The ANECS Group comprises Spurdle A.B., Webb P., Young J. (Queensland Institute of Medical Research); Consumer representative: McQuire L.; Clinical Collaborators: NSW: Baron-Hay S., Bell D., Bonaventura A., Brand A., Braye S., Carter J., Chan F., Dalrymple C., Ferrier A. (deceased), Gard G., Hacker N., Hogg R., Houghton R., Marsden D., McIlroy K., Otton G., Pather S., Proietto A., Robertson G., Scurry J., Sharma R., Wain G., Wong F.; Qld: Armes J., Crandon A., Cummings M., Land R., Nicklin J., Perrin L., Obermair A., Ward B.; SA: Davy M., Dodd T., Miller J., Oehler M., Paramasivum S., Pierides J., Whitehead F.; Tas: Blomfield P., Challis D.; Vic: Neesham D., Pyman J., Quinn M., Rome R., Weitzer M.; WA: Brennan B., Hammond I., Leung Y., McCartney A. (deceased), Stewart C., Thompson J.; Project Managers: O'Brien S., Moore S.; Laboratory Manager: Ferguson K.; Pathology Support: Walsh M.; Admin Support: Cicero R., Green L., Griffith J., Jackman L., Ranieri B.; Laboratory Assistants: O'Brien M., Schultz P.; Research Nurses: Alexander B., Baxter C., Croy H., Fitzgerald A., Herron E., Hill C., Jones M., Maidens J., Marshall A., Martin K., Mayhew J., Minehan E., Roffe D., Shirley H., Steane H., Stenlake A., Ward A., Webb S., White J.
for the Australian National Endometrial Cancer Study Group
Aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs) have been associated with reduced risk of a number of cancer types, however, previous studies of endometrial cancer have yielded inconclusive results. We analyzed data from the Australian National Endometrial Cancer Study (ANECS), a population-based case–control study (1,398 cases, 740 controls). We systematically reviewed all the evidence linking aspirin/NSAIDs use with endometrial cancer and conducted a meta-analysis. For ANECS, unconditional logistic regression was used to estimate odds ratios (OR) adjusting for potential confounders. For the systematic review, we searched Pubmed, Embase, Web of Science and conducted a review of citations from retrieved articles. The meta-analysis risk estimates were pooled using a random-effects model. In our case–control study, women who had ever used aspirin in the last 5 years had a significantly lower risk of endometrial cancer OR = 0.78 [95% confidence interval (CI): 0.63-0.97]. There was a significant inverse dose–response (p-trend <0.001) such that women who reported using ≥2 aspirin/week had almost half the risk OR = 0.54 (0.38–0.78). No significant associations were observed between use of half-aspirin/day, non-aspirin NSAIDs or paracetamol and endometrial cancer risk. The results were similar when examined by cancer subtype. Nine studies were included in the meta-analysis. The overall pooled risk estimate for any versus no use of aspirin was 0.87 (0.79–0.96) with no evidence of heterogeneity. The pooled risk estimate for obese women (BMI ≥ 30 kg/m2) was 0.72 (0.58–0.90) but there was no association for non-obese women. Overall these results suggest that aspirin may reduce the risk of endometrial cancer, particularly among obese women.
