View Article PDF

Multiple myeloma is one of the most common haematological cancers worldwide1 and involves proliferation of plasma cells within the bone marrow. It primarily affects older adults. Although myeloma remains incurable, survival has greatly increased over the past few decades,2 particularly in younger patients, due to stem cell transplantation and the availability of increasingly effective drugs.

For unknown reasons, incidence rates are consistently higher in men than women,1 black populations world-wide1,3 and in Māori.4 Older age and a positive family history5 are also regarded as risk factors, but little consensus has been reached about the importance of environmental or occupational factors. A systematic review of meta-analyses of risk factors for myeloma found positive associations with farming, pesticide exposure and exposure to farm animals, whereas exposure to benzene showed no association.6 With respect to lifestyle factors, obesity has been shown to be a risk factor,7 but no significant relationship has been found for cigarette smoking.8 In contrast, ever drinking alcohol was shown to be protective, but current alcohol consumption was not.9

A New Zealand haematologist expressed concern to one of the authors about a possible increased incidence of myeloma in a rural region of New Zealand and suggested it needed investigation (personal communication). The risks from farming-related exposures, including exposure to farm animals, are uncertain, and these activities are common in New Zealand and elsewhere. Therefore, we aimed to investigate whether this rural region had a significantly higher incidence rate than other regions, and whether there were any other regions with high myeloma incidence in New Zealand, to see whether these might inform further aetiological investigations.

Methods

In New Zealand, all cancer diagnoses made since July 1994 are notified to the New Zealand Cancer Registry (NZCR) under the Cancer Registry Act 1993. All new diagnoses of multiple myeloma or plasmacytoma (ICD-10 code C90) registered with the NZCR between 1991 and 2016 were extracted from data supplied by the Statistical Services of the Ministry of Health. For simplicity, ICD-10 code C90 is hereafter referred to as ‘myeloma’. Over the period covered by this study, the NZCR changed from ICD-9 to ICD-10 cancer codes. We found that some retrospective conversions from ICD-9 to ICD-10 for myeloma and plasmacytoma were inaccurate, so codes prior to 2000 (when ICD-10 was introduced) were checked and corrected as necessary. Only confirmed diagnoses of myeloma or plasmacytoma were included in the analyses.

Data for sex, date of diagnosis, date of birth and ethnicity were used as recorded by the NZCR or National Health Index (NHI) databases of the Ministry of Health. The ethnic classifications used prioritised ethnicity groups, as defined by Stats NZ, and whereby each individual is allocated to a single ethnicity on the basis of the following priority: Māori, Pacific peoples, Asian, other groups except New Zealand European and, finally, New Zealand European. For these analyses, ethnicity was ultimately categorised into two groups: Māori and non-Māori. All people missing an ethnic classification were included in the non-Māori category for analysis.

To avoid using geographical groupings such as district health boards (DHBs) that, because of different specialist facilities available within DHBs, may be related to myeloma diagnosis and therefore incidence rates, we used the standard 74 Territorial Local Authority (TLA) categories for 2006 (an intermediate year between 1991 and 2016 for which TLA data were available). TLAs are the second tier of local government, under the 16 regions, in New Zealand.10 These 74 TLA regions comprised 16 city councils, 57 district councils and the Chatham Islands Council. City councils administer the larger urban areas and district councils serve rural and smaller urban areas. The original geographical area of concern mapped to a specific TLA. The cumulative population from 1991 (ie, the population summed for each year from 1991 to end 2016) in the TLAs ranged from 94,333 in Kaikoura to 10.6m in Auckland City, reflecting person-years of exposure.

For some analyses, TLAs with district councils covering both rural and small urban regions were further split into small urban or rural areas, based on average population densities (<15 people per km2 for rural areas and 15–80 people per km2 for small urban areas), or they were split into the North Island and South Island regions.

As the data for myeloma were over-dispersed (as indicated by a likelihood ratio test), negative binomial regression was used to estimate incidence rates, incidence rate ratios, corresponding 95% confidence intervals (95% CIs) and p-values for comparisons between TLAs, after adjustment for the following possible confounding factors: ethnicity (Māori versus non-Māori), sex, year of diagnosis and age at diagnosis. As this analysis was precipitated by concern that a region might have a high myeloma incidence, our a priori intention was to examine whether this was supported. Therefore, adjustment for multiple testing did not apply for this region of interest, and a two-sided p<0.05 was considered to statistically confirm this difference with the average (mean) rate. Other pairwise analyses of TLAs were also performed using two-sided p-values, which were unadjusted for multiple testing and should be interpreted as exploratory in nature.

In addition to TLAs, other larger geographical groupings may also be of aetiological interest for disease incidence. Negative binomial regression models were used to explore associations between the North Island and South Island of New Zealand and between rural, small urban and urban TLAs, as well as associations involving age, sex, year and ethnicity. Where they were considered plausible, interactions were also investigated using the same approach.

All analyses were carried out in Stata 1511 and R 3.6.112 (using tmap 2.3.2).

Results

In New Zealand, between 1 January 1991 and 31 December 2016, 7,083 diagnoses of myeloma were reported to the NZCR (Table 1). This equates to an age-standardised rate of 5.3/100,000 in 2016. Over half of these were in men (56.9%), and 53.3% of diagnoses occurred in people aged 70 years and over, with the latter making up 9.5% of the population in the 2013 census. Approximately 9.1% were Māori compared to 15% of the total population in the 2013 Census data.13

Table 1: Characteristics of patients diagnosed with myeloma between 1 January 1991 and 31 December 2016.

Figure 1 shows the incidence rate ratios (IRR) for myeloma for each TLA. Although the relative myeloma incidence in Gore (the southernmost, darkest blue region in Figure 1) was higher than average, this was not statistically significant when compared to the average rate (RR=1.42; 95% CI 0.95–2.13). When comparing myeloma incidence in individual TLAs with the mean incidence in New Zealand, only Clutha (the southernmost pale-yellow region in Figure 1) had a significantly different rate, and this was significantly lower than the average. When compared to the Clutha TLA (the TLA with the lowest rate adjoining the Gore TLA), many regions had significantly higher IRR for myeloma, but there was no clear spatial pattern or latitude gradient to this.

Figure 1: Adjusted* myeloma incidence rate ratios (IRR) by New Zealand Territorial Land Authority.**

*Adjusted for age, sex, ethnicity, year of diagnosis.
**Data for 2006 TLA boundaries obtained from https://koordinates.com/layer/1247-nz-territorial-authorities-2006-census/ [last accessed 09-12-2020].

