In an initial step, data were analyzed with descriptive statistics and Pearson correlation was used to investigate simple correlations between different variables. Appropriate linearity tests and lack of fit tests were performed to ensure that these relationships were indeed linear and adequate. Secondly, in a more in depth analysis the relationship between PSA levels and the different potential covariates was analyzed with a multiple regression model. Since the histogram of PSA is severely skewed due to the prevalence of lower PSA values in a normal population (Figure 2), PSA values were transformed with a natural log transformation, resulting in a more realistic normality assumption of the log PSA than the original PSA scale. Selection of the variables included in the model for log PSA was based on the AIC criterion and the adjusted r-square. With respect to the variable active inflammation, it might not seem realistic from a clinical point of view to assume that the difference between any consecutive levels of the active inflammation was the same or the increase was linear in nature. For this reason, the choice was made to treat inflammation as a categorical rather than a continuous variable. Furthermore this choice resulted in a better fit of the model.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Histogram for prostate-specific antigen (A) and log prostate-specific antigen (B) with a normal density curve superimposed. PSA: Prostate-specific antigen.

The patients were between 45 years and 90 years old (mean age: 69.1 ± 8.6 years) (Table 1) and had an indication for TURP. Mean preoperative PSA value was 4.7 ± 5.4 ng/mL (0.2 to 32.5 ng/mL). Mean IPSS was 15.7 ± 6.9 (0 to 32). Mean peak flow was 10.7 ± 6.5 mL/s (1 to 47 mL/s). Post residual volume ranged from 0 to 890 mL with a mean value of 83.8 ± 130.3 mL. When indicated (n = 154), pressure flow analysis was executed and showed a mean value of PdetQ_max of 96.3 ± 34.4 cmH₂O (10 to 220 cmH₂O). The mean resected prostate weight was 39.4 ± 27.3 g (3 to 189 g). Mean operating time was 31.1 ± 11.9 min (10 to 110 min).

We investigated the relationship between log PSA and several parameters, including IPSS, maximum uroflow, post voided residual volume, pressure flowmetry parameter PdetQ_max and the weight of the resected prostate tissue (Table 2). A significant correlation between log PSA and the resected prostate weight was found (r = 0.54; P < 0.001). Furthermore, log PSA was correlated with PdetQ_max (r = 0.17; P = 0.032). A negative correlation was encountered between log PSA and IPSS (r = -0.14; P = 0.04). No correlation was observed between log PSA and peak flow on uroflowmetry (r = 0.00; P = 0.99). The correlation between log PSA and the post residual volume (r = 0.08; P = 0.28) was not significant. Additionally, we did not find any correlation between IPSS and PdetQ_max (r = 0.051; P = 0.53). Last but not least, a significant correlation between the weight of the resected tissue and PdetQ_max (r = 0.39; P < 0.001) was observed. These relationships were confirmed in a multiple regression model (Table 3).

We also investigated the relationship between inflammation (active and chronic) and log PSA, IPSS, age and the resected prostate tissue weight (Tables 4 and 5). A significant correlation between log PSA and active (r = 0.21; P < 0.0001) and chronic (r = 0.19; P = 0.005) inflammation was observed. Age was related to active (r = 0.24; P < 0.0001) but not to chronic (r = 0.09; P = 0.08) inflammation. Similarly, the weight of the resected prostate tissue was related to active inflammation (r = 0.13; P = 0.011) but not to chronic inflammation (r = 0.1; P = 0.34). However, IPSS was not correlated with active inflammation (r = 0.03; P = 0.6) or chronic inflammation (r = - 0.03; P = 0.91). Additionally, categorization of IPSS into 3 categories did not result in any association with active and chronic inflammations (Table 6).

We also evaluated the influence of pre-operatively placed suprapubic catheters, trans-urethral catheters or previous prostate puncture biopsies and TURP on the relationship with active and chronic inflammation (Table 7). Only a weak negative correlation was found (r = -0.16; P = 0.003) between active inflammation and the presence of a suprapubic catheter. No significant correlation was observed with the other variables (Table 7). However, for some categories of degree of inflammation no data were obtained, complicating comparisons.