Endometrial cancer is the sixth most common cancer in women worldwide with an incidence rate of 8.2 per 100,000 per year and a mortality rate of 2.0 per 100,000 in 2008.1 Early menarche, late menopause, nulliparity, use of estrogen-only hormone replacement therapy (HRT) and obesity have all been consistently identified as risk factors for endometrial cancer. This suggests that greater lifetime exposure to estrogens, unopposed by progesterone, is important in the etiology of this cancer.2, 3 More recently it has been proposed that chronic inflammation may work with or in addition to estrogen exposure in the development of endometrial cancer.4, 5
Chronic inflammation has been associated with the development of cancer at a number of sites and it is thought that the inflammatory process itself provides the prerequisite environment for the development of malignancies in general.6 Mediators of the inflammatory response such as the enzyme cyclooxygenase (COX) lead to the production of prostaglandins, which together with inflammatory cytokines may suppress the cell-mediated immune responses and promote angiogenesis.6 These factors may also impact on cell growth resulting in induction of cell proliferation and inhibition of apoptosis. Furthermore, chronic inflammation may also induce DNA damage.6
Two different COX enzymes exist known as COX-1 and COX-2.7 COX-1 is expressed constitutively in most tissues and plays a role in the production of prostaglandins that control physiological functions.8 In contrast, COX-2 is undetectable in most normal tissues but is induced by various inflammatory stimuli.8 All classical nonsteroidal anti-inflammatory drugs (NSAIDs) are able to inhibit both COX-1 and COX-2 to a varying extent, and therefore reduce prostaglandins throughout the body.8 Inhibition of prostaglandin synthesis is considered the predominant mechanism by which NSAIDs act as anti-inflammatory agents, but it remains unclear as to whether COX inhibition is the only anti-cancer mechanism involved.9 Nevertheless evidence clearly shows a chemopreventive effect for aspirin (acetylsalicylic acid) and other NSAIDs on colorectal cancer, and probably for other cancer types including oesophageal, gastric, lung, breast and prostate.10
Studies evaluating the potential chemoprotective effects of aspirin and non-aspirin NSAIDs on endometrial cancer risk have produced inconclusive results. To date, six studies have observed a non-significant inverse association between use of aspirin and/or non-aspirin NSAIDs and endometrial cancer risk overall,11–16 and two studies reported no association.17, 18 Six studies also stratified by obesity and two of these found a significant inverse association among obese women but not non-obese women.14, 18
In view of the inconclusive results to date, we decided to evaluate this relationship using data from the Australian National Endometrial Cancer Study (ANECS). Unlike previous studies, ANECS examined endometrial cancer separately by histological subtype allowing us to additionally explore whether any association differs between the more common low-grade endometrioid tumors (Type I) and the more aggressive high-grade tumors (Type II). ANECS also collected data on paracetamol (acetaminophen) use. Like aspirin and other NSAIDs, paracetamol is a commonly used antipyretic and analgesic, although it has very weak anti-inflammatory activity and in vitro is a weak inhibitor of both COX-1 and COX-2.19 Our hypothesis was that aspirin and non-aspirin NSAIDs use, but not paracetamol use, would be associated with a decreased risk of endometrial cancer.
We have also conducted a systematic literature review and a meta-analysis to better quantify the magnitude of the association between aspirin/NSAIDs use and endometrial cancer risk, both overall and among obese and non-obese women.
Material and Methods
The ANECS is an Australia-wide population-based case–control study. The study was approved by the Human Research Ethics Committees at the Queensland Institute of Medical Research and all participating institutions. Cases were women aged 18–79 years resident in Australia with histologically confirmed epithelial endometrial cancer newly diagnosed between July 2005 (May 2005 in Queensland) and December 2007. They were recruited by nurses who liaised with the treatment clinics and state-based cancer registries across Australia.
Of 2,707 women identified, 66 died before contact was made, physicians refused permission to contact 108 (because they were too sick or unable to give informed consent) and 82 women could not be contacted. A further 220 were excluded on the basis of language difficulties, mental incapacity or because they were too sick. The remaining 2,231 women were invited to participate and, of these, 1,497 (67%) agreed to take part. After recruitment, information on tumor site and histology was extracted from the diagnostic histopathology reports. A further 39 women were excluded because they did not have primary epithelial endometrial cancer or were first diagnosed outside the study period. Of the remaining 1,458 women, almost all (96%) completed an interview.
Control women were randomly selected from the national electoral roll (enrolment in Australia is compulsory) to match the Australian state and age distribution (in 5-year bands) of the cases. Selected women were mailed an invitation letter and information brochure and, where possible, followed up by telephone. Women with prior hysterectomy or a history of endometrial cancer were excluded. Of 1,496 eligible women contacted, 92 women were excluded due to illness, language difficulties or inability to give informed consent. Of the remaining 1,404 women, 740 (53%) participated in the study.
After obtaining written informed consent, information was collected via a standardized telephone interview for both cases and controls. Data collected included demographic, medical, hormonal, reproductive and other potential etiological factors for endometrial cancer.
Participants were asked about common medications they might have taken for conditions such as arthritis, headaches or menstrual pain. Specifically women were asked how often (on average over the last 5 years) they had taken aspirin, other nonsteroidal anti-inflammatory drugs or paracetamol. The response categories were never, occasionally, <once a month, 2–3 times per month, once a week, 2–3 times a week, 4–7 times a week and twice or more per day. For aspirin, we included a response option for use of half a tablet per day.
Cases and controls, and subgroups of controls with different levels of aspirin use, were compared using Chi-squared (χ2) tests for categorical variables and t-tests for continuous variables. Risk estimates were calculated as odds ratios (OR) with 95% confidence intervals (CI) using unconditional multiple logistic regression models. All significance tests were 2-sided and p-value<0.05 was taken as significant.
