Article Text

other Versions

PDF

Ethnic differences in disease presentation of uterine cancer in New Zealand women
  1. Ridvan Tupai Firestone1,
  2. Lis Ellison-Loschmann2,
  3. Andrew N Shelling3,
  4. Alec Ekeroma4,
  5. Bettina A Ikenasio-Thorpe5,
  6. Neil Pearce6,
  7. Mona Jeffreys7
  1. 1Research Fellow, Centre for Public Health Research, Massey University, Wellington, New Zealand
  2. 2Research Fellow, Centre for Public Health Research, Massey University, Wellington, New Zealand
  3. 3Associate Professor, Department of Obstetrics and Gynecology, Faculty of Medical and Health Sciences, University of Auckland, Auckland, New Zealand
  4. 4Head of Department, Pacific Women's Health Research and Development Unit, Department of Obstetrics and Gynaecology, University of Auckland, Auckland, New Zealand
  5. 5Lead Researcher, Pacific Health Research and Development Foundation/Pacific Women's Health Research and Development Unit, Auckland, New Zealand
  6. 6Professor, Centre for Public Health Research, Massey University, Wellington, New Zealand
  7. 7Senior Lecturer, Department of Social Medicine, University of Bristol, Bristol, UK
  1. Correspondence to Dr Ridvan Tupai Firestone, Centre for Public Health Research, Massey University, Private Box 756, Wellington, New Zealand; R.T.Firestone{at}massey.ac.nz

Abstract

Objectives Little is known about the ethnic differences in disease presentation of uterine cancer in New Zealand women. The objectives of this study were two-fold: (1) to estimate the incidence and mortality of uterine cancer among women in New Zealand and (2) to examine the association of ethnicity and socioeconomic status with tumour stage and grade, at presentation of uterine cancer.

Methods Retrospective survey of cancer cases identified from the New Zealand Cancer Registry. The authors analysed all 3203 uterine cancer cases registered with the New Zealand Cancer Registry during the period 1 January 1997 to 31 December 2006. Ethnic groups were defined based on the self-identified ethnicity recorded on the cancer registry: Ma–ori, Pacific and non-M–aori non-Pacific women. Socioeconomic status was categorised as quintiles of the New Zealand Deprivation Index 2006. The mortality to incidence ratio was used as a measure of prognosis. Logistic regression was used to estimate age, ethnic and deprivation adjusted odds ratios (ORs) and 95% confidence intervals (CIs).

Results Pacific and M–aori women have higher incidence (32.4 and 17.7 per 100 000 women, respectively) and mortality rates of uterine cancer (12.1 and 7.4 per 100 000 women, respectively). Women in the most deprived areas are more likely to present with an advanced stage of uterine cancer (OR 1.64, 95% CI 1.09–2.48). M–aori and Pacific women are less likely to present with well-differentiated tumours (OR 0.69, 95% CI 0.52–0.92 and OR 0.72, 95% CI 0.52–0.99, respectively).

Conclusions M–aori and Pacific women, and those from lower socioeconomic areas, are more likely to present with advanced uterine cancer.

Statistics from Altmetric.com

Introduction

The majority of uterine cancers are derived from the endometrial tissue within the inner layer of the uterus; hence they are referred to as endometrial cancers. These are the most frequently diagnosed genital cancers, accounting for 4% of all female cancers worldwide, and they are the fourth most common cancer in women from developed countries.1 2 In New Zealand in 2005, the age-standardised rate for cancer of the uterus was higher (13.8 per 100 000) than those for cancer of the ovary (10.5 per 100 000) and of the cervix (6.2 per 100 000).3 Age-standardised incidence rates are projected to remain stable over a 15-year period (1997–2012),4 a trend that was also projected in the USA.5 In New Zealand, mortalities from uterine cancer have declined since the 1970s, but there was a slight increase reported in the late 1990s.4 Currently, the age-standardised mortality rate from cancer of the uterus (2.6 per 100 000) is intermediary, compared to cancers of the ovary (6.1 per 100 000) and cervix (1.9 per 100 000).3 Uterine cancer has a relatively good prognosis (death to registration ratio, 0.21).3

Key message points

  • There is substantial variation in the incidence of uterine cancer by ethnicity, with the highest incidence and late stage of disease noted primarily in Pacific and Ma–ori women.

