Schildkraut, Joellen M.*†; Moorman, Patricia G.*†; Halabi, Susan‡; Calingaert, Brian*; Marks, Jeffrey R.§; Berchuck, Andrew¶
From the *Duke Comprehensive Cancer Center, Cancer Prevention, Detection and Control Research Program, the †Department of Community and Family Medicine, the ‡Department of Biostatistics and Bioinformatics, the §Department of Experimental Surgery, and ¶Gynecology/Division of Gynecologic Oncology, Duke University Medical Center, Durham, North Carolina.
Submitted 29 December 2004; accepted 14 July 2005.
Supported by the National Institutes of Health (grant no. 1-R01-CA76016) and the Department of Defense (DOD) (grant no. DAMD17-02-1-0666).
Correspondence: Joellen M. Schildkraut, Duke University Medical Center, Box 2949, Durham NC 27710. E-mail: firstname.lastname@example.org.
Background: Previous epidemiologic research suggests that analgesic use may reduce the risk of ovarian cancer, although results are not consistent.
Methods: In a population-based, case–control study, we analyzed data from 586 ovarian cancer cases and 627 matched controls in North Carolina for the relationship between analgesic use and ovarian cancer risk. Logistic regression analysis was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) while adjusting for potential confounders.
Results: Use of any nonsteroidal antiinflammatory drugs, including aspirin, within 5 years of diagnosis/interview was found to be associated with a reduction in the risk of ovarian cancer (adjusted OR = 0.72; 95% CI = 0.56–0.92). For use of acetaminophen, the OR was 0.78 (95% CI = 0.56–1.08).
Conclusions: These data support an inverse relationship between the use of both nonsteroidal antiinflammatory drugs and acetaminophen and the risk of ovarian cancer.
Inhibition of cyclooxygenase (COX)-1 and COX-2 enzymes in the synthesis of prostaglandins is thought to be the major mechanism responsible for the antiinflammatory and antineoplastic effects of nonsteroidal antiinflammatory drugs (NSAIDs).1 In addition, NSAIDs may suppress ovulation and affect cell proliferation, angiogenesis, and apoptosis of the epithelium.2–6 Acetaminophen, another commonly used analgesic, has weak antiinflammatory activity and may have an antigonadotrophic effect.7,8 It has also been proposed that acetaminophen inhibits ovarian carcinogenesis through the depletion of glutathione leading to necrosis.9
Several studies have described an association between aspirin or other NSAID use and the risk of ovarian cancer, with some reporting a reduction in risk10–12 and others finding no association.7,13,14 Furthermore, 3 studies have reported an inverse relationship between ovarian cancer and acetaminophen use7,14,15 and 2 did not.10,16 No previous study has specifically examined an association between COX-2 inhibitors and ovarian cancer.
We explored the relationship between analgesic use and epithelial ovarian cancer in a case–control study.
We analyzed data from the North Carolina Ovarian Cancer Study, a population-based, case–control study of women with newly diagnosed epithelial ovarian cancer. Cases were age 20 to 74 years and diagnosed with invasive or borderline epithelial ovarian cancer between 1 January 1999 and 31 March 2003. Cases had no history of ovarian cancer, and they resided within a 48-county area of North Carolina. All cases underwent standardized pathologic and histologic review by the study pathologist. The response rate among eligible cases was 76%.
Controls were identified from the same 48-county region as the cases and were frequency-matched to the ovarian cancer cases on the basis of race (black or nonblack) and age (5-year age categories) using list-assisted random digit dialing. Controls had to have at least 1 intact ovary and no prior diagnosis of ovarian cancer. Among eligible controls who were identified, 73% agreed to be contacted and be sent additional study information. Among those sent additional study information, the response rate was 64% with an overall response rate of 47%.
The study protocol was approved by the Duke University Medical Center Institutional Review Board and the human subjects committees at each of the participating institutions. All participants signed a written informed consent form.