In examinations of myeloma incidence by island of residence (North Island versus South Island), incidence varied considerably (Table 2), and there were significant interactions for island of residence and Māori ethnicity (p for interaction=0.002) and sex with Māori ethnicity (p for interaction=0.008). The highest point estimate of myeloma incidence was for Māori men in the North Island: this incidence rate was significantly higher than for North Island non-Māori men (IRR=1.45; 95% CI 1.29–1.64) and North Island Māori women (IRR=1.27; 95% CI 1.07–1.49). Women on both islands and in both ethnic groups had significantly lower myeloma rates than men of the same ethnic group and in the same area. Māori women in the North Island also had significantly higher myeloma incidence than South Island Māori women (IRR=2.04; 95% CI 1.25–3.33). The lowest point estimate of myeloma incidence was for Māori women in the South Island. However, there was no significant difference in myeloma incidence between non-Māori and Māori for either men or women in the South Island (IRR=0.91, 95% CI 0.65–1.27 and IRR=1.12, 95% CI 0.69–1.82, respectively).

Table 2: Adjusted* incidence rate ratios (IRR) of myeloma incidence by ethnic group, North Island or South Island region of residence and sex.

* Adjusted for age and year of diagnosis.

As the original TLA of concern was a rural area, we also investigated whether the higher myeloma incidence was repeated in other rural or small urban areas. There was no significant difference in myeloma incidence by rural, small urban or urban location for either ethnic group or sex (Table 3). Both the lowest and the highest incidence of myeloma occurred in rural areas, and all TLAs (except Horowhenua district) in the two lowest risk categories were rural TLAs with low population densities (see Appendix Table 1). Furthermore, of the TLAs in the two highest risk categories, one was urban (Palmerston North City), two were small urban (Waipa and Wanganui districts) and the remainder were rural.

Table 3: Adjusted* incidence rate ratios (IRR) of myeloma incidence by ethnic group, population density of residence and sex.

* Adjusted for age and year of diagnosis

Discussion

To the best of our knowledge, we have carried out the first analysis of the regional variation of myeloma incidence in New Zealand. This analysis was precipitated by a haematologist’s perceptions of a higher than expected number of myeloma cases in one area of New Zealand, so our a priori intention was to examine whether this was indeed a higher rate, while exploring other geographical associations that could also inform future aetiological investigations.

The standard 74 TLA areas (2006) in New Zealand included 16 urban areas; the rest were rural/small urban. When investigating potential spatial clustering by TLA region, we found that Clutha (a rural TLA) had significantly lower incidence rates than the mean, and Gore and Ōpotiki (both rural TLAs) had non-significantly higher incidence rates compared to the mean. It is of interest to note that the regions with the highest and lowest incidence of myeloma lie next to each other on the map (Figure 1) and there is no obvious pattern or gradient of regional IRRs. In addition, myeloma incidence varied considerably by residency in the North Island or South Island, with significant interaction effects with island of residence and both ethnic group and sex. The highest myeloma incidence was for Māori men in the North Island, and the lowest was for Māori women in the South Island. There was no significant association of rurality or population density with myeloma incidence.

In addition to their public health value in the assessment of a public or clinical fear,14 geographical investigations can help to identify research questions for further research. Farming is common in New Zealand, and as myeloma development has been linked to the occupational exposure of farming in some studies (possibly mediated through exposure to farm animals or pesticides), it was important to assess whether this effect could be identified in New Zealand. Individual occupational exposure data were not available in the routinely collected datasets. However, as both the lowest and the highest incidences of myeloma occurred in rural areas, and that all TLAs (except Horowhenua) in the two lowest risk categories were rural TLAs with low population densities (see Appendix Table 1), it seems unlikely that there is an elevated risk of myeloma from farming in New Zealand.

New Zealand is a small country in terms of population size, but it has considerable diversity in both its population and physical geography, which ranges from remote rural locations to mid-size metropolitan cities such as Auckland (2018 population 1.57 million). Although many of the TLAs represent small populations, they have the advantage of being a standard geographical measure that is independent of our health system. As another way of separating geographical boundaries into larger areas of New Zealand, the TLAs were grouped into North Island or South Island regions. This classification showed significant differences for incidence of myeloma, and a significant effect modification was observed by both ethnicity and sex.

It has been reported that Māori have a higher myeloma incidence in New Zealand.4 But we found this relationship to be restricted to North Island Māori: there were no significant differences in myeloma incidence between male Māori and male non-Māori, or between female Māori and female non-Māori, in the South Island. According to the Wai 2575 Māori Health Trends Report,15 ethnicity data collection has been sufficiently reliable since 1996 to have confidence in Māori versus non-Māori comparisons. When our ethnic analyses were restricted to 1996 and thereafter, no substantive differences in either point estimates or interpretations were seen. Therefore, we presented analyses using all data from 1991.

These results suggest that some factor other than ethnicity is driving the differences in myeloma incidence by ethnic group. Deprivation is one possibility that warrants further discussion. The New Zealand Index of Deprivation16 is an area-based, not individual-level, measure of socioeconomic deprivation and includes Census17 area measures such as income, unemployment, education and home ownership. Deprivation is worse in the North Island than the South Island of New Zealand for both Māori and non-Māori, and it is worse in Māori than non-Māori in both islands.18 However, the biggest discrepancy is between North Island and South Island Māori, with Māori in the North Island 4.8 times more likely to be in the most deprived area compared to South Island Māori.

A small number of studies in other countries have found significant variation in myeloma incidence with both rurality and latitude of residence, but the direction of association has been inconsistent. A recent study of the geographic distribution of myeloma patients in Canada19 found that large metropolitan cities and high-latitude regions had a lower incidence of myeloma, compared to the high incidence rates that were found in smaller cities and rural areas largely located in the fertile croplands of Canada, as well as some around bodies of water. These results did not appear to be driven by ethnicity but were not adjusted for age, sex or socioeconomic status. In addition, an older Canadian study found that exposure to livestock did not increase risk.20 In contrast, a small study of 63 cases of multiple myeloma in Iran showed a higher incidence in urban areas.21

In New Zealand, all cancer diagnoses (except non-melanoma skin cancer) made since July 1994 are notified to the NZCR under the Cancer Registry Act 1993. Prior to 1994, cancers were notified to the NZCR but with variable completeness. Cancers reliant on the public hospital system for diagnosis (eg, myeloma) were notified, whereas those diagnosed in general practice (eg, melanoma) or in private hospitals were less likely to be notified. Currently, the NZCR collection is regarded as practically complete. However, the NZCR does not collect socioeconomic data or occupational exposure data relevant to myeloma, so this has been inferred from rurality of residence rather than from individual record data.