  • There also appears to be a socioeconomic gradient in the risk of uterine cancer with higher incidence in more deprived groups, but the reasons for this are unclear.

Few studies have examined the association between socioeconomic status and uterine cancer. It has been reported that the incidence of uterine cancer was more common in women of higher socioeconomic status, in Westernised countries.5 6 However, in New Zealand, the incidence and mortality from uterine cancer is higher in women from lower socioeconomic groups.3 7,,9 This article reports the recent incidence of, and mortality from, uterine cancer in New Zealand women. It also assesses the contributions of ethnicity and socioeconomic status to tumour stage and grade, at presentation of uterine cancer.

Methods

We identified all women registered on the New Zealand Cancer Registry (NZCR) with a primary diagnosis of corpus uteri and uterus part unspecified cancer [International Classification of Diseases (ICD) ICD-10 codes C54-C55, ICD-9 code 182] during the period 1 January 1997 to 31 December 2006. The data extracted from the NZCR included tumour grade and stage, histology subtypes and basic demographic information, including age at diagnosis, ethnicity, and age at death.

Uterine cancer tumour grade was characterised according to the NZCR criteria, using four categories: well differentiated, moderately differentiated, poorly differentiated, and undetermined or unknown grades. The New Zealand Health Information Service (NZHIS) used the numeric extent of disease (stage) codes, assigned by cancer registrars, which were applied to registrations up to and including 1998.10 From 1999, the extent of disease coding was standardised using the Surveillance, Epidemiology and End Results (SEER) Guide to Summary Staging.11 Thus combining the SEER guide and the numeric code, we categorised tumour stage into four categories: local, regional, distant and not known.

For descriptive analyses, age at diagnosis was divided into 15 5-year age bands from 15–19 years to 85 years and older. For logistic regression analyses, age was included as a continuous variable.

Ethnicity was classified using the standard New Zealand prioritisation system which gives the highest priority to Māori ethnicity (i.e. women who reported being Māori and also being in one or more other ethnic groups were classified as Māori) followed by Pacific ethnicity. Ethnicity was then classified into three categories: Māori, Pacific (i.e. Samoan, Cook Island Māori, Tongan, Niuean, Tokelauan, Fijian, other Pacific Island not listed and not further defined) and non-Māori non-Pacific women (the majority of whom are European in origin, but this group also includes Asian, Middle Eastern, Latin American/Hispanic and African women, as well as those for whom ethnicity was not stated).10

To measure socioeconomic status, we converted domicile codes provided by the NZHIS to the New Zealand Deprivation Index 2006 (NZDep2006) as a standardised measure of socioeconomic deprivation. Based on the 2006 New Zealand Census, the index combines nine census variables, and provides a summary deprivation score from 1 to 10 for small area units, which contain a median of 90 people. A score of 1 is allocated to the least deprived 10% of areas and 10 is allocated to the most deprived 10% of areas.12 For our analyses, deciles were grouped into quintiles: 1–2 (least deprived); 3–4; 5–6; 7–8; 9–10 (most deprived).

Analysis

The analyses were performed with the Stata (version 8.2) statistical package (StataCorp, College Station, TX, USA). Chi-square (χ2) tests were used to examine ethnic differences for descriptive characteristics. Logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for the associations between ethnicity, socioeconomic deprivation and tumour features (stage and grade). The level of statistical significance was accepted at the conventional level of p≤0.05. Two logistic models were run for each dependent variable (grade and stage of tumour); the first model involved estimation of the age-adjusted OR for ethnicity (or deprivation), and the second model included adjustment for deprivation (or ethnicity).

