Association of Nonsteroidal Anti-Inflammatory Drugs, Prostate Specific Antigen and Prostate Volume

Article Outline

• Abstract
• Materials and Methods
  • Study Population
  • Data Collection
  • Data Analysis
• Results
• Discussion
• Conclusions
• References
• Copyright

Purpose

Nonsteroidal anti-inflammatory drugs such as aspirin prevent cardiovascular disease and several prior studies suggest that nonsteroidal anti-inflammatory drugs also decrease prostate inflammation and prostate cancer risk. We investigated the association between nonsteroidal anti-inflammatory drug use, prostate specific antigen and prostate volume, hypothesizing that there would be lower prostate specific antigen and prostate volume with nonsteroidal anti-inflammatory drug use.

Materials and Methods

The Nashville Men's Health Study uses a multicenter, rapid recruitment protocol to collect clinical, biological, behavioral and body measurement data on 1,277 men older than 40 years who are scheduled for diagnostic prostate biopsy. Nonsteroidal anti-inflammatory drug use was ascertained by survey and clinical interview. Medical charts were reviewed to ascertain current prostate specific antigen, prostate volume and clinical diagnoses following biopsy.

Results

Approximately 46% of patients reported receiving nonsteroidal anti-inflammatory drugs, primarily aspirin (37%). After adjusting for age, race and other factors prostate volume was similar between aspirin users and nonusers (47.6 vs 46.0 ml, p = 0.16). In contrast, prostate specific antigen was significantly lower in aspirin users (7.3 vs 8.0 ng/ml, p = 0.01). The association between prostate specific antigen and aspirin was significant in men with latent prostate cancer (6.1 vs 7.3 ng/ml, p <0.01), marginal in patients with high grade prostatic intraepithelial neoplasia (5.0 vs 5.9 ng/ml, p = 0.09) and nonsignificant in those with a negative biopsy (5.6 vs 5.7 ng/ml, p = 0.64). The strongest prostate specific antigen-aspirin association was in men with cancer and a prostate volume of 60 ml or more (7.3 vs 12.7 ng/ml, p <0.01).

Conclusions

Prostate specific antigen was significantly lower in aspirin users with latent cancer. Prostate volume was not associated with nonsteroidal anti-inflammatory drug use. Results suggest that aspirin may affect prostate cancer detection, suggesting a potential detection bias to address in future studies of nonsteroidal anti-inflammatory drugs and prostate cancer prevention.

Key Words: prostate, prostatic neoplasms, prostate-specific antigen, anti-inflammatory agents, non-steroidal, aspirin

Abbreviations and Acronyms: BMI, body mass index, BPH, benign prostatic hyperplasia, COX, cyclooxygenase, CVD, cardiovascular disease, NSAID, nonsteroidal anti-inflammatory drug, PG, prostaglandin, PIN, high grade prostatic intraepithelial neoplasia, PSA, prostate specific antigen, WHR, waist-to-hip ratio

Prostate histology consistent with prostate inflammation is common. Inflammatory infiltrates may be found in 43% of BPH cases¹ and chronic inflammation is associated with lower urinary tract symptoms and a larger prostate volume.¹, ² Free radicals induce and derive from inflammatory and immune responses, and they trigger PG synthesis via the COX pathway.³ The COX-1 and inducible COX-2 isoforms are rate limiting in the conversion of arachidonic acid to PGE2 and other PGs that are known to increase cell proliferation in prostate and other tissues.³ Several endogenous or exogenous factors may induce prostate tissue inflammation and PG synthesis, such as infection, trauma due to urine reflux and prostatic calculi, sex hormone imbalance and obesity. To varying degrees these factors are also associated with prostate cancer risk and progression.⁴ Similarly COX-2 expression and PGE2 levels are higher in prostate cancer and high grade PIN tissue than in benign tissue,⁵, ⁶ and genetic variants in COX-2 have been associated with prostate cancer risk.⁷