All analyses were adjusted for age in years at diagnosis/interview (<50, 50–59, 60–69, ≥70 years), age at menarche (9–11, 12, 13, 14, ≥15 years), parity (0, 1, ≥2 pregnancies of ≥6 months duration), duration of oral contraceptive (OC) use (never, <12 months, 12–59 months, 60–119 months, ≥120 months), HRT use (<3 or ≥ 3 months), body mass index (BMI) 1 year before diagnosis/recruitment (<25, 25–29.9, 30–34.9, ≥35 kg/m2), Type 2 diabetes (yes, no) and smoking status (never, past, current). Other potential confounders that were considered, but not included in the final models because they did not change risk estimates by >10%, were place of residence, education and menopausal status. Tests for linear trend were performed using the categorical variable of interest as a continuous term.
Using the criteria specified by Silverberg et al.,20 we defined Type I endometrial cancer as low-grade endometrioid and mucinous endometrial cancers, and Type II as all other epithelial subtypes including serous and clear cell cancers, high-grade endometrioid cancers and carcinosarcomas. Women with non-epithelial tumors, such as endometrial stromal sarcomas were excluded from all analyses. We analyzed data for all cases combined, as well as for Type I and Type II cases separately.
Statistical analyses were performed using SPSS Version 19.0 (IBM, SPSS for Windows, SPSS, Chicago) and Stata Version 11.0 (StataCorp. 2009. Stata Statistical Software: Release 11. College Station, TX: StataCorp LP).
Systematic review and meta-analysis
All studies that permitted assessment of the association between use of aspirin/non-aspirin NSAIDs and histologically confirmed, newly diagnosed endometrial cancer were eligible for inclusion in the systematic review and meta-analysis. Eligible studies were identified by using PubMed (U.S. Center for Biotechnology Information, U.S. National Library of Medicine, Bethesda, MD); Embase (Elsevier, Amsterdam, The Netherlands); and Web of Science (online academic citation index provided by Thomson Reuters), as well as manually checking citations from retrieved articles, from inception to December 2011. The search strategy included the MeSH terms “Aspirin” OR “Anti-Inflammatory Agents, NonSteroidal” AND “Risk Factors” AND “Endometrial Neoplasms,” as well as closely associated text terms. We read the abstracts of all identified studies to exclude those that were clearly not relevant. The full texts of the remaining articles were read to determine whether they met the study inclusion criteria. Risk estimates for aspirin use versus no aspirin use from individual studies were pooled using a random-effects model,21 and in addition, we stratified by study type (cohort, case–control) and BMI, as previous studies had reported different associations for obese versus non-obese women.
ANECS: Australian National Endometrial Cancer Study; BMI: body mass index; CI: confidence interval; COX: cyclooxygenase; HRT: hormone replacement therapy; NSAIDs: nonsteroidal anti-inflammatory drugs; OC: oral contraceptive; OR: odds ratio; RR: relative risk
ANECS case–control study
A total of 1,398 cases (Type I = 1077, Type II = 321) and 740 population-based controls were included in ANECS. The groups did not differ with respect to age (61.1 years for controls and 61.3 years for cases), education status or menopausal status (Table 1). Cases were significantly more likely to have commenced menarche at a younger age, to be nulliparous, to be obese and to have Type 2 diabetes. Controls were significantly more likely to be current or ex-smokers, to have taken OC and to have used HRT for at least 3 months (Table 1). Although estrogen exposure is a well-known risk factor for endometrial cancer, use of estrogen-only therapy was rare in our study population. In this study, HRT use was primarily estrogen plus progestin with only about 5% of women using estrogen-only forms of HRT.
Table 1. Descriptive characteristics of 1398 women with endometrial cancer and 740 population-based controls (from the Australian National Endometrial Cancer Study)
We observed an inverse association (OR = 0.78; 95% CI: 0.63–0.97) between ever use of aspirin in the last 5 years versus no use and risk of endometrial cancer (Table 2). Although there was no association for women who reported occasional use OR = 0.99 (95% CI: 0.73–1.34), overall there was a significant inverse association with increasing frequency of aspirin use (p-trend <0.001), such that women who reported use of ≥2 aspirin/week had almost half the risk of endometrial cancer OR = 0.54 (95% CI: 0.38–0.78) compared with never aspirin users. The results were similar when we considered Type I and Type II endometrial cancers separately (Table 2).