Uterine cancer incidence (or mortality) was estimated as the number of new registered uterine cases (or uterine cancer deaths) (aged 15+ years) during 1997–2006 per 100 000 person-years. For the denominator, we obtained census population estimates from 1997 to 2006 for each ethnic group: Pacific, Māori and non-Māori non-Pacific peoples. Age standardisation of incidence and mortality rates was conducted using Segi's 1960 World Population.13 Mortality to incidence (M:I) ratios were calculated for each ethnic group, as a measure of prognosis.

Missing data (n=14) were excluded from the analyses, where domicile codes could not be matched to the NZDep2006 index. Information on other potentially important known risk factors, such as non-use of the oral contraceptive pill, type and duration of hormone replacement therapy, age at menarche, co-existing medical conditions (e.g. diabetes and hypertension) and other important reproductive factors were not available from this database. Smoking history is recorded on the NZCR database, but more than 70% of the data was incomplete, and thus it was not included in the analyses.

Results

There were 3203 women who had registered on the NZCR with a primary diagnosis of uterine cancer from 1997 to 2006. This comprised 2564 non-Māori non-Pacific, 366 Māori, and 273 Pacific women. The unadjusted average age at first diagnosis was 63.7 years (56.9 years for Māori women, 58.1 years for Pacific women, compared to 65.3 years for non-Māori non-Pacific women) (p<0.001). Table 1 shows the distributions of clinical and demographic characteristics, by ethnicity. In this study, a large proportion of New Zealand women diagnosed with uterine cancer were living in lower socioeconomic deprivation areas (NZDep2006 7–0 decile groups). Across all ethnic groups, the majority of women were diagnosed with ‘regional’ stage of disease, although Pacific women had the highest chance of being diagnosed with distant metastases, as well as not having sufficient diagnostic tests to allow stage to be classified. For tumour grade, 38% of all women had tumours coded as ‘undetermined’, ‘unknown’, ‘not supplied’ or ‘not applicable’ and therefore they do not have a histological grade assigned. However, among those who did have histology performed, many were well-differentiated tumours. Endometrioid adenocarcinoma was the most commonly recorded histological subtype of uterine cancer for all women (Table 1).

Table 1

Distributions of uterine cancer characteristics, by ethnicity

In every age group, Pacific women (followed by Māori) have higher incidence (Table 2) and mortality rates for uterine cancer (Table 3), compared to non-Māori non-Pacific women. Table 4 shows the M:I ratios using the age-standardised rates, estimated as a measure of prognosis. Māori and Pacific women have higher values of the M:I ratio than non-Māori non-Pacific women, which suggest Māori women with uterine cancer do not have a prognosis.

Table 2

Age-specific incidence rates per 100 000 of uterine cancer, by ethnicity, 1997–2006

Table 3

Age-specific mortality rates per 100 000 of uterine cancer, by ethnicity, 1997–2006

Table 4

Mortality to incidence ratio, by ethnicity

Age and ethnicity adjusted ORs showing the associations between socioeconomic position and tumour features are shown in Table 5. The age-adjusted analyses suggested that women from more deprived areas were less likely to present with well-differentiated tumours, but this effect was attenuated following adjustment for ethnicity. Women in the most deprived areas were more likely to present with an advanced stage of uterine cancer, and the association was attenuated after adjustment for ethnicity. The age and deprivation adjusted models (Table 6) showed that Māori and Pacific women are less likely to present with well-differentiated tumours, and that little of the effect appears to be due to confounding by deprivation. Pacific women were two and a half times more likely to present with ‘advanced’ stage uterine tumours, compared to non-Māori non-Pacific women.