Multiple papers have been published indicating the relationship between PSA levels and BOO. In patients with clinical BPH, PSA levels are shown to be positively correlated with the five year cumulative risk of invasive BPH treatment[15]. Additionally, in men with BPH, prostate volume and PSA have been evaluated as predictors of acute urinary retention[16,17]. Consequently, elevated PSA levels are predictive for the need of BPH-related surgery. PSA has also been shown to be a strong predictor of future prostate growth[18], which is related to a higher risk of acute urinary retention and, subsequently, to the need for BPH-related surgery. A broader approach for the use of PSA seems justified. In a multicenter study in men suffering from LUTS, a correlation between PSA and the category of BOO was observed[19], Above a PSA-level of 4 ng/mL, mild or definite BOO was observed (in 89% of the cases) and below 2 ng/mL, the chances of a patient not suffering from BOO were one in three[19]. In previous papers, we have shown that BOO can be expected in a very particular group of patients with elevated and/or rising PSA, (multiple) negative multisided prostate biopsies and minor LUTS (mild + moderate IPSS)[20,21]. In a retrospective analysis of 82 consecutive patients, 95.9% were clearly obstructed (PdetQ_max≥ 40 cm H₂O)[18]. Similar results were obtained in a prospective analysis (n = 33), with a mean PdetQ_max of 80.3 cm H₂O[21]. A positive correlation between PSA velocity and PdetQ_max was found (r = 0.5014; P = 0.006)[22]. As already mentioned, rising PSA levels can also be observed in patients with prostate cancer. Notwithstanding the presence of a tumor in these patients, the elevated PSA levels can also be influenced by BOO, implying that PSA should be included in the pre-treatment workup and the post-therapy evaluation of these patients. Therefore, PSA should be considered an additional indicator in the BOO decision tree.

We found some interesting correlations in this prospective analysis of 222 consecutive patients undergoing TURP for bothersome LUTS. A significant correlation was shown between log PSA and PdetQ_max. In line with the literature[22], log PSA was also correlated with prostate volume (the resected prostate weight) (r = 0.54; P < 0.001). Interestingly, a negative correlation between IPSS and log PSA was observed (r = -0.14; P = 0.04), confirming the findings of our previous studies. This negative correlation could partly be explained by the correlation between high PSA levels and high detrusor pressures due to an increased compensation level of the detrusor for the urethral resistance, resulting in high flow rates and less symptoms.

On the other hand, no statistically significant correlation was observed between log PSA and post voiding residual volume, and between IPSS and the weight of the resected tissue. This might mean that the response of the detrusor to obstruction is more related to the symptoms than the degree of obstruction as such. Last but not least, a highly significant correlation between the weight of the resected tissue and PdetQ_max (r = 0.39; P < 0.001) was observed, implying that bigger prostates are more frequently associated with obstruction and therefore are more prone to BPH-related surgery.

The role of chronic prostate inflammation in PSA elevation in asymptomatic men is still unclear. Many papers have covered this subject with various and sometimes contradictory outcomes and conclusions. In a prospective analysis of a small group of asymptomatic patients (n = 51) who underwent prostate puncture biopsies, a statistically significant correlation was found between the inflammatory process and the PSA values (P = 0.02)[23]. Similarly, in a retrospective analysis of 238 men, a correlation between PSA and inflammation was observed[24]. In an analysis of the prostate puncture biopsy of 80 asymptomatic patients, the extent of inflammation was positively correlated with total PSA levels (P < 0.001)[25]. In all these studies, histological examination was performed on prostate puncture biopsies in asymptomatic men, limiting the results obtained.

In contrast to the previously mentioned reports, multiple papers claimed the opposite and concluded that there is no correlation between PSA and inflammation. In a retrospective analysis of cancer negative prostate biopsies (n = 233), it was concluded that the degree of chronic inflammation did not correlate with PSA levels[26].

Chronic prostatitis was encountered in 68.3% of 284 patients with negative prostate puncture biopsies, while active prostatitis was observed in 8.4% of patients[27]. However, inflammation did not correlate with total PSA levels. In a prostate puncture biopsy driven evaluation (n = 49), the presence of inflammation did not correlate statistically with the PSA levels[28]. Finally, Nickel et al[29] studied a cohort of 80 patients without a history of prostatitis who underwent TURP. After histological examination on the resected tissue, prostatic inflammation was found to be extremely common. On the other hand, no correlation was observed between inflammation and PSA levels.