These results have shown no convincing evidence of a significant regional variability by TLA or rurality, although the wide confidence intervals for some effects do not allow us to rule out important geographical differences. Despite caveats around some data, we have provided a baseline of the geographical burden of myeloma in New Zealand and have suggested several directions for further epidemiological analysis.

Appendix

Appendix Table 1: Myeloma incidence rates and rate ratios by TLA adjusted for age, ethnicity and sex, and urban/rural designation.

*Coloured by incidence rate ratio. Blue <=1.0, green 1-<1.5, black 1.5-<2.0, orange 2.0-<2.5, red>=2.5
**1=urban, 2=small urban (pop density 15-80 /km2), 3 =rural (pop density <15/km2)

Summary

Abstract

AIMS: To investigate regional variation in myeloma incidence in New Zealand in order to inform aetiological investigations. METHODS: All new registrations of myeloma (1991–2016) were extracted from the New Zealand Cancer Registry. Ethnic classifications used prioritised ethnicity. For geographical groupings, 74 Territorial Local Authority (TLA) categories for 2006 and population densities were used. Negative binomial regression was used to estimate incidence rate ratios, 95% confidence intervals and p-values. RESULTS: Between 1 January 1991 and 31 December 2016, 7,083 myelomas were registered. The Clutha TLA had a significantly lower incidence than the New Zealand average. Compared to Clutha, many regions had a significantly higher incidence, but there was no clear spatial pattern. The highest incidence rate was for Māori men in the North Island. Women had significantly lower incidence than men of the same ethnic group and in the same area. CONCLUSIONS: As both extremes of myeloma incidence occurred in rural areas, and as all TLAs (except one, Horowhenua) in the two lowest risk categories were rural, it seems unlikely that farming confers an increased risk. Results suggest that some other factor is driving the differences in myeloma incidence by ethnic group. We have provided a baseline of the geographical burden of myeloma in New Zealand.

Aim

Method

Results

Conclusion

Author Information

Mary Jane Sneyd: Senior Research Fellow and Deputy Director of the Hugh Adam Cancer Epidemiology Unit, Dept of Preventive and Social Medicine, University of Otago. Andrew Gray: Biostatistician, Biostatistics Centre, University of Otago. Ian Morison: Haematologist and Professor of Pathology, Department of Pathology, University of Otago.

Acknowledgements

Financial support: The Richdale Trust.

Correspondence

Mary Jane Sneyd, Senior Research Fellow and Deputy Director of the Hugh Adam Cancer Epidemiology Unit, Dept of Preventive and Social Medicine, University of Otago, PO Box 56, Dunedin 9054, New Zealand. +64 3 4797226

Correspondence Email

mary-jane.sneyd@otago.ac.nz

Competing Interests

Dr Sneyd reports grants from the University of Otago, during the conduct of the study.

1. GLOBOCAN 2012 v1.0 Cancer Incidence and Mortality Worldwide [database on the Internet]. IARC. 2013 [cited 29/10/14]. Available from: http://globocan.iarc.fr.

2. Renshaw C, Ketley N, Moller H, Davies E. Trends in the incidence and survival of multiple myeloma in South East England 1985-2004. BMC Cancer. 2010;10:74-81.

3. Samy E, Ross J, Bolton E, et al. Variation in incidence and survival by ethnicity for patients with myeloma in England (2002-2008). Leukemia & Lymphoma. 2015;56(9):2660-7.

4. Sneyd M, Cox B, Morison I. Trends in myeloma incidence, mortality and survival in New Zealand (1985-2016). Cancer Epidemiol. 2019;60:55-9.

5. Alexander D, Mink P, Adami H, et al. Multiple myeloma: a review of the epidemiologic literature. Int J Cancer. 2007;120(Suppl 12):40-61.

6. Sergentanis T, Zagouri F, Tsilimidos G, et al. Risk factors for multiple myeloma: a systematic review of meta-analyses. Clinical lymphoma, myeloma and leukemia. 2015;15(10):563-77.

7. Wallin A, Larsson S. Body mass index and risk of multiple myeloma: a meta-analysis of prospective studies. Eur J Cancer. 2011;47:1606-15.

8. Psaltopoulou T, Sergentanis T, Kanellias N, et al. Tobacco smoking and risk of multiple myeloma: a meta-analysis of 40 observational studies. Int J Cancer. 2013;132:2413-31.

9. Psaltopoulou T, Sergentanis T, Sergentanis I, et al. Alcohol intake, alcoholic beverage type and multiple myeloma risk: a meta-analysis of 26 observational studies. Leukemia and Lymphoma. 2015;56(5):1484-501.

10. Statistics New Zealand. Territorial Authority 2015.

11. StataCorp. Stata Statistical Software: Release 15. 2017.

12. R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2019.

13. Statistics New Zealand. Ethnic groups in New Zealand. Wellington, New Zealand: Statistics New Zealans; 2014 [cited 2019].

14. Kristensen P, Svendsen K, Grimsrud T. Incidence of lymphohaematopoietic cancer at a university laboratory: a cluster investigation. Eur J Epidemiol. 2008;23:11-5.

15. Ministry of Health. Wai 2575 Maori Health Trends Report. Wellington: Ministry of Health 2019.

16. Atkinson J, Salmond C, Crampton P. NZDep2013 Index of Deprivation. Wellington: Department of Public Health, University of Otago, Wellington 2014.

17. Data tables. 2013 Census [database on the Internet]. Statistics New Zealand. 2015 [cited June, 2018].

18. Te Ropu Rangahau Hauora a Eru Pomare. DHB Maori health profiles and profile summaries. Wellington: University of Otago 2018.

19. Tsang M, Le M, Ghazawi F, et al. Multiple myeloma epidemiolgy and patient geographic distribution in Canada: a population study. Cancer. 2019;125:2435-44.

20. Fritschi L, Johnson K, Kliewer E, Fry R. Canadian Cancer Registries Epidemiology Research Group. Animal-related occupations and the risk of leukemia, myeloma, and non-Hodgkin's lymphoma in Canada. Cancer Causes Control. 2002;13:563-71.

21. Rajabli N, Naeimi-Tabeie M, Jahangirrad A, et al. Epidemiology of leukemia and multiple myeloma in Golestan, Iran. Asian Pac J Cancer Prev. 2013;14(4):2333-6.

