Prostate inflammation not a predictor of cancer

January 1, 2007

Chicago-Initial analyses from a retrospective study involving almost 600 men undergoing prostate needle biopsy show no association between inflammation on initial or subsequent biopsy and prostate cancer. However, the investigators from Harvard Medical School are withholding any conclusions while they continue their research.

Chicago-Initial analyses from a retrospective study involving almost 600 men undergoing prostate needle biopsy show no association between inflammation on initial or subsequent biopsy and prostate cancer. However, the investigators from Harvard Medical School are withholding any conclusions while they continue their research.

"With the population of men at risk for prostate cancer increasing as longevity is increasing, it is important to identify potentially modifiable risk factors for prostate cancer," first author William W. Bassett, MD, said at the American College of Surgeons clinical congress here.

Dr. Bassett observed that chronic inflammation established via an infectious agent is recognized as a precursor to stomach and hepatic cancer. However, the role of chronic inflammation in prostate cancer carcinogenesis is less clear, as previous studies have yielded mixed results.

"Many of the reported investigations had a retrospective cohort design, and so are subject to recall, follow-up, and misclassification bias. In our study design that analyzes outcome data based on pathological samples, we hoped to circumvent some of those potential confounders," he said.

The preliminary analyses were based on 591 men who underwent needle biopsy between 1990 and 2004 at Brigham and Women's Hospital. All men had a PSA >4.0 ng/dL and/or abnormal digital rectal exam.

At initial biopsy, inflammation was present in 55 men (13%), and prostate cancer was diagnosed in 159 men (27%). Those 159 men comprised two men (<1%) with inflammation and 157 men (26%) without inflammation.

Within the group, 244 men had a subsequent biopsy over a mean follow-up of about 2.5 years. Of these men, 31 (13%) had inflammation at initial biopsy. Time to event analysis was conducted using Kaplan-Meier survival techniques to determine whether inflammation at initial biopsy predicted subsequent cancer diagnosis, but no significant relationship was found.

"The p-value for the log rank test in a Kaplan-Meier survivor analysis was 0.650, which does not even represent marginal statistical significance," Dr. Bassett said. "However, it is possible that a study with greater power may have shown separation of the curves between the inflamed and non-inflamed groups or a difference in time to development of prostate cancer. As our study accrues more patients with inflammation, it will gain increased power that will allow us to strengthen our conclusions."

Further analyses of the association between inflammation and prostate cancer on follow-up biopsy were conducted using Cox models. The results were similar to the Kaplan-Meier analysis in identifying that inflammation was not predictive of time to cancer development in the 244 patients.

Univariate and multivariate analyses also were conducted to investigate potential associations between various clinical and pathologic factors and prostate inflammation. The covariates included digital rectal exam, PSA, PSA density, prostate volume, and age. With the men categorized according to whether inflammation was present or absent on initial prostate biopsy, the only feature that differed significantly between the two groups was prostate volume, which was significantly greater in men with an inflamed versus non-inflamed prostate: 63 versus 45 cc, respectively.

In univariate analysis considering data from the first four biopsies, men with an inflamed prostate had both a significantly higher PSA and prostate volume. Prostate volume and PSA were jointly significant in a multivariate model including both covariates.

"These results are consistent with previous reports showing a positive association between inflammation on needle biopsy and serum PSA and confirm that prostate volume correlates significantly with PSA. We believe that relationship underlies the association between prostate volume and inflammation in our study," Dr. Bassett said.