We performed a prospective analysis in 222 patients undergoing TURP according to the EAU guidelines[30] but without proven infection, known prostate cancer and a history of chronic prostatitis. We made a histological distinction between active and chronic inflammation, each with 4 subcategories. Additionally, since we used TURP tissue and not prostate puncture biopsy derived tissue, we were able to evaluate a substantial prostatic tissue volume. Our results showed that most patients have some degree of inflammation. Especially in the active inflammation group, a correlation between log PSA and inflammatory parameters was observed (r = 0.21; P < 0.0001). However, we did not encounter a statistically relevant correlation between symptoms (IPSS) and inflammation. The correlation between age and active inflammation (r = 0.24; P < 0.0001) is interesting to note, which was not the case for the chronic inflammation group (r = 0.09; P = 0.08). A possible explanation for this finding could be the accumulation of debris in the prostate during the lifetime, resulting in more frequent active inflammations. This finding may also be the explanation for the correlation between prostate weight (TURP) and active (r = 0.13; P = 0.011) and chronic (r = 0.1; P = 0.34) inflammation.

In conclusion, elevated and/or rising PSA levels with regard to underlying prostate cancer and prostatic inflammations are well known. In this paper, we have shown that PSA could also be indicative of BOO, not only with respect to observational data but also with regard to statistically relevant correlations between PSA and PdetQ_max. Therefore, PSA levels should be taken into account and could even be used as a biomarker when a treatment strategy is determined or executed for patients suffering from clinical BPH. Additionally, the results obtained in this study indicate that inflammation can be correlated with PSA levels. This implies that elevated PSA levels should also be considered as predictive for prostate inflammation as well as for prostate cancer.

Since its introduction, prostate-specific antigen (PSA) has played a key role in prostatic evaluation. However, it is not a cancer specific parameter but merely organ specific. Consequently, elevated PSA levels are a challenging problem for urologists in assessing or excluding potential life-threatening prostate cancer. A better understanding of other possible mechanisms causing PSA elevation is required. Therefore, the aim of this study was to investigate the relationship between PSA and the degree of prostate inflammation and to investigate whether PSA levels can be used as a biomarker for bladder outlet obstruction (BOO).

The cause of elevated PSA serum levels is not always prostate cancer. In the area of interpreting PSA elevation, the research hotspot is to have an insight in to other mechanisms that can be responsible for a PSA elevation, apart from prostate cancer.

In this study the authors showed a correlation between BOO (PdetQ_max) and PSA (logarithmic). Furthermore, the authors demonstrated a weak correlation between PSA (logarithmic) and active as well as chronic prostatic inflammation.

PSA could also be indicative of BOO, not only with respect to observational data but also with regard to statistically relevant correlations between PSA and PdetQ_max. Therefore, PSA levels should also be taken into account and could even be used as a biomarker when a treatment strategy is determined or executed for patients suffering from clinical benign prostate hyperplasia (BPH). Additionally, the results obtained in this study indicate that inflammation can be correlated with PSA levels. This implies that elevated PSA levels should also be considered as predictive for prostate inflammation as well as for prostate cancer.

BOO: a blockage at the base of the bladder that reduces or prevents the flow of urine into the urethra, the tube that carries urine out of the body. This condition is most common in aging men. It is often caused by BPH. As a man ages, his chance of developing these diseases increases dramatically; Prostate inflammation: inflammation of the prostate gland. There are four types of prostatitis: acute bacterial prostatitis, chronic bacterial prostatitis, chronic prostatitis without infection, asymptomatic inflammatory prostatitis; PSA (prostate-specific antigen): a protein manufactured exclusively by the prostate gland. PSA is produced for the ejaculate where it liquefies the semen and allows sperm cells to swim freely. Elevated levels of PSA in blood serum are associated with benign prostatic hyperplasia and prostate cancer.

This prospective analysis investigates the association of high PSA levels and bladder outflow obstruction and prostatic inflammation. Its results, as described in the discussion, add to the literature and the controversy that exists in these issues among other studies.