Contact diana@nzma.org.nz
for the PDF of this article

View Article PDF

Multiple myeloma is one of the most common haematological cancers worldwide1 and involves proliferation of plasma cells within the bone marrow. It primarily affects older adults. Although myeloma remains incurable, survival has greatly increased over the past few decades,2 particularly in younger patients, due to stem cell transplantation and the availability of increasingly effective drugs.

For unknown reasons, incidence rates are consistently higher in men than women,1 black populations world-wide1,3 and in Māori.4 Older age and a positive family history5 are also regarded as risk factors, but little consensus has been reached about the importance of environmental or occupational factors. A systematic review of meta-analyses of risk factors for myeloma found positive associations with farming, pesticide exposure and exposure to farm animals, whereas exposure to benzene showed no association.6 With respect to lifestyle factors, obesity has been shown to be a risk factor,7 but no significant relationship has been found for cigarette smoking.8 In contrast, ever drinking alcohol was shown to be protective, but current alcohol consumption was not.9

A New Zealand haematologist expressed concern to one of the authors about a possible increased incidence of myeloma in a rural region of New Zealand and suggested it needed investigation (personal communication). The risks from farming-related exposures, including exposure to farm animals, are uncertain, and these activities are common in New Zealand and elsewhere. Therefore, we aimed to investigate whether this rural region had a significantly higher incidence rate than other regions, and whether there were any other regions with high myeloma incidence in New Zealand, to see whether these might inform further aetiological investigations.

Methods

In New Zealand, all cancer diagnoses made since July 1994 are notified to the New Zealand Cancer Registry (NZCR) under the Cancer Registry Act 1993. All new diagnoses of multiple myeloma or plasmacytoma (ICD-10 code C90) registered with the NZCR between 1991 and 2016 were extracted from data supplied by the Statistical Services of the Ministry of Health. For simplicity, ICD-10 code C90 is hereafter referred to as ‘myeloma’. Over the period covered by this study, the NZCR changed from ICD-9 to ICD-10 cancer codes. We found that some retrospective conversions from ICD-9 to ICD-10 for myeloma and plasmacytoma were inaccurate, so codes prior to 2000 (when ICD-10 was introduced) were checked and corrected as necessary. Only confirmed diagnoses of myeloma or plasmacytoma were included in the analyses.

Data for sex, date of diagnosis, date of birth and ethnicity were used as recorded by the NZCR or National Health Index (NHI) databases of the Ministry of Health. The ethnic classifications used prioritised ethnicity groups, as defined by Stats NZ, and whereby each individual is allocated to a single ethnicity on the basis of the following priority: Māori, Pacific peoples, Asian, other groups except New Zealand European and, finally, New Zealand European. For these analyses, ethnicity was ultimately categorised into two groups: Māori and non-Māori. All people missing an ethnic classification were included in the non-Māori category for analysis.

To avoid using geographical groupings such as district health boards (DHBs) that, because of different specialist facilities available within DHBs, may be related to myeloma diagnosis and therefore incidence rates, we used the standard 74 Territorial Local Authority (TLA) categories for 2006 (an intermediate year between 1991 and 2016 for which TLA data were available). TLAs are the second tier of local government, under the 16 regions, in New Zealand.10 These 74 TLA regions comprised 16 city councils, 57 district councils and the Chatham Islands Council. City councils administer the larger urban areas and district councils serve rural and smaller urban areas. The original geographical area of concern mapped to a specific TLA. The cumulative population from 1991 (ie, the population summed for each year from 1991 to end 2016) in the TLAs ranged from 94,333 in Kaikoura to 10.6m in Auckland City, reflecting person-years of exposure.

For some analyses, TLAs with district councils covering both rural and small urban regions were further split into small urban or rural areas, based on average population densities (<15 people per km2 for rural areas and 15–80 people per km2 for small urban areas), or they were split into the North Island and South Island regions.

As the data for myeloma were over-dispersed (as indicated by a likelihood ratio test), negative binomial regression was used to estimate incidence rates, incidence rate ratios, corresponding 95% confidence intervals (95% CIs) and p-values for comparisons between TLAs, after adjustment for the following possible confounding factors: ethnicity (Māori versus non-Māori), sex, year of diagnosis and age at diagnosis. As this analysis was precipitated by concern that a region might have a high myeloma incidence, our a priori intention was to examine whether this was supported. Therefore, adjustment for multiple testing did not apply for this region of interest, and a two-sided p<0.05 was considered to statistically confirm this difference with the average (mean) rate. Other pairwise analyses of TLAs were also performed using two-sided p-values, which were unadjusted for multiple testing and should be interpreted as exploratory in nature.

In addition to TLAs, other larger geographical groupings may also be of aetiological interest for disease incidence. Negative binomial regression models were used to explore associations between the North Island and South Island of New Zealand and between rural, small urban and urban TLAs, as well as associations involving age, sex, year and ethnicity. Where they were considered plausible, interactions were also investigated using the same approach.

All analyses were carried out in Stata 1511 and R 3.6.112 (using tmap 2.3.2).

Results

In New Zealand, between 1 January 1991 and 31 December 2016, 7,083 diagnoses of myeloma were reported to the NZCR (Table 1). This equates to an age-standardised rate of 5.3/100,000 in 2016. Over half of these were in men (56.9%), and 53.3% of diagnoses occurred in people aged 70 years and over, with the latter making up 9.5% of the population in the 2013 census. Approximately 9.1% were Māori compared to 15% of the total population in the 2013 Census data.13

Table 1: Characteristics of patients diagnosed with myeloma between 1 January 1991 and 31 December 2016.

Figure 1 shows the incidence rate ratios (IRR) for myeloma for each TLA. Although the relative myeloma incidence in Gore (the southernmost, darkest blue region in Figure 1) was higher than average, this was not statistically significant when compared to the average rate (RR=1.42; 95% CI 0.95–2.13). When comparing myeloma incidence in individual TLAs with the mean incidence in New Zealand, only Clutha (the southernmost pale-yellow region in Figure 1) had a significantly different rate, and this was significantly lower than the average. When compared to the Clutha TLA (the TLA with the lowest rate adjoining the Gore TLA), many regions had significantly higher IRR for myeloma, but there was no clear spatial pattern or latitude gradient to this.

Figure 1: Adjusted* myeloma incidence rate ratios (IRR) by New Zealand Territorial Land Authority.**

*Adjusted for age, sex, ethnicity, year of diagnosis.
**Data for 2006 TLA boundaries obtained from https://koordinates.com/layer/1247-nz-territorial-authorities-2006-census/ [last accessed 09-12-2020].