To the extent that inflammation generates an environment conducive to prostate tumor growth NSAIDs may decrease the prostate cancer risk. NSAIDs such as aspirin inhibit COX activity and arachidonic acid metabolism. A recent meta-analysis estimated that NSAID use is associated with a 10% to 30% decrease in prostate cancer risk.⁸ Subsequent investigations using the Baltimore Longitudinal Study of Aging,⁹ the American Cancer Society Cancer Prevention Study II¹⁰ and pharmaceutical claim administrative data¹¹ also indicate protective associations between NSAID use and prostate cancer risk. However, several prior studies showed no association, little dose-response effect, or inconsistent associations with disease stage or grade.⁸, ⁹, ¹⁰

Inconsistencies such as these across studies may derive from any relationship between NSAID use and the clinical methods used to detect early stage prostate cancer in a population. Any decrease in prostate inflammation or infiltration into the glandular epithelium with NSAID use may decrease PSA below the level of clinical suspicion. NSAID administration decreased PSA velocity in 2 small trials of patients with biochemical failure following prostate cancer treatment.¹², ¹³ In contrast, PSA was slightly but significantly higher with nonaspirin NSAID use in healthy men older than 70 years participating in the Baltimore Longitudinal Study of Aging.⁹ In addition, decreased prostate inflammation with NSAID use may also decrease prostate volume. Aside from the potential benefits of decreasing the severity of lower urinary tract symptoms a volume decrease may improve PSA as a marker of cancer or increase the efficacy of sampling cancer at prostate biopsy.¹⁴

We investigated the association of NSAID use, PSA and prostate volume, hypothesizing that NSAID users have lower PSA and prostate volume. Results may have implications for prostate cancer detection, clinical decision making following cancer diagnosis and research into prostate cancer chemoprevention.

Back to Article Outline

Materials and Methods 

Study Population 

All participants provided written informed consent with guarantees of confidentiality before data collection, in accordance with the Vanderbilt University institutional review board. Men scheduled for diagnostic prostate biopsy between 2002 and 2008 at Vanderbilt University Medical Center, Tennessee Valley Veteran's Administration Hospital, Nashville, Tennessee or Urology Associates, a private urology clinic, were approached for recruitment. Eligible participants were 40 years old or older and had no prior prostate cancer diagnosis. Approximately 95% of eligible men approached for recruitment agreed to participate and the study population included 1,372 eligible consenting patients.

Data Collection 

Measures of body size and weight were collected by a trained research staff member at recruitment using standardized protocols. Participants wore a hospital gown or other light clothing and did not wear shoes. Chart review included age, race, PSA history and prostate needle biopsy result, including cancer, high grade PIN, negative, or a suspicious, atypical or other lesion. Gleason scores at biopsy were also ascertained in patients diagnosed with cancer following pathological review of the biopsy specimen. Prostate volume in cm³ was measured by transrectal ultrasound during prostate biopsy. The family history of prostate cancer was ascertained from the surgical chart and by a structured research questionnaire that was administered to each participant upon recruitment. Patients with a family history recorded on the chart and those who reported a family history on the questionnaire were considered positive for a family history. The study population included 1,277 patients (93%) with data available on PSA and prostate volume.

As a part of the prebiopsy clinic visit all patients recorded all current medications on a medication record form. Patients were instructed to list all current medications at home or bring the medications to this clinic visit. Medications reported on the medication record form were entered into the patient computerized medical record. Before biopsy the surgeon reviewed all survey responses with the patient to confirm that the patient had been ingesting the listed drugs. Additionally, the surgeon specifically queried the patient about drugs such as aspirin or warfarin that might affect bleeding or clotting during the biopsy procedure.

All prescription and nonprescription NSAIDs were abstracted from the medical record following patient consent. We also collected from the chart all data on prescription and nonprescription medications to treat CVD, diabetes, hyperlipidemia and BPH. The frequency and duration of use of these drugs were not systematically recorded and, therefore, they were not included in this analysis. Drugs formulated as a combination of 2 or more active drugs were classified by the component drugs.