Table 2. Association between average use of aspirin, non-aspirin NSAIDs and paracetamol in the last 5 years, and risk of endometrial cancer (from the Australian National Endometrial Cancer Study)
There was no evidence of a risk reduction (OR = 0.98; 95% CI: 0.62–1.55) among women who used only half an aspirin tablet per day. Among population controls, aspirin users (standard adult doses) were older on average (62.1 years) than non-users (59.6 years), however, women who reported taking half an aspirin per day were significantly older than either of the other groups (69.1 years) and were almost all post-menopausal (97.1%). Consistent with being an older group, these women were also significantly more likely to have never used OCs (38.2%) compared with those using standard adult doses of aspirin (18.5%) and never aspirin users (16.3%). There were no significant differences between the groups with respect to age at menarche, parity, HRT use, obesity, Type 2 diabetes, smoking status, non-aspirin NSAIDs use or paracetamol use.
We considered whether the association between aspirin use and endometrial cancer risk differed between groups stratified by BMI, parity and OC use, as these factors are all strongly associated with endometrial cancer risk. When data were stratified by BMI, the adjusted OR for use of aspirin versus no aspirin among non-obese women (BMI<30) was 0.82 (95% CI: 0.63–1.06), similar to the estimate among obese women (BMI ≥30) where the adjusted OR was 0.79 (95% CI: 0.53–1.17) In contrast, use of ≥2 aspirin per week was associated with a significantly reduced risk of endometrial cancer among non-obese women (OR = 0.47; 95% CI: 0.30–0.74), but this was not the case among obese women (OR = 0.79; 95% CI: 0.42–1.49), although these estimates were not significantly different (p = 0.45).
Although we found no significant associations between aspirin use and risk of endometrial cancer among women with less than two pregnancies (OR = 1.44; 95% CI: 0.85–2.46), the adjusted OR for women who had two or more pregnancies was 0.68 (95% CI: 0.54–0.87) and this difference was statistically significant (p < 0.001). A similar difference was seen for use of ≥2 aspirin per week (p = 0.02). The associations between ever aspirin use and risk of endometrial cancer did not differ for women who had never used OC (OR = 0.75; 95% CI: 0.47–1.18), women who had used OC <5 years (OR = 0.76; 95% CI: 0.53–1.09), or women who had used OC ≥5 years (OR = 0.82; 95% CI: 0.59–1.14).
For non-aspirin NSAIDs, we found a borderline inverse association (OR = 0.81; 95% CI: 0.66–1.00) for ever use in the last 5 years versus no use, but no trend with increasing use (p-trend = 0.39). The use of paracetamol in the last 5 years was not significantly associated with endometrial cancer (Table 2). The results for non-aspirin NSAIDs and paracetamol were similar when we considered Type I and Type II endometrial cancers separately (Table 2).
Systematic review and meta-analysis
We identified eight previous studies including four cohort, two population-based and two hospital-based case–control studies that presented risk estimates for use of aspirin versus no use and risk of endometrial cancer (Table 3). Four of these studies provided results separately for non-aspirin NSAIDs and all reported null or generally weak inverse associations (Table 3). Because the data evaluating the association between non-aspirin NSAIDs and endometrial cancer are so limited, and because non-aspirin NSAIDs comprise a heterogeneous group of drugs whose precise effects are more difficult to define than the actions of aspirin, we decided to restrict further analysis to studies looking at aspirin use and risk of endometrial cancer only.
Table 3. Characteristics of the nine studies included in the meta-analysis
With respect to aspirin use, ANECS found a significant inverse association, six of the previous studies found a non-significant risk reduction among aspirin users11–16 while two studies reported no association.17, 18 Combining the results of all nine studies gave a pooled relative risk (RR) of 0.87 (95% CI: 0.79–0.96) and there was no evidence of significant heterogeneity (p = 0.57). After stratifying by study design, the pooled RRs were 0.91 (95% CI: 0.80–1.03) for the four cohort studies (p = 0.55) and 0.82 (95% CI: 0.71–0.96) for the five case–control studies (p = 0.46) (Fig. 1).
Five studies (including ANECS) provided risk estimates for the association between aspirin use versus no use and risk of endometrial cancer stratified by BMI11, 14, 17, 18 (Table 3). Two other studies reported that after stratification by BMI there were no patterns of relationships between aspirin use and endometrial cancer risk.12, 15 The pooled results from five studies for obese women (BMI ≥ 30 kg/m2) gave an RR of 0.72 (95% CI: 0.58–0.90) and there was no significant heterogeneity between results of individual studies (p = 0.43). However, the pooled RR for non-obese women (BMI < 30 kg/m2) showed no reduction in risk (RR 1.08; 95% CI: 0.82–1.43) and there was significant heterogeneity between study estimates (p = 0.03) (Fig. 2).