In examinations of myeloma incidence by island of residence (North Island versus South Island), incidence varied considerably (Table 2), and there were significant interactions for island of residence and Māori ethnicity (p for interaction=0.002) and sex with Māori ethnicity (p for interaction=0.008). The highest point estimate of myeloma incidence was for Māori men in the North Island: this incidence rate was significantly higher than for North Island non-Māori men (IRR=1.45; 95% CI 1.29–1.64) and North Island Māori women (IRR=1.27; 95% CI 1.07–1.49). Women on both islands and in both ethnic groups had significantly lower myeloma rates than men of the same ethnic group and in the same area. Māori women in the North Island also had significantly higher myeloma incidence than South Island Māori women (IRR=2.04; 95% CI 1.25–3.33). The lowest point estimate of myeloma incidence was for Māori women in the South Island. However, there was no significant difference in myeloma incidence between non-Māori and Māori for either men or women in the South Island (IRR=0.91, 95% CI 0.65–1.27 and IRR=1.12, 95% CI 0.69–1.82, respectively).

Table 2: Adjusted* incidence rate ratios (IRR) of myeloma incidence by ethnic group, North Island or South Island region of residence and sex.

* Adjusted for age and year of diagnosis.

As the original TLA of concern was a rural area, we also investigated whether the higher myeloma incidence was repeated in other rural or small urban areas. There was no significant difference in myeloma incidence by rural, small urban or urban location for either ethnic group or sex (Table 3). Both the lowest and the highest incidence of myeloma occurred in rural areas, and all TLAs (except Horowhenua district) in the two lowest risk categories were rural TLAs with low population densities (see Appendix Table 1). Furthermore, of the TLAs in the two highest risk categories, one was urban (Palmerston North City), two were small urban (Waipa and Wanganui districts) and the remainder were rural.

Table 3: Adjusted* incidence rate ratios (IRR) of myeloma incidence by ethnic group, population density of residence and sex.

* Adjusted for age and year of diagnosis

Discussion

To the best of our knowledge, we have carried out the first analysis of the regional variation of myeloma incidence in New Zealand. This analysis was precipitated by a haematologist’s perceptions of a higher than expected number of myeloma cases in one area of New Zealand, so our a priori intention was to examine whether this was indeed a higher rate, while exploring other geographical associations that could also inform future aetiological investigations.

The standard 74 TLA areas (2006) in New Zealand included 16 urban areas; the rest were rural/small urban. When investigating potential spatial clustering by TLA region, we found that Clutha (a rural TLA) had significantly lower incidence rates than the mean, and Gore and Ōpotiki (both rural TLAs) had non-significantly higher incidence rates compared to the mean. It is of interest to note that the regions with the highest and lowest incidence of myeloma lie next to each other on the map (Figure 1) and there is no obvious pattern or gradient of regional IRRs. In addition, myeloma incidence varied considerably by residency in the North Island or South Island, with significant interaction effects with island of residence and both ethnic group and sex. The highest myeloma incidence was for Māori men in the North Island, and the lowest was for Māori women in the South Island. There was no significant association of rurality or population density with myeloma incidence.

In addition to their public health value in the assessment of a public or clinical fear,14 geographical investigations can help to identify research questions for further research. Farming is common in New Zealand, and as myeloma development has been linked to the occupational exposure of farming in some studies (possibly mediated through exposure to farm animals or pesticides), it was important to assess whether this effect could be identified in New Zealand. Individual occupational exposure data were not available in the routinely collected datasets. However, as both the lowest and the highest incidences of myeloma occurred in rural areas, and that all TLAs (except Horowhenua) in the two lowest risk categories were rural TLAs with low population densities (see Appendix Table 1), it seems unlikely that there is an elevated risk of myeloma from farming in New Zealand.

New Zealand is a small country in terms of population size, but it has considerable diversity in both its population and physical geography, which ranges from remote rural locations to mid-size metropolitan cities such as Auckland (2018 population 1.57 million). Although many of the TLAs represent small populations, they have the advantage of being a standard geographical measure that is independent of our health system. As another way of separating geographical boundaries into larger areas of New Zealand, the TLAs were grouped into North Island or South Island regions. This classification showed significant differences for incidence of myeloma, and a significant effect modification was observed by both ethnicity and sex.

It has been reported that Māori have a higher myeloma incidence in New Zealand.4 But we found this relationship to be restricted to North Island Māori: there were no significant differences in myeloma incidence between male Māori and male non-Māori, or between female Māori and female non-Māori, in the South Island. According to the Wai 2575 Māori Health Trends Report,15 ethnicity data collection has been sufficiently reliable since 1996 to have confidence in Māori versus non-Māori comparisons. When our ethnic analyses were restricted to 1996 and thereafter, no substantive differences in either point estimates or interpretations were seen. Therefore, we presented analyses using all data from 1991.

These results suggest that some factor other than ethnicity is driving the differences in myeloma incidence by ethnic group. Deprivation is one possibility that warrants further discussion. The New Zealand Index of Deprivation16 is an area-based, not individual-level, measure of socioeconomic deprivation and includes Census17 area measures such as income, unemployment, education and home ownership. Deprivation is worse in the North Island than the South Island of New Zealand for both Māori and non-Māori, and it is worse in Māori than non-Māori in both islands.18 However, the biggest discrepancy is between North Island and South Island Māori, with Māori in the North Island 4.8 times more likely to be in the most deprived area compared to South Island Māori.

A small number of studies in other countries have found significant variation in myeloma incidence with both rurality and latitude of residence, but the direction of association has been inconsistent. A recent study of the geographic distribution of myeloma patients in Canada19 found that large metropolitan cities and high-latitude regions had a lower incidence of myeloma, compared to the high incidence rates that were found in smaller cities and rural areas largely located in the fertile croplands of Canada, as well as some around bodies of water. These results did not appear to be driven by ethnicity but were not adjusted for age, sex or socioeconomic status. In addition, an older Canadian study found that exposure to livestock did not increase risk.20 In contrast, a small study of 63 cases of multiple myeloma in Iran showed a higher incidence in urban areas.21

In New Zealand, all cancer diagnoses (except non-melanoma skin cancer) made since July 1994 are notified to the NZCR under the Cancer Registry Act 1993. Prior to 1994, cancers were notified to the NZCR but with variable completeness. Cancers reliant on the public hospital system for diagnosis (eg, myeloma) were notified, whereas those diagnosed in general practice (eg, melanoma) or in private hospitals were less likely to be notified. Currently, the NZCR collection is regarded as practically complete. However, the NZCR does not collect socioeconomic data or occupational exposure data relevant to myeloma, so this has been inferred from rurality of residence rather than from individual record data.