Data Analysis 

Descriptive analyses included the chi-square test between population characteristics and NSAID use, and the Spearman correlation coefficient between PSA and prostate volume. In the primary analytical approach we compared mean PSA or prostate volume between men using vs not using aspirin, a selective COX-2 inhibitor or another NSAID. The distributions of PSA and prostate volume data were natural log transformed before analysis. In a linear model mean PSA and volume scores across NSAID categories were adjusted for patient age, race (black/white), family history (yes/no), number of prior PSA tests (1 to 3 or more), BMI, WHR, height, finasteride/dutasteride use (yes/no), other BPH treatment (yes/no), diabetes (yes/no), CVD (yes/no), hyperlipidemia (yes/no), and a diagnosis of PIN, atypical findings suspicious for cancer, or low or high grade cancer following biopsy. High grade cancer was defined by a biopsy Gleason sum score of 7 or higher. PSA and prostate volume scores were back transformed and geometric mean values are reported. Tests for trend in adjusted PSA or prostate volume were determined by the significance of a continuous variable for each categorization in a multivariable linear regression model. Tests for interaction were determined by cross-product term with NSAID use in the presence of each main effect and other covariates.

Back to Article Outline

Results 

NSAID use was reported in 586 patients (45.9%) (table 1). Aspirin was the most common NSAID reported (468 men or 36.7%), while only 45 (3.5%) reported using celecoxib and other selective COX-2 inhibitors. Other NSAIDs reported by patients included diclofenac in 5, etodolac in 8, ibuprofen in 40, indomethacin in 12, meloxicam in 17, nabumetone in 2, naproxen in 38, oxaprozin in 1, piroxicam in 1, sulindac in 1 and leflunomide in 1.

Table 1. Study population characteristics

	No. Pts (%)⁎
Age:
40–Younger than 50	48 (3.8%)
50–Younger than 60	292 (22.9%)
60–Younger than 70	518 (40.6%)
70–Younger than 80	344 (26.9%)
80–Younger than 90	75 (5.9%)
Race:
Black	144 (11.4%)
White	1,125 (88.6%)
Family history:
Yes	258 (20.2%)
No/unsure	1,019 (79.8%)
No. prior PSA tests:
1	498 (39.0%)
2	402 (31.5%)
3 or More	377 (29.5%)
NSAIDs:
Yes	586 (45.9%)
No	691 (54.1%)
Aspirin†	468 (36.7%)
Cox-2 inhibitors	45 (3.5%)
Other	91 (7.1%)
Finasteride/dutasteride:
Yes	63 (4.9%)
No	1,214 (95.1%)
Other BPH treatment:
Yes	309 (24.2%)
No	968 (75.8%)
Diabetes treatment:
Yes	172 (13.5%)
No	1,105 (86.5%)
Lipid treatment:
Yes	520 (40.7%)
No	757 (59.3%)
CVD treatment:
Yes	726 (56.9%)
No	551 (43.2%)
Biopsy outcome:
Gleason 8–10	82 (6.4%)
Gleason 7	123 (9.6%)
Gleason 6	259 (20.3%)
Atypia/suspicious	27 (2.1%)
PIN	148 (11.6%)
Neg	638 (50.0%)
Vol (ml):
Less than 40	545 (42.7%)
40–Less than 60	409 (32.0%)
60 or More	323 (25.3%)
PSA (ng/ml):
Less than 4.0	229 (17.9%)
4.0–Less than 6.0	491 (38.5%)
6.0–Less than 8.0	248 (19.4%)
8.0–Less than 10	112 (8.8%)
10 or More	197 (15.4%)

Several demographic and health related factors were associated with NSAID use, including increasing age (18.8% less than 50 years and 56.0% 80 years or greater, p <0.01), higher BMI (40.4% less than 25 kg/m² and 49.5% 35 or greater, p <0.01) and white race (47.4% white and 35.4% black, p = 0.03). NSAID use was significantly associated with medication for hyperlipidemia (yes in 61.9% of patients and no in 34.8%, p <0.01) and CVD (yes in 53.3% and no in 36.1%, p <0.01), and it was marginally associated with medication for diabetes (yes in 54.1% and no in 44.6%, p = 0.06). NSAID use was not significantly associated with receiving a steroid reductase inhibitor (p = 0.93) or with cancer grade at biopsy (p = 0.84).