Table 5

Association between deprivation level and tumour features: grade and stage

Table 6

Association between ethnicity and tumour features: grade and stage

Discussion

Our study has shown that cancer of the uterus is a common gynaecological cancer affecting women in New Zealand in the older age group, and it is more common in Pacific and Māori (intermediary) women. Within this particular group of women, approximately half resided in the lowest deprived areas. Recent cancer analyses in New Zealand showed an increasing trend (by 17%) of endometrial incidence rates among those aged 25+ years old in the low-income group, compared to a 15% decrease in incidence in the high income group, from the early 1980s to 2004.14 Our study further supports the notion of notable inequalities by socioeconomic status among women with uterine cancer. New Zealand women with uterine cancer have a reasonably good prognosis (M:I=0.29) when diagnosed. This is probably because more than 70% of tumours were detected at an early stage, when treatment can be more effective. The M:I ratios appear to be elevated for Māori and Pacific women; however, this may be reflective of later diagnosis in Māori and Pacific women.

Recently, it had been reported that there was no strong evidence of an increasing trend in uterine cancer rates in all New Zealand women, except among Pacific women where rates have increased to 69% (95% CI 55.1–84.2) in 2001–2004 from 89% (95% CI 17.3–57.6) in 1981–1986.14 Our study also demonstrates higher age-standardised rates of uterine cancer among Pacific women from 1997 to 2006, particularly among women in menopause.15 However, few studies have examined ethnic differences in uterine cancer among groups in New Zealand or internationally. Our study confirms that there is substantial variation in the incidence of uterine cancer by ethnicity, with the highest incidence in Pacific and Māori women. For Pacific women, the rate was almost double that of non-Māori non-Pacific women. This differs from results reported elsewhere from studies that have specifically looked at ethnic differences in uterine cancer.16 A USA study found that White American women had the highest age-adjusted rate (116.1 per 100 000 women) compared to African-American (87.8 per 100 000 women), native Hawaiian (106.7 per 100 000 women), Japanese-American (71.3 per 100 000 women) and Latina women (73.4 per 100 000 women).16 The same authors also reported that among African-American women the incidence rate of advanced disease was 80% higher than that for White American women.16 Similarly, in the current study, there were significantly higher proportions of Pacific women who presented with advanced stage disease, which may explain in part, their higher mortality, compared to non-Māori non-Pacific women.

There are several possible explanations for the higher rate of uterine cancer among Pacific women in New Zealand. First, it has been suggested that Pacific women of this generation are not common users of the combined oral contraceptive pill, which is known to be a protective factor for uterine cancer.5 The reasons for this include the lack of adequate education on contraceptive use, and the fact that Pacific (and Māori) women are significantly less likely to use contraceptives in general.17 However, the low use of oral contraception in Pacific women is more likely to be offset by the hormonal effects of frequent pregnancies.18 Furthermore, previous studies have found that hormone contraceptive users are typically characterised as being of White/European ethnicity, having a higher level of education, being thinner and older.19 The availability of ethnic-specific data on factors such as oral contraceptive use, type and duration over a life-course would provide necessary knowledge for preventive medicine, not just in this area, but for reproductive cancers in general. Second, higher rates of uterine cancer among Pacific women could be explained, in part, by obesity, which has been consistently reported as an independent risk factor for uterine cancer.15 20 21 The biological mechanism underlying this relationship remains unclear; however, the potential explanation includes the role of obesity in the production of peripheral oestrogens, primarily through the conversion of androstenedione to estrone by aromatase in adipose tissue, increasing the risk of uterine hyperplasia and hence cancer of the uterus.22 23 Furthermore, the link between obesity and polycystic ovarian syndrome (PCOS), leading to the production of high levels of unopposed estradiol, has also been reported as a significant risk factor in uterine hyperplasia. However, the prevalence of PCOS among Pacific women has not been adequately examined. In Auckland, New Zealand a cross-sectional study reported little or no symptoms of PCOS among Pacific Island women compared to other ethnic groups (European, Maori, Indian and Asian), yet the majority of Pacific women in that study were morbidly obese and had the highest rates of insulin resistance and lipid abnormalities.24 Moreover, prior research has suggested that having a high body mass index (BMI) from the mid-teens and for a period of 30 years thereafter, and closer to the time of diagnosis, is a strong risk predictor of uterine cancer. Obtaining accurate data on obesity and BMI information requires a life-course epidemiological approach that could provide important information specific to at-risk ethnic groups.25 26