These results have shown no convincing evidence of a significant regional variability by TLA or rurality, although the wide confidence intervals for some effects do not allow us to rule out important geographical differences. Despite caveats around some data, we have provided a baseline of the geographical burden of myeloma in New Zealand and have suggested several directions for further epidemiological analysis.

Appendix

Appendix Table 1: Myeloma incidence rates and rate ratios by TLA adjusted for age, ethnicity and sex, and urban/rural designation.

*Coloured by incidence rate ratio. Blue <=1.0, green 1-<1.5, black 1.5-<2.0, orange 2.0-<2.5, red>=2.5
**1=urban, 2=small urban (pop density 15-80 /km2), 3 =rural (pop density <15/km2)

Summary

Abstract

AIMS: To investigate regional variation in myeloma incidence in New Zealand in order to inform aetiological investigations. METHODS: All new registrations of myeloma (1991–2016) were extracted from the New Zealand Cancer Registry. Ethnic classifications used prioritised ethnicity. For geographical groupings, 74 Territorial Local Authority (TLA) categories for 2006 and population densities were used. Negative binomial regression was used to estimate incidence rate ratios, 95% confidence intervals and p-values. RESULTS: Between 1 January 1991 and 31 December 2016, 7,083 myelomas were registered. The Clutha TLA had a significantly lower incidence than the New Zealand average. Compared to Clutha, many regions had a significantly higher incidence, but there was no clear spatial pattern. The highest incidence rate was for Māori men in the North Island. Women had significantly lower incidence than men of the same ethnic group and in the same area. CONCLUSIONS: As both extremes of myeloma incidence occurred in rural areas, and as all TLAs (except one, Horowhenua) in the two lowest risk categories were rural, it seems unlikely that farming confers an increased risk. Results suggest that some other factor is driving the differences in myeloma incidence by ethnic group. We have provided a baseline of the geographical burden of myeloma in New Zealand.

Aim

Method

Results

Conclusion

Author Information

Mary Jane Sneyd: Senior Research Fellow and Deputy Director of the Hugh Adam Cancer Epidemiology Unit, Dept of Preventive and Social Medicine, University of Otago. Andrew Gray: Biostatistician, Biostatistics Centre, University of Otago. Ian Morison: Haematologist and Professor of Pathology, Department of Pathology, University of Otago.

Acknowledgements

Financial support: The Richdale Trust.

Correspondence

Mary Jane Sneyd, Senior Research Fellow and Deputy Director of the Hugh Adam Cancer Epidemiology Unit, Dept of Preventive and Social Medicine, University of Otago, PO Box 56, Dunedin 9054, New Zealand. +64 3 4797226

Correspondence Email

mary-jane.sneyd@otago.ac.nz

Competing Interests

Dr Sneyd reports grants from the University of Otago, during the conduct of the study.

1. GLOBOCAN 2012 v1.0 Cancer Incidence and Mortality Worldwide [database on the Internet]. IARC. 2013 [cited 29/10/14]. Available from: http://globocan.iarc.fr.

2. Renshaw C, Ketley N, Moller H, Davies E. Trends in the incidence and survival of multiple myeloma in South East England 1985-2004. BMC Cancer. 2010;10:74-81.

3. Samy E, Ross J, Bolton E, et al. Variation in incidence and survival by ethnicity for patients with myeloma in England (2002-2008). Leukemia & Lymphoma. 2015;56(9):2660-7.

4. Sneyd M, Cox B, Morison I. Trends in myeloma incidence, mortality and survival in New Zealand (1985-2016). Cancer Epidemiol. 2019;60:55-9.

5. Alexander D, Mink P, Adami H, et al. Multiple myeloma: a review of the epidemiologic literature. Int J Cancer. 2007;120(Suppl 12):40-61.

6. Sergentanis T, Zagouri F, Tsilimidos G, et al. Risk factors for multiple myeloma: a systematic review of meta-analyses. Clinical lymphoma, myeloma and leukemia. 2015;15(10):563-77.

7. Wallin A, Larsson S. Body mass index and risk of multiple myeloma: a meta-analysis of prospective studies. Eur J Cancer. 2011;47:1606-15.

8. Psaltopoulou T, Sergentanis T, Kanellias N, et al. Tobacco smoking and risk of multiple myeloma: a meta-analysis of 40 observational studies. Int J Cancer. 2013;132:2413-31.

9. Psaltopoulou T, Sergentanis T, Sergentanis I, et al. Alcohol intake, alcoholic beverage type and multiple myeloma risk: a meta-analysis of 26 observational studies. Leukemia and Lymphoma. 2015;56(5):1484-501.

10. Statistics New Zealand. Territorial Authority 2015.

11. StataCorp. Stata Statistical Software: Release 15. 2017.

12. R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2019.

13. Statistics New Zealand. Ethnic groups in New Zealand. Wellington, New Zealand: Statistics New Zealans; 2014 [cited 2019].

14. Kristensen P, Svendsen K, Grimsrud T. Incidence of lymphohaematopoietic cancer at a university laboratory: a cluster investigation. Eur J Epidemiol. 2008;23:11-5.

15. Ministry of Health. Wai 2575 Maori Health Trends Report. Wellington: Ministry of Health 2019.

16. Atkinson J, Salmond C, Crampton P. NZDep2013 Index of Deprivation. Wellington: Department of Public Health, University of Otago, Wellington 2014.

17. Data tables. 2013 Census [database on the Internet]. Statistics New Zealand. 2015 [cited June, 2018].

18. Te Ropu Rangahau Hauora a Eru Pomare. DHB Maori health profiles and profile summaries. Wellington: University of Otago 2018.

19. Tsang M, Le M, Ghazawi F, et al. Multiple myeloma epidemiolgy and patient geographic distribution in Canada: a population study. Cancer. 2019;125:2435-44.

20. Fritschi L, Johnson K, Kliewer E, Fry R. Canadian Cancer Registries Epidemiology Research Group. Animal-related occupations and the risk of leukemia, myeloma, and non-Hodgkin's lymphoma in Canada. Cancer Causes Control. 2002;13:563-71.

21. Rajabli N, Naeimi-Tabeie M, Jahangirrad A, et al. Epidemiology of leukemia and multiple myeloma in Golestan, Iran. Asian Pac J Cancer Prev. 2013;14(4):2333-6.

Contact diana@nzma.org.nz
for the PDF of this article

View Article PDF

Multiple myeloma is one of the most common haematological cancers worldwide1 and involves proliferation of plasma cells within the bone marrow. It primarily affects older adults. Although myeloma remains incurable, survival has greatly increased over the past few decades,2 particularly in younger patients, due to stem cell transplantation and the availability of increasingly effective drugs.

