact factors which might improve with such experience and provide benefit remain unclear. 
We did not identify any significant impact of UDS findings on the risk for rUTI. While others have reported that UDS characteristics such as DO and elevated storage pressures may be risk factors for UTI, this has not been universally noted. Our UDS were routinely done in the supine position (for ease) at a fill rate of 30 mL/min with a low threshold to reduce the filling rate to 10 mL/min when the storage pressure appeared to increase (eg detrusor overactivity (DO) or impaired compliance). Furthermore, approximately 50% of our patients were on an overactive bladder medication. As such, our UDS results may be different than what others have found in this patient population. 
Several limitation are noteworthy. The majority of urine cultures in our patients were ordered by physicians outside of our department. Consequently, we were unable to confirm the symptoms present at the time of each culture as to provide stringent definition of infection within the study. We also cannot account for any cultures/infections that might not have been documented in our electronic medical record (e.g. patient went elsewhere) which may underestimate the number of UTIs.

We found factors such as female gender, African American race, and more recent SCI were risk factors for rUTI in SCI patients using CIC for bladder management. We did not identify any urodynamic parameters that predicted rUTI.

Kavanagh A, Baverstock R, Campeau L, et al. Canadian Urological Association guideline: Diagnosis, management, and surveillance of neurogenic lower urinary tract dysfunction - Full text. Can Urol Assoc J. 2019;13(6):E157-E176. doi:10.5489/cuaj.5912
Siroky MB. Pathogenesis of bacteriuria and infection in the spinal cord injured patient. Am J Med. 2002;113 Suppl 1A:67S-79S. doi:10.1016/s0002-9343(02)01061-6