In this large Australian case–control study of endometrial cancer, we observed a significant inverse association between ever use versus no use of aspirin in the last 5 years and risk of endometrial cancer. Overall there was a significant inverse association with increasing aspirin use such that women who reported use of ≥2 aspirin per week had almost half the risk of endometrial cancer as never users. However, there was no association with use of half tablet of aspirin/day. There was little evidence of a reduced risk for users of non-aspirin NSAIDs and there was no association between paracetamol use and risk of endometrial cancer. The results were similar when restricted to Type I and Type II endometrial cancers.
Given our significant findings with regard to aspirin use, the lack of an association among users of half a tablet/day (equivalent to approximately 3–4 tablets/week) and risk of endometrial cancer was unexpected and required further evaluation. To assess whether this could be an artifact due to confounding by some other characteristic associated with use of half tablet of aspirin/day, we compared aspirin users and non-users. Users of half an aspirin tablet per day were found to be significantly older (post-menopausal) and less likely to have used OC than users of standard adult doses of aspirin or non-aspirin users. The age difference is probably explained by the fact that low-dose aspirin is mainly used in the treatment and prevention of cardiovascular or cerebrovascular disease, conditions more common in older people. Adjusting for age and OC use separately made little difference to the risk estimates for aspirin, so we concluded that residual confounding by these variables was unlikely to explain the marked differences between the associations with standard aspirin use and use of half an aspirin per day.
The standard adult dose of aspirin is 300–600 mg and this has antipyretic, analgesic, anti-platelet and anti-inflammatory effects.22 In contrast, the major action associated with the use of ½ 300 mg tablet/day is the inhibition of platelet aggregation.22 Although not directly comparable by dose, ANECS results relating to half an aspirin tablet/day were consistent with the findings of the Women's Health Study,23 a large-scale long-term randomized controlled trial, where it was found that use of 100 mg every second day for an average of 10 years of treatment did not lower risk of all cancers combined, or of breast, colorectal or other site-specific cancers in women, including endometrial cancer.23 These data have led to suggestions that the actual dose might have been insufficient or that giving it on alternate days was ineffective.24
Our systematic review identified eight previous studies that looked at the association between aspirin use and risk of endometrial cancer. Pooling the results of these individual studies with our ANECS data suggested that aspirin use was associated with a 13% reduced risk of endometrial cancer overall. When stratified by BMI, the pooled estimate among obese women (BMI ≥ 30 kg/m2) found a significant risk reduction overall with no significant heterogeneity between studies. However, among non-obese women (BMI < 30) the pooled estimate showed no significant association and there was significant heterogeneity between the individual studies. The reduction in risk for obese women is of interest as obesity is itself associated with a chronic low-grade state of inflammation. This inflammation has been attributed to inflammatory cytokine production, increased fatty acids and an increase in immune cells such as macrophages that themselves produce inflammatory mediators.25 It is therefore possible that the anti-inflammatory effects of aspirin would be more evident in obese women than non-obese women.
The difference in findings in our case–control study for aspirin and non-aspirin NSAIDs is of interest. All NSAIDs act through inhibiting prostaglandin synthesis via the COX pathways26 and epidemiologic evidence implicates both COX-1 and COX-2 in the chemopreventive roles of NSAIDS.26 A protective role for both aspirin and non-aspirin NSAIDs in the etiology of endometrial cancer is thus biologically plausible. Aspirin, non-aspirin NSAIDs and COX-2 selective inhibitors have been shown in vitro to markedly inhibit the proliferation of endometrial cancer cells in a time and dose dependent manner.27, 28
The use of aspirin and non-aspirin NSAIDs has been widely investigated as potentially preventive for a number of tumors. The evidence is strongest for colorectal cancer where both cohort and case–control studies indicate the incidence of colorectal cancer is about 40% lower in people who take NSAIDs regularly compared with non-users.10 In randomized trials, aspirin taken for several years has been shown to reduce both incidence and mortality due to colorectal cancer.24 Cuzick et al. have reported that an international consensus panel regarded the anti-tumor effect of aspirin and sulindac (another NSAID) as “very probable,” but the effect for other NSAIDs was listed as “possible” due to lack of evidence.10 Other cancers where aspirin/NSAIDs show promise of chemoprevention include breast,29, 30 prostate,31, 32 lung,33, 34 stomach35, 36 and oesophageal.35, 37, 38 Epidemiological evidence also suggests a protective effect for pancreatic cancer39 and melanoma.40 Interestingly, a number of these studies29–31, 37–39 found that while aspirin was associated with reduced risk of cancer, there was borderline or no association with non-aspirin NSAIDs, consistent with the findings of our ANECS analysis.