For unknown reasons, incidence rates are consistently higher in men than women,1 black populations world-wide1,3 and in Māori.4 Older age and a positive family history5 are also regarded as risk factors, but little consensus has been reached about the importance of environmental or occupational factors. A systematic review of meta-analyses of risk factors for myeloma found positive associations with farming, pesticide exposure and exposure to farm animals, whereas exposure to benzene showed no association.6 With respect to lifestyle factors, obesity has been shown to be a risk factor,7 but no significant relationship has been found for cigarette smoking.8 In contrast, ever drinking alcohol was shown to be protective, but current alcohol consumption was not.9

A New Zealand haematologist expressed concern to one of the authors about a possible increased incidence of myeloma in a rural region of New Zealand and suggested it needed investigation (personal communication). The risks from farming-related exposures, including exposure to farm animals, are uncertain, and these activities are common in New Zealand and elsewhere. Therefore, we aimed to investigate whether this rural region had a significantly higher incidence rate than other regions, and whether there were any other regions with high myeloma incidence in New Zealand, to see whether these might inform further aetiological investigations.

Methods

In New Zealand, all cancer diagnoses made since July 1994 are notified to the New Zealand Cancer Registry (NZCR) under the Cancer Registry Act 1993. All new diagnoses of multiple myeloma or plasmacytoma (ICD-10 code C90) registered with the NZCR between 1991 and 2016 were extracted from data supplied by the Statistical Services of the Ministry of Health. For simplicity, ICD-10 code C90 is hereafter referred to as ‘myeloma’. Over the period covered by this study, the NZCR changed from ICD-9 to ICD-10 cancer codes. We found that some retrospective conversions from ICD-9 to ICD-10 for myeloma and plasmacytoma were inaccurate, so codes prior to 2000 (when ICD-10 was introduced) were checked and corrected as necessary. Only confirmed diagnoses of myeloma or plasmacytoma were included in the analyses.

Data for sex, date of diagnosis, date of birth and ethnicity were used as recorded by the NZCR or National Health Index (NHI) databases of the Ministry of Health. The ethnic classifications used prioritised ethnicity groups, as defined by Stats NZ, and whereby each individual is allocated to a single ethnicity on the basis of the following priority: Māori, Pacific peoples, Asian, other groups except New Zealand European and, finally, New Zealand European. For these analyses, ethnicity was ultimately categorised into two groups: Māori and non-Māori. All people missing an ethnic classification were included in the non-Māori category for analysis.

To avoid using geographical groupings such as district health boards (DHBs) that, because of different specialist facilities available within DHBs, may be related to myeloma diagnosis and therefore incidence rates, we used the standard 74 Territorial Local Authority (TLA) categories for 2006 (an intermediate year between 1991 and 2016 for which TLA data were available). TLAs are the second tier of local government, under the 16 regions, in New Zealand.10 These 74 TLA regions comprised 16 city councils, 57 district councils and the Chatham Islands Council. City councils administer the larger urban areas and district councils serve rural and smaller urban areas. The original geographical area of concern mapped to a specific TLA. The cumulative population from 1991 (ie, the population summed for each year from 1991 to end 2016) in the TLAs ranged from 94,333 in Kaikoura to 10.6m in Auckland City, reflecting person-years of exposure.

For some analyses, TLAs with district councils covering both rural and small urban regions were further split into small urban or rural areas, based on average population densities (<15 people per km2 for rural areas and 15–80 people per km2 for small urban areas), or they were split into the North Island and South Island regions.

As the data for myeloma were over-dispersed (as indicated by a likelihood ratio test), negative binomial regression was used to estimate incidence rates, incidence rate ratios, corresponding 95% confidence intervals (95% CIs) and p-values for comparisons between TLAs, after adjustment for the following possible confounding factors: ethnicity (Māori versus non-Māori), sex, year of diagnosis and age at diagnosis. As this analysis was precipitated by concern that a region might have a high myeloma incidence, our a priori intention was to examine whether this was supported. Therefore, adjustment for multiple testing did not apply for this region of interest, and a two-sided p<0.05 was considered to statistically confirm this difference with the average (mean) rate. Other pairwise analyses of TLAs were also performed using two-sided p-values, which were unadjusted for multiple testing and should be interpreted as exploratory in nature.

In addition to TLAs, other larger geographical groupings may also be of aetiological interest for disease incidence. Negative binomial regression models were used to explore associations between the North Island and South Island of New Zealand and between rural, small urban and urban TLAs, as well as associations involving age, sex, year and ethnicity. Where they were considered plausible, interactions were also investigated using the same approach.

All analyses were carried out in Stata 1511 and R 3.6.112 (using tmap 2.3.2).

Results

In New Zealand, between 1 January 1991 and 31 December 2016, 7,083 diagnoses of myeloma were reported to the NZCR (Table 1). This equates to an age-standardised rate of 5.3/100,000 in 2016. Over half of these were in men (56.9%), and 53.3% of diagnoses occurred in people aged 70 years and over, with the latter making up 9.5% of the population in the 2013 census. Approximately 9.1% were Māori compared to 15% of the total population in the 2013 Census data.13

Table 1: Characteristics of patients diagnosed with myeloma between 1 January 1991 and 31 December 2016.

Figure 1 shows the incidence rate ratios (IRR) for myeloma for each TLA. Although the relative myeloma incidence in Gore (the southernmost, darkest blue region in Figure 1) was higher than average, this was not statistically significant when compared to the average rate (RR=1.42; 95% CI 0.95–2.13). When comparing myeloma incidence in individual TLAs with the mean incidence in New Zealand, only Clutha (the southernmost pale-yellow region in Figure 1) had a significantly different rate, and this was significantly lower than the average. When compared to the Clutha TLA (the TLA with the lowest rate adjoining the Gore TLA), many regions had significantly higher IRR for myeloma, but there was no clear spatial pattern or latitude gradient to this.

Figure 1: Adjusted* myeloma incidence rate ratios (IRR) by New Zealand Territorial Land Authority.**

*Adjusted for age, sex, ethnicity, year of diagnosis.
**Data for 2006 TLA boundaries obtained from https://koordinates.com/layer/1247-nz-territorial-authorities-2006-census/ [last accessed 09-12-2020].

