Hypothesis / aims of study
The urinary microbiome has only recently been described, and its potential relations with health and disease are still under investigation. It is reasonable to assume that functional abnormalities of the lower urinary tract, such as detrusor overactivity or bladder outlet obstruction, could lead to significant changes in the urinary microbiome. Our aim was to conduct a pilot study of a selected group of men with lower urinary tract symptoms and correlate their clinical and urodynamic findings with the ability to detect and characterize the bladder microbiome.
Study design, materials and methods
We prospectively selected male patients over 50 years of age with lower urinary tract symptoms who would be submitted to urodynamic study due to investigation of LUTS (Lower Urinary Tract Symptoms). We collected prostatic specific antigen (PSA) value, post void residual volume (PVR) and prostate volume measured by ultrasound. They also self-completed the IPSS (International Prostate Symptom score) and Overactive Bladder V-8 questionnaires (OAB V8) as a way of quantify their symptoms. They were submitted to a conventional urodynamic study as recommended by the ICS (International Continence society) and then classified as having or not detrusor overactivity and bladder outlet obstruction. Before the urodynamic study, a urine sample of 50 ml was collected by aseptic bladder catheterization and subjected to DNA extraction and quantification. To perform the sequencing, we amplified the V4 hypervariable region of the 16s rRNA gene through Polymerase Chain Reaction (PCR) and subsequent sequencing in the Illumina® MiSeq system. The sequences were clustered into operational taxonomic units (OTUs) based on 97% similarity and were identified against the SILVA v132 reference database. Then, we evaluated for correlation between OTUs richness and number of reads with patients’ symptom scores, age, PVR, prostate volume, and urodynamic diagnosis.
Exclusion criteria for this study are the presence of any other abnormality in the urodynamic study than bladder outlet obstruction or detrusor overactivity, presence of urethral stenosis, diabetes, obesity (IMC > 30), smoking, alcoholism, previous pelvic surgery, any known neurological disease, use of diuretics, use of any medications that may have an impact on the lower urinary tract function, urinary tract infections in the last 12 months, use of antibiotics for any reason in the last 6 months, acute or chronic prostatitis and chronic pelvic pain.
Based on the sample size of published male urinary bacterial microbiome studies (n = 10 per group in Nelson et al, 2010), a total of 12 samples was felt to be practical for fast recruitment considering an initial study.
We collected urine samples from 49 patients. Due to amount and/or quality of DNA, 16S rRNA amplification was possible to be done in 13 samples (26.5%), but only 12 (24.5%) yielded results after sequencing. These are described in Table 1 and Figure 1. We did not find any correlation of successful sequencing with the parameters of interest like age (63.17±10.63 for positive sequencing vs 64.87±8.50 for negative sequencing, p=0.574), PVR (49 (5-350) for positive sequencing vs 45 (6-925) for negative sequencing, p=0.723), IPSS (18.58±6.68 for positive sequencing vs 17.19±7.26 for negative sequencing, p=0.560), OAB V-8 (17.67±7.43 for positive sequencing vs 15.32±7.54 for negative sequencing, p=0,351) , prostate volume (42.5±21.5 for positive sequencing vs 38.5 ±16.9 for negative sequencing p=0.868), PSA (2.02±1.97 for positive sequencing vs 2.11±1.49 for negative sequencing, p=0.508), presence of detrusor overactivity (7/12 for positive sequencing vs 14/37 for negative sequencing. P=0.362) or bladder outlet obstruction (8/12 for positive sequencing vs 23/37 for negative sequencing. P=0.950). All of our patients had an IPSS score of 9 or more, denoting at least moderate LUTS.
About the caractherization of the bladder urinary micobiome, the predominant phyla were Proteobacteria (10/12 samples), Firmicutes (9/12 samples) and Actinobacteria (8/12 samples). The prevalent families were Enterobacteriaceae (10/12 samples), Burkholderiaceae (7/12 samples), Sphingomonadaceae (7/12 samples). The Staphylococcus genus was present in 58.3% samples. We found 7/12 samples (58,3%) with a dominant genus (>50% of the reads belonging to the same genus), thus these samples were considered as having a low diversity. The Shannon index is H'=0,918 (0,003 - 2,436) confirming the low diversity for this group of samples. Again, we did not find any correlation of bacterial microbiome composition or diversity with detrusor overactivity, bladder outlet obstruction or any other evaluated parameter (p>0.05).
Interpretation of results
We had a low rate of bacterial identification in urine samples (24.5%), although it was comparable to that of a previous study with a similar samples collection method (aseptic bladder catheterization). Bladder asseptic catheterization is described as the best method to evaluate the bladder urinary microbiome.
The bladder is a low biomass environment and is difficult to identify the microbiome, eventhough we believe that we have a microbiome in all patients. Strategies to improve this detection rate should be apply in the main study.
This is the first study that correlates urodynamic findings and bladder microbiome. Although we have no correlation in this pilot study, this issue should be better evaluated with a more robust study that we are carry on nowadays.
The main finding is the low diversity in this group of patients with LUTS. Low diversity, with a dominant genus, could increase the chance to develop a symptomatic urinary tract infection or chronic inflammation as described in patients with moderate and severe LUTS. Notably, the range of the Shannon index (measure of the microbial diversity and expressed as H') is high, with two samples with H'>2, that is associated with good diversity. It could be a reason why some patients with LUTS have recurrent urinary tract infections and others don´t have it. But we need a more robust study with different design to confirm this association.
As weak points, we have a small sample (this is a pilot study) and we did not have patients with mild symptoms.
We were able to detect and characterize the bladder urinary microbiome in only 24.5% of the patients through 16S rRNA sequencing in this inital study. The detection rate has no association with age, PSA, symptoms severity, prostate volume, detrusor overactivity or bladder outlet obstruction. We found that the bladder microbiome of most patients with moderate and severe LUTS is characterized by low bacterial diversity and this finding could be linked to some clinical caractheristics of this group of patients, like recurrent urinary tract infections or chronic inflammation (prostatitis, benign prostatic hyperplasia). Further robust studies with a larger number of patients and additional urodynamic parameters are needed to obtain definitive evidence of this pattern. Additionally, this pilot study reveals the need to seek strategies to increase the detection rates of the bladder microbiome in order to be more representative of the entire population, and should ideally include patients with mild symptoms.