PSA was approximately 9% lower in aspirin vs nonaspirin users after adjusting for age, race, family history, biopsy outcome and BPH treatment (7.3 vs 8.0 ng/ml, p = 0.01, table 2). Differences in PSA with aspirin use persisted after controlling for obesity and treatment for other conditions (7.1 vs 7.8 ng/ml, p = 0.05), and also with further adjustment for prostate volume (6.8 vs 7.5 ng/ml, p = 0.01). Results were similar when we repeated analysis after excluding any patient receiving an NSAID other than the one in question, including aspirin in 450 (PSA 7.2 vs 8.1 ng/ml, p = 0.01), COX-2 inhibitor in 27 (PSA 9.9 vs 9.0 ng/ml, p = 0.51) and other in 91 (PSA 8.7 vs 8.9 ng/ml, p = 0.69). In contrast, prostate volume was not significantly associated with NSAID use.

Table 2. Adjusted mean PSA and prostate volume by NSAID use

Drug Class	No. Pts	Adjusted⁎/Additionally Adjusted† Vol (ml)	Adjusted⁎/Additionally Adjusted† PSA (ng/ml)
NSAID:
Yes	586	48.5/47.6	7.4/7.3
No	691	47.0/46.0	8.1/7.8
p Value		0.21/0.17	0.04/0.09
Aspirin:
Yes	468	48.4/47.7	7.3/7.1
No	809	47.2/46.1	8.0/7.8
p Value		0.35/0.21	0.01/0.05
COX-2 inhibitor:
Yes	45	48.0/45.4	8.7/8.3
No	1,232	47.6/46.7	7.8/7.5
p Value		0.90/0.67	0.29/0.34
Other:
Yes	91	49.3/48.0	7.9/7.5
No	1,186	47.4/46.6	7.8/7.5
p Value		0.33/0.53	0.88/0.98

Table 3 shows the association between PSA and aspirin by age, prostate volume and diagnostic biopsy outcome. PSA was lower with aspirin use regardless of age (p interaction = 0.82). However, differences in PSA associated with aspirin were significant in men with latent cancer (p = 0.02), marginal in patients with PIN (p = 0.09) and negligible in men with negative biopsy (p = 0.64). Furthermore, PSA was lower in aspirin users with a prostate volume of greater than 60 ml (8.6 vs 9.7 ng/ml, p = 0.06) but not in men with a volume of between 40 and 60 ml (p interaction <0.01).

Table 3. PSA and aspirin association by age, biopsy outcome, and prostate volume

	Aspirin		No Aspirin		PSA p Value
	No. Pts	PSA (ng/ml)⁎	No. Pts	PSA (ng/ml)
Age:
40–Younger than 50	3	14.9	45	15.1	0.98
50–Younger than 60	72	7.9	220	8.1	0.77
60–Younger than 70	209	6.6	309	7.4	0.07
70–Younger than 80	148	7.5	196	8.5	0.11
80–Younger than 90	36	8.1	39	8.7	0.76
p Interaction	0.82
Biopsy status:
Neg	233	5.6	405	5.7	0.64
PIN	53	5.0	95	5.9	0.09
Ca	173	6.1	291	7.3	0.02
Gleason 6 Ca	99	4.8	160	5.7	0.01
Gleason 7–10 Ca	74	14.9	131	16.2	0.56
p Interaction	0.09
Vol (ml):
Less than 40	175	6.5	370	7.2	0.13
40–Less than 60	160	7.5	249	7.6	0.87
60 or More	133	8.6	190	9.7	0.06
p Interaction	<0.01

We looked at the PSA and aspirin association in prostate volume and diagnostic categories (table 4). PSA was somewhat lower in aspirin users in each category. Differences in PSA with aspirin use were statistically significant in men with cancer who had a prostate volume of 60 ml or more (7.3 vs 12.7 ng/ml, p <0.01).