A 90-minute questionnaire was administered in person to obtain information on known and suspected ovarian cancer risk factors. The median time from diagnosis to interview was 4.5 months. Participants were asked to provide information about medications for pain or inflammation that they had taken regularly for at least 3 months in the 5 years before diagnosis (for cases) or interview (for controls). The indication for use of the medication as well as the frequency and duration of use was obtained. To assist in recall, we provided a pictorial display of prescription and over-the-counter medications, including NSAIDs (but not COX-2 inhibitors), and acetaminophen. During the first 2 years of data collection, aspirin use was not ascertained. These subjects were classified as nonusers in the primary analysis; the analysis then was repeated after omitting these subjects. Because cases and controls were similarly queried about aspirin use, the effect of including subjects for whom information on aspirin use was not available, and thus misclassifying some users as nonusers, should have attenuated the relationship between aspirin use and ovarian cancer risk.
Associations between case–control status and use within 5 years of diagnosis or interview were determined separately for NSAIDs and for acetaminophen. NSAIDs and acetaminophen use were further categorized by duration and frequency into 3 groups: less than 8 times a month, at least 8 times a month but for less than 3 years, at least 8 times a month for 3 to 5 years. NSAID use included aspirin, nonselective NSAIDs, and COX-2 selective NSAIDs; aspirin and COX-2 inhibitors also were evaluated as separate categories. We used unconditional logistic regression analysis to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for the association between the use of NSAIDs or acetaminophen and ovarian cancer while controlling for potential confounders.
The age and race distributions of the cases and controls were similar (Table 1). Other characteristics of the cases and controls are found in Table 1, in which differences on several established ovarian cancer risk factors are noted. There were not differences in indications for use (such as headache and menstrual cramps) among cases and controls who reported using NSAIDs or acetaminophen (data not shown). Thus, these indications are unlikely to be involved in the relationship between analgesic use and ovarian cancer in our study.
Analytic results for the association between analgesic use and ovarian cancer risk are shown in Table 2. The adjusted OR of 0.72 (95% CI = 0.56–0.92) is consistent with an inverse relationship between NSAID use and ovarian cancer. There was no evidence for differences in the association according to frequency or duration of use. However, the association with any NSAID use did appear somewhat stronger for those who used both prescription and over-the-counter formulations (0.44; 0.25–0.79). Among users of NSAIDs, the most common type used was ibuprofen (57%) followed by naproxen (19%). Stratification by pre- and postmenopausal status did not reveal any differences in this relationship (data not shown).
Separate analyses were also conducted to estimate the relationships of aspirin use and COX-2 inhibitors with ovarian cancer risk. The magnitude of the ORs for these associations indicates a similar reduction in ovarian cancer risk to that found with any NSAID use. When we limited the sample to those subjects who were asked whether they used aspirin (312 cases and 279 controls), the adjusted OR for aspirin use was somewhat stronger (0.40; 0.23–0.68) than when all subjects were included in the analysis.
Similar to NSAIDs, acetaminophen use was associated with a reduction in ovarian cancer risk (0.78; 0.56–1.08). There was no evidence for differences in the association according to frequency of use.
The results of a multivariable logistic regression model that simultaneously included both NSAID and acetaminophen use suggested that each category of analgesics was independently and inversely associated with ovarian cancer (Table 2). The associations between NSAIDs or acetaminophen and ovarian cancer did not substantially change when the analyses were restricted to ovarian cancer cases with invasive disease or to the serous ovarian cancer histologic subtype (data not shown).
The data from the North Carolina Ovarian Cancer Study support an inverse relationship between NSAID use and ovarian cancer that did not appear to be confounded by other epidemiologic factors associated with ovarian cancer or analgesic use. This is the first study to report an inverse relationship specifically between COX-2 inhibitors and ovarian cancer. Acetaminophen use also was inversely related to ovarian cancer risk, but the association appeared slightly weaker than that for the use of any NSAIDs. Our data did not reveal a relationship between duration of analgesic use and ovarian cancer, although our ability to detect an effect was limited due to the 5-year timeframe before diagnosis or interview for which exposure data were collected.
The inconsistency in results from previous studies of analgesic use and ovarian cancer may be due to the fact that some studies limited their focus to specific drugs rather than all NSAID use;7,10,13,14 others have taken a more comprehensive exposure assessment of all formulations,11,12 like in the current study. In the less comprehensive studies, women who used other NSAIDs or NSAIDs available through prescription were categorized as nonusers, which may have attenuated the true association. Differences in the associations across studies may be a result of differences in the definition of an analgesic user versus a nonuser, which affects the classification of the exposure among study subjects and thus the referent group in the analysis.