Aspirin inhibits both COX-1 and COX-2 although it preferentially inhibits COX-1,26 and is different from other NSAIDs in that its effects on COX are irreversible.41 Aspirin has several additional mechanisms of action which may contribute to its anti-cancer effect.9 It is now thought that aspirin can influence cellular processes such as apoptosis and angiogenesis (which are crucial for the development and growth of malignancies) independently of the COX pathways.9 It has recently been reported that new insights into the mechanisms of aspirin-mediated growth inhibition may indicate advantages of aspirin over other NSAIDs for treating malignant tumors.9, 42
We found interpretation of previous epidemiological studies was complicated by the different definitions of aspirin use. The case–control studies asked for aspirin use in the period 6 months to 5 years before diagnosis/interview. In contrast, three of the four cohort studies asked about aspirin use at baseline only, and did not update exposure information during 7–15 years of follow-up.12, 15, 16 This limited information does not take into account that the exposure could have changed over time and may explain the weaker associations seen in the cohort studies.
There were some limitations to our ANECS analysis including the relatively low response rate. Due to privacy restrictions, we were not permitted to access any information about non-participants. One reason for non-participation was that women were too sick to be included, therefore, non-participating cases may have had more advanced disease than those who did not participate. However, their non-participation would only have affected our results if the associations differed between women with early- and late-stage cancers, and the pattern observed for aspirin use in our results did not vary according to stage of disease (data not shown). The fact that we observed an association with aspirin use and not with other analgesic medications also suggests that our case group was not systematically biased with respect to their analgesic use.
We also acknowledge that case–control studies are subject to recall bias, and to minimize this we used highly structured questionnaires and the telephone interviews were conducted by trained research nurses. We also limited questions on medication use to average use over the last 5 years, and standard prompts were provided covering common brand names to assist with recall. To deal with confounding, we adjusted for known risk factors for endometrial cancer. Adjustment for age, age at menarche, parity, OC use, HRT use, Type 2 diabetes and smoking status had little effect on our results, while adjustment for BMI strengthened the association slightly. Our results were also very similar to results obtained by two large cohort studies.12, 15 Overall, we believe that the results we obtained were highly unlikely to have been seriously affected by bias or residual confounding.
In conclusion, the results of our population-based case–control study support the hypothesis that aspirin use (but not non-aspirin NSAIDs) may reduce the risk of endometrial cancer. The findings are further strengthened by the results of the meta-analysis. The fact that half tablet of aspirin per day was not associated with risk may indicate that dosage is important. However, it is acknowledged that data on the risk-benefit profile for aspirin and cancer prevention in general are insufficient, and future large-scale studies are needed to address unanswered questions relating to dose and duration of therapy.10, 43 As endometrial cancer is strongly associated with obesity, the significant inverse association between aspirin use and risk of endometrial cancer in obese women in the meta-analysis is of interest and warrants further investigation.
The authors acknowledge the cooperation of the following institutions. NSW: John Hunter Hospital, Liverpool Hospital, Mater Misericordiae Hospital (Sydney), Mater Misericordiae Hospital (Newcastle), Newcastle Private Hospital, North Shore Private Hospital, Royal Hospital for Women, Royal Prince Alfred Hospital, Royal North Shore Hospital, Royal Prince Alfred Hospital, St George Hospital; Westmead Hospital, Westmead Private Hospital; Qld: Brisbane Private Hospital, Greenslopes Hospital, Mater Misericordiae Hospitals, Royal Brisbane and Women's Hospital, Wesley Hospital, Queensland Cancer Registry; SA: Adelaide Pathology Partners, Burnside Hospital, Calvary Hospital, Flinders Medical Centre, Queen Elizabeth Hospital, Royal Adelaide Hospital, South Australian Cancer Registry; Tas: Launceston Hospital, North West Regional Hospitals, Royal Hobart Hospital; Vic: Freemasons Hospital, Melbourne Pathology Services, Mercy Hospital for Women, Royal Women's Hospital, Victorian Cancer Registry; WA: King Edward Memorial Hospital, St John of God Hospitals Subiaco & Murdoch, Western Australian Cancer Registry.