Table 4. PSA and aspirin association by prostate volume and biopsy outcome

Vol (biopsy) (ml)	Aspirin		No Aspirin		PSA p Value
	No. Pts	PSA (ng/ml)⁎	No. Pts	PSA (ng/ml)
Less than 40:
Neg	71	3.9	154	4.2	0.45
PIN	20	4.3	34	5.7	0.17
Ca	81	6.5	172	6.9	0.56
Gleason 6 Ca	41	5.5	94	6.0	0.46
Gleason 7–10 Ca	40	7.9	78	8.1	0.84
40–Less than 60:
Neg	82	5.6	142	5.4	0.45
PIN	16	5.4	31	6.4	0.41
Ca	58	7.5	71	8.9	0.15
Gleason 6 Ca	34	6.3	38	6.7	0.69
Gleason 7–10 Ca	24	9.8	33	12.8	0.14
60 or More:
Neg	80	6.4	109	6.7	0.71
PIN	17	6.0	30	8.2	0.15
Ca	34	7.3	48	12.7	<0.01
Gleason 6 Ca	24	5.5	28	7.5	0.09
Gleason 7–10 Ca	10	15.5	20	27.9	0.02

Back to Article Outline

Discussion 

We found that men on aspirin had approximately 9% lower PSA. Two prior small trials targeting patients with prostate cancer with biochemical failure showed that PSA velocity decreased with selective COX-2 inhibitor administration.¹², ¹³ However, a prior cross-sectional analysis demonstrated that PSA was slightly higher with NSAID use in healthy older men with low PSA.⁹ Our results are in part consistent with these prior studies since aspirin was associated with lower PSA in men with concurrent prostate cancer or PIN but less so in men with a negative biopsy. However, the association between aspirin use and PSA in biopsy negative cases was not modified by age (data not shown) and we did not find an association between selective COX-2 inhibitors and PSA.

Aspirin is widely recommended to decrease the risk of CVD. Similarly we found that aspirin use was significantly associated with age, BMI, and treatment for CVD and hyperlipidemia. PSA and prostate volume are also associated with obesity¹⁵ and medications used to treat these chronic conditions may affect PSA expression.¹⁶ However, controlling for BMI, WHR and the use of medications to treat CVD, hyperlipidemia and diabetes had little effect on the association between aspirin use and PSA, suggesting that aspirin is not simply a proxy for these comorbid conditions or their associated treatments.

Prostate enlargement and histological BPH are among the most common conditions associated with aging and they are often seen with concurrent acute or chronic inflammation. However, the effect of aspirin on PSA appears to be separate from the clinical parameters associated with BPH. For example, PSA was lower in aspirin users regardless of age. Furthermore, prostate volume was not associated with aspirin use and, although prostate volume and PSA correlated moderately (NSAID user vs nonuser r_s 0.24, p <0.01 vs r_s 0.25, p <0.01), aspirin remained associated with lower PSA after controlling for volume. Finally, aspirin was not associated with PSA in men without latent prostate cancer. Inflammation in prostate tissue comprises multiple cellular and tissue phenotypes, and our results suggest that the effect of aspirin on PSA may be independent of the cellular or inflammatory processes associated with prostate enlargement.

In contrast, PSA was significantly lower with aspirin use in men with latent cancer. COX enzymes may be induced with conversion to cancer,⁵, ⁶ perhaps providing a target for aspirin to decrease PG synthesis and the inflammatory response. Any decrease in PGs or inflammatory infiltration into the glandular epithelium with aspirin use may be sufficient to decrease PSA without affecting overall volume. A stronger association between aspirin and PSA in men with a larger prostate volume may reflect a larger tumor volume or perhaps the effect of aspirin on PSA is more evident when prostate tissue is in a hyperproliferative or hyperinflammatory state.⁵ Furthermore, these potential mechanisms do not exclude the possibility that aspirin also more directly effects carcinogenesis, perhaps through inhibiting testosterone secretion, inducing apoptosis in prostate cancer cells or inhibiting the vascular endothelial growth factor expression associated with angiogenesis.¹⁷, ¹⁸