Our data were subject to the inherent inaccuracy in the self-report of analgesic use and the retrospective nature in assessing exposure. We would expect this measurement error to be different in the cases and controls, because women with ovarian cancer may be more likely to report factors such as the use of medication. If the true use of analgesics is actually greater than that reported among control women, the relationship we observed between analgesic use and ovarian cancer is likely attenuated toward the null.
The assessment of NSAID use only within the last 5 years was designed to reduce recall bias, but it also may have limited our ability to address the relationship between duration of analgesic use and ovarian cancer. A long-term user who stopped using analgesics more than 5 years ago also would be classified as a nonuser, which could lead to attenuation in the associations being assessed. We did not obtain the date or age of first or last use for the medications and were not able to determine the relationship between recency and latency of use and the diagnosis of ovarian cancer. We did not detect any case–control differences in the frequency of various indications for analgesic use among those who used either NSAIDs or acetaminophen, which makes it unlikely that confounding by indication would explain the associations.
Further research is needed to better characterize this inverse relationship with ovarian cancer in terms of the timing, dose, and duration of analgesic use.
This study would not have been possible without the cooperation of the North Carolina Central Tumor Registry and all of the staff of the North Carolina Ovarian Cancer study. We also thank Christine Lankevich for her management of the data collection for the North Carolina Study and Rex C. Bentley for review of the pathology in the North Carolina Ovarian Cancer Study.
1. Altinoz MA, Korkmaz R. NF-kappaB, macrophage migration inhibitory factor and cyclooxygenase-inhibitions as likely mechanisms behind the acetaminophen- and NSAID-prevention of the ovarian cancer. Neoplasma
2. Smith ER, Daly MB, Xu XX. A mechanism for cox-2 inhibitor anti-inflammatory activity in chemoprevention of epithelial cancers. Cancer Epidemiol Biomarkers Prev
3. Dannenberg AJ, Altorki NK, Boyle JO, et al. Cyclo-oxygenase 2: a pharmacological target for the prevention of cancer. Lancet Oncol
4. Xu XC. COX-2 inhibitors in cancer treatment and prevention, a recent development. Anti Cancer Drugs
5. Norman RJ, Wu R. The potential danger of COX-2 inhibitors. Fertil Steril
6. Pall M, Friden BE, Brannstrom M. Induction of delayed follicular rupture in the human by the selective COX-2 inhibitor rofecoxib: a randomized double-blind study. Human Reprod
7. Cramer DW, Harlow BL, Titus-Ernstoff L, et al. Over-the-counter analgesics and risk of ovarian cancer. Lancet
8. Sprem M, Babic D, Abramic M, et al. Glutathione and glutathione S-transferases as early markers for ovarian carcinomas: case series. Croat Med J
9. Lawson JA, Fisher MA, Simmons CA, et al. Inhibition of Fas receptor (CD95)-induced hepatic caspase activation and apoptosis by acetaminophen in mice. Toxicol Appl Pharmacol
10. Fairfield KM, Hunter DJ, Fuchs CS, et al. Aspirin, other NSAIDs, and ovarian cancer risk (United States). Cancer Causes Control
11. Sorensen HT, Friis S, Norgard B, et al. Risk of cancer in a large cohort of nonaspirin NSAID users: a population-based study. Br J Cancer
12. Rosenberg L, Palmer JR, Rao RS, et al. A case–control study of analgesic use and ovarian cancer. Cancer Epidemiol Biomarkers Prev
13. Tavani A, Gallus S, La Vecchia C, et al. Aspirin and ovarian cancer: an Italian case–control study. Ann Oncol
14. Moysich K, Mettlin C, Piver M, et al. Regular use of analgesic drugs and ovarian cancer risk. Cancer Epidemiol Biomarkers Prev
15. Rodriguez G, Walmer D, Cline M, et al. Effect of progestin on the ovarian epithelium of macaques: cancer prevention through apoptosis? J Soc Gynecol Invest
16. Rosenberg L, Louik C, Shapiro S. Nonsteroidal antiinflammatory drug use and reduced risk of large bowel carcinoma. Cancer
© 2006 Lippincott Williams & Wilkins, Inc.