The higher rates of uterine cancer among Pacific and Māori women are puzzling because they reportedly have higher fertility rates (median age 27.7 years and 26.0 years, respectively) compared to all women (32.0 years) and they tend to have larger families.18

There is clear evidence from previous research that pregnancy (ever vs never)27 has a protective effect on uterine cancer, which increases with increasing parity.28 However, this protective effect is either not present in Pacific and Māori women, or is outweighed by other uterine cancer risk factors. Pacific and Māori women are diagnosed with uterine cancer at a younger age, which could be an indicator of familial predisposition.29 Some studies have reported differences in the genes responsible for hormonally responsive cancers, such as prostate and breast cancer, but these studies need to be extended to further large and well-characterised association studies.30 Given the different rates of diagnosis of uterine cancer, it may be important to consider whether specific genes are involved in hormone function and metabolism, and may act as predisposition genes for specific ethnic populations.

We found that women living in the most deprived areas were less likely to have well-differentiated tumours, albeit with more advanced stage of the disease, suggesting a socioeconomic gradient in risk of women living in more deprived areas, and this raises questions about access to diagnostic services for these groups of women. Furthermore, Māori and Pacific women were also less likely to have well-differentiated tumours compared to non-Māori non-Pacific women. Although this association remained significant for Māori, for Pacific women the association reduced when adjusted for deprivation. Pacific women were significantly more likely to have an advanced stage of disease at initial diagnosis (age and deprivation adjusted) compared to non-Māori non-Pacific women. Apart from the usual risk factors (e.g. obesity, poor access to health services, co-morbid conditions) it is possible that Pacific women are presenting with a more aggressive histological subtype. This hypothesis appears to be valid for African-American women who had a much higher risk of developing cancer with more aggressive histology in a USA study.5

Limitations of the study

The limitations of this study are: (1) the lack of specific histological information [e.g. International Federation of Gynecology and Obstetrics (FIGO) codes]; (2) the lack of information on other potential risk factors (e.g. oral contraceptive use, smoking history, parity vs nulliparity) that are not recorded in the NZCR; (3) the precision of our ethnic-specific rate estimates may be hampered by the small number of Pacific and Māori cases;31 although the NZCR endeavours to differentiate between New Zealand resident and non-resident registrants, there is still the potential that some non-resident registrants have been included in our analyses, resulting in inflated incidence rates. Furthermore, some cancer registrations with unspecified ethnicity were included in the non-Māori non-Pacific group, which would have reduced the observed differences between ethnic groups; and (4) the final limitation relates to a considerable proportion of women in our sample with missing ‘tumour graded’ data. This could be due to errors in coding, or that histopathology samples were not taken, or that the different information sources (clinical, radiological, histological, autopsy or death certificate) providing cancer diagnosis were incomplete. The lack of complete information on tumour grade at diagnosis will affect the precision of the estimates produced by this study, introducing the possibility of selection bias.

Conclusions

This study has provided some useful findings comparing New Zealand women with uterine cancer, by ethnic group. We have shown that there are ethnic differences in uterine cancer incidence and mortality, and that the differences in stage and grading of tumours are not completely explained by age, ethnicity and deprivation. More information around the known and unknown risk factors of this disease would sufficiently complete the knowledge gaps, especially for high-risk ethnic groups. Our study highlights some significant differences in disease presentation, by ethnic groups. In light of the findings from our study, further investigations are necessary to examine other explanatory or causal factors of uterine cancer. These factors could include occupational exposures, domestic and external environmental factors, diet composition, lifestyle and possibly genetics.

Acknowledgments

Dr R T Firestone was funded by a Pacific Postdoctoral Fellowship from the Health Research Council of New Zealand.

References

View Abstract

Footnotes

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Highlights from this issue
    British Medical Journal Publishing Group