To the extent that aspirin lowers PSA men with latent prostate cancer may not be referred for diagnostic biopsy. An overall difference of approximately 1 ng/ml in PSA with aspirin use may have the greatest impact on men with borderline PSA, eg greater than 2.5 ng/ml. Cancer detected in these patients is often localized and may be most responsive to early treatment. Furthermore, several prior studies indicate a slight decrease in risk associated with NSAID use.⁸ However, our data suggest that such a protective association could also be caused by differential prostate cancer detection with respect to aspirin use. While we could not determine whether aspirin affects referral patterns, in men referred for biopsy we found that aspirin was not associated with cancer grade at biopsy. Finally, aspirin use could potentially affect treatment decisions because PSA at diagnosis is included in the predictive models of pathological stage¹⁹ used to inform treatment decision making. Thus, our results may have broad clinical and research implications that will require further investigation to characterize.

Strengths of this study include the evaluation of PSA and volume in clinically relevant ranges. Furthermore, evaluating patients seeking diagnostic biopsy allowed us to identify those with PIN and those without latent cancer or PIN at biopsy. BMI and WHR were measured by trained staff using standardized protocols. NSAIDs and other ascertained drugs included nonprescription drugs that would not be captured through prescription or billing records and medication assessment was done before biopsy or patient knowledge of disease status, removing the potential for recall bias. All analyses controlled for finasteride/dutasteride use and other treatment for BPH or CVD. Surgeons routinely queried patients specifically for drugs such as NSAIDs to avoid complications during the biopsy procedure, greatly increasing our ability to identify current NSAID users and the type of NSAID.

Study limitations include the cross-sectional design. Also, we were not able to capture with confidence data on the duration or dose of NSAIDs. However, aspirin use was clearly the most common NSAID, consistent with CVD prevention. Indeed, aspirin was significantly associated with CVD and hyperlipidemia treatment, and it may be reasonable to assume that most patients were ingesting aspirin on a regular basis, consistent with guidelines to prevent CVD or other chronic conditions. Results of prior studies have been inconsistent regarding the prostate cancer risk and the duration or dose of NSAIDs.⁹, ¹⁰, ¹¹ The frequency of selective COX-2 inhibitors or other nonaspirin NSAID use was insufficient to analyze thoroughly and we cannot exclude the possibility of an association with nonaspirin NSAIDS.

Back to Article Outline

Conclusions 

PSA was significantly lower in men with latent cancer who were on aspirin. In contrast to the hypothesis, prostate volume was not associated with NSAID use. Our results suggest that aspirin use may affect prostate cancer detection and they suggest a potential detection bias to address in future studies of NSAID use and prostate cancer prevention.

Back to Article Outline

References 

1. Roehrborn CG. Definition of at-risk patients: baseline variables. BJU Int. 2006;97(suppl.):7
   • View In Article
   • CrossRef
2. Nickel JC, Roehrborn CG, O'Leary MP, Bostwick DG, Somerville MC, Rittmaster RS. The relationship between prostate inflammation and lower urinary tract symptoms: examination of baseline data from the REDUCE Trial. Eur Urol. 2007;54:1379
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (464 KB)
   • CrossRef
3. Lucia MS, Torkko KC. Inflammation as a target for prostate cancer chemoprevention: pathological and laboratory rationale. J Urol. 2004;171:S30
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (268 KB)
   • CrossRef
4. Palapattu GS, Sutcliffe S, Bastian PJ, Platz EA, De Marzo AM, Isaacs WB, et al. Prostate carcinogenesis and inflammation: emerging insights. Carcinogenesis. 2005;26:1170
   • View In Article
   • MEDLINE
   • CrossRef
5. Wang W, Bergh A, Damber JE. Cyclooxygenase-2 expression correlates with local chronic inflammation and tumor neovascularization in human prostate cancer. Clin Cancer Res. 2005;11:3250
   • View In Article
   • MEDLINE
   • CrossRef
6. Denkert C, Thoma A, Niesporek S, Weichert W, Koch I, Noske A, et al. Overexpression of cyclooxygenase-2 in human prostate carcinoma and prostatic intraepithelial neoplasia-association with increased expression of Polo-like kinase-1. Prostate. 2007;67:361
   • View In Article
   • MEDLINE
   • CrossRef
7. Shahedi K, Lindstrom S, Zheng SL, Wiklund F, Adolfsson J, Sun J, et al. Genetic variation in the COX-2 gene and the association with prostate cancer risk. Int J Cancer. 2006;119:668
   • View In Article
   • MEDLINE
   • CrossRef
8. Mahmud S, Franco E, Aprikian A. Prostate cancer and use of nonsteroidal anti-inflammatory drugs: systematic review and meta-analysis. Br J Cancer. 2004;90:93
   • View In Article
   • MEDLINE
   • CrossRef
9. Platz EA, Rohrmann S, Pearson JD, Corrada MM, Watson DJ, De Marzo AM, et al. Nonsteroidal anti-inflammatory drugs and risk of prostate cancer in the Baltimore Longitudinal Study of Aging. Cancer Epidemiol Biomarkers Prev. 2005;14:390
   • View In Article
   • MEDLINE
   • CrossRef
10. Jacobs EJ, Rodriguez C, Mondul AM, Connell CJ, Henley SJ, Calle EE, et al. A large cohort study of aspirin and other nonsteroidal anti-inflammatory drugs and prostate cancer incidence. J Natl Cancer Inst. 2005;97:975
    • View In Article
    • CrossRef
11. Garcia Rodriguez LA, Gonzalez-Perez A. Inverse association between nonsteroidal anti-inflammatory drugs and prostate cancer. Cancer Epidemiol Biomarkers Prev. 2004;13:649
    • View In Article
    • MEDLINE
12. Smith MR, Manola J, Kaufman DS, Oh WK, Bubley GJ, Kantoff PW. Celecoxib versus placebo for men with prostate cancer and a rising serum prostate-specific antigen after radical prostatectomy and/or radiation therapy. J Clin Oncol. 2006;24:2723
    • View In Article
    • CrossRef
13. Pruthi RS, Derksen JE, Moore D, Carson CC, Grigson G, Watkins C, et al. Phase II trial of celecoxib in prostate-specific antigen recurrent prostate cancer after definitive radiation therapy or radical prostatectomy. Clin Cancer Res. 2006;12:2172
    • View In Article
    • MEDLINE
    • CrossRef
14. Kulkarni GS, Al-Azab R, Lockwood G, Toi A, Evans A, Trachtenberg J, et al. Evidence for a biopsy derived grade artifact among larger prostate glands. J Urol. 2006;175:505
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (71 KB)
    • CrossRef
15. Fowke JH, Signorello LB, Chang SS, Matthews CE, Buchowski MS, Cookson MS, et al. Effects of obesity and height on PSA and percent free PSA levels among African-American and Caucasian men. Cancer. 2006;107:2361
    • View In Article
    • MEDLINE
    • CrossRef
16. Yang CC, Ku CY, Wei S, Shiau CW, Chen CS, Pinzone JJ, et al. Peroxisome proliferator-activated receptor γ-independent repression of prostate-specific antigen expression by thiazolidinediones in prostate cancer cells. Mol Pharmacol. 2006;69:1564
    • View In Article
    • MEDLINE
    • CrossRef
17. Conte D, Romanelli F, Fillo S, Guidetti L, Isidori A, Franceschi F, et al. Aspirin inhibits androgen response to chorionic gonadotropin in humans. Am J Physiol Endocrinol Metab. 1999;277:E1032
    • View In Article
18. Liu XH, Kirschenbaum A, Yu K, Yao S, Levine AC. Cyclooxygenase-2 suppresses hypoxia-induced apoptosis via a combination of direct and indirect inhibition of p53 activity in a human prostate cancer cell line. J Biol Chem. 2005;280:3817
    • View In Article
    • MEDLINE
    • CrossRef
19. Makarov DV, Trock BJ, Humphreys EB, Mangold LA, Walsh PC, Epstein JI, et al. Updated nomogram to predict pathologic stage of prostate cancer given prostate-specific antigen level, clinical stage, and biopsy Gleason score (Partin tables) based on cases from 2000 to 2005. Urology. 2007;69:1095
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (106 KB)
    • CrossRef