In examinations of myeloma incidence by island of residence (North Island versus South Island), incidence varied considerably (Table 2), and there were significant interactions for island of residence and Māori ethnicity (p for interaction=0.002) and sex with Māori ethnicity (p for interaction=0.008). The highest point estimate of myeloma incidence was for Māori men in the North Island: this incidence rate was significantly higher than for North Island non-Māori men (IRR=1.45; 95% CI 1.29–1.64) and North Island Māori women (IRR=1.27; 95% CI 1.07–1.49). Women on both islands and in both ethnic groups had significantly lower myeloma rates than men of the same ethnic group and in the same area. Māori women in the North Island also had significantly higher myeloma incidence than South Island Māori women (IRR=2.04; 95% CI 1.25–3.33). The lowest point estimate of myeloma incidence was for Māori women in the South Island. However, there was no significant difference in myeloma incidence between non-Māori and Māori for either men or women in the South Island (IRR=0.91, 95% CI 0.65–1.27 and IRR=1.12, 95% CI 0.69–1.82, respectively).

Table 2: Adjusted* incidence rate ratios (IRR) of myeloma incidence by ethnic group, North Island or South Island region of residence and sex.

* Adjusted for age and year of diagnosis.

As the original TLA of concern was a rural area, we also investigated whether the higher myeloma incidence was repeated in other rural or small urban areas. There was no significant difference in myeloma incidence by rural, small urban or urban location for either ethnic group or sex (Table 3). Both the lowest and the highest incidence of myeloma occurred in rural areas, and all TLAs (except Horowhenua district) in the two lowest risk categories were rural TLAs with low population densities (see Appendix Table 1). Furthermore, of the TLAs in the two highest risk categories, one was urban (Palmerston North City), two were small urban (Waipa and Wanganui districts) and the remainder were rural.

Table 3: Adjusted* incidence rate ratios (IRR) of myeloma incidence by ethnic group, population density of residence and sex.

* Adjusted for age and year of diagnosis

Discussion

To the best of our knowledge, we have carried out the first analysis of the regional variation of myeloma incidence in New Zealand. This analysis was precipitated by a haematologist’s perceptions of a higher than expected number of myeloma cases in one area of New Zealand, so our a priori intention was to examine whether this was indeed a higher rate, while exploring other geographical associations that could also inform future aetiological investigations.

The standard 74 TLA areas (2006) in New Zealand included 16 urban areas; the rest were rural/small urban. When investigating potential spatial clustering by TLA region, we found that Clutha (a rural TLA) had significantly lower incidence rates than the mean, and Gore and Ōpotiki (both rural TLAs) had non-significantly higher incidence rates compared to the mean. It is of interest to note that the regions with the highest and lowest incidence of myeloma lie next to each other on the map (Figure 1) and there is no obvious pattern or gradient of regional IRRs. In addition, myeloma incidence varied considerably by residency in the North Island or South Island, with significant interaction effects with island of residence and both ethnic group and sex. The highest myeloma incidence was for Māori men in the North Island, and the lowest was for Māori women in the South Island. There was no significant association of rurality or population density with myeloma incidence.

In addition to their public health value in the assessment of a public or clinical fear,14 geographical investigations can help to identify research questions for further research. Farming is common in New Zealand, and as myeloma development has been linked to the occupational exposure of farming in some studies (possibly mediated through exposure to farm animals or pesticides), it was important to assess whether this effect could be identified in New Zealand. Individual occupational exposure data were not available in the routinely collected datasets. However, as both the lowest and the highest incidences of myeloma occurred in rural areas, and that all TLAs (except Horowhenua) in the two lowest risk categories were rural TLAs with low population densities (see Appendix Table 1), it seems unlikely that there is an elevated risk of myeloma from farming in New Zealand.

New Zealand is a small country in terms of population size, but it has considerable diversity in both its population and physical geography, which ranges from remote rural locations to mid-size metropolitan cities such as Auckland (2018 population 1.57 million). Although many of the TLAs represent small populations, they have the advantage of being a standard geographical measure that is independent of our health system. As another way of separating geographical boundaries into larger areas of New Zealand, the TLAs were grouped into North Island or South Island regions. This classification showed significant differences for incidence of myeloma, and a significant effect modification was observed by both ethnicity and sex.

It has been reported that Māori have a higher myeloma incidence in New Zealand.4 But we found this relationship to be restricted to North Island Māori: there were no significant differences in myeloma incidence between male Māori and male non-Māori, or between female Māori and female non-Māori, in the South Island. According to the Wai 2575 Māori Health Trends Report,15 ethnicity data collection has been sufficiently reliable since 1996 to have confidence in Māori versus non-Māori comparisons. When our ethnic analyses were restricted to 1996 and thereafter, no substantive differences in either point estimates or interpretations were seen. Therefore, we presented analyses using all data from 1991.

These results suggest that some factor other than ethnicity is driving the differences in myeloma incidence by ethnic group. Deprivation is one possibility that warrants further discussion. The New Zealand Index of Deprivation16 is an area-based, not individual-level, measure of socioeconomic deprivation and includes Census17 area measures such as income, unemployment, education and home ownership. Deprivation is worse in the North Island than the South Island of New Zealand for both Māori and non-Māori, and it is worse in Māori than non-Māori in both islands.18 However, the biggest discrepancy is between North Island and South Island Māori, with Māori in the North Island 4.8 times more likely to be in the most deprived area compared to South Island Māori.

A small number of studies in other countries have found significant variation in myeloma incidence with both rurality and latitude of residence, but the direction of association has been inconsistent. A recent study of the geographic distribution of myeloma patients in Canada19 found that large metropolitan cities and high-latitude regions had a lower incidence of myeloma, compared to the high incidence rates that were found in smaller cities and rural areas largely located in the fertile croplands of Canada, as well as some around bodies of water. These results did not appear to be driven by ethnicity but were not adjusted for age, sex or socioeconomic status. In addition, an older Canadian study found that exposure to livestock did not increase risk.20 In contrast, a small study of 63 cases of multiple myeloma in Iran showed a higher incidence in urban areas.21

In New Zealand, all cancer diagnoses (except non-melanoma skin cancer) made since July 1994 are notified to the NZCR under the Cancer Registry Act 1993. Prior to 1994, cancers were notified to the NZCR but with variable completeness. Cancers reliant on the public hospital system for diagnosis (eg, myeloma) were notified, whereas those diagnosed in general practice (eg, melanoma) or in private hospitals were less likely to be notified. Currently, the NZCR collection is regarded as practically complete. However, the NZCR does not collect socioeconomic data or occupational exposure data relevant to myeloma, so this has been inferred from rurality of residence rather than from individual record data.

These results have shown no convincing evidence of a significant regional variability by TLA or rurality, although the wide confidence intervals for some effects do not allow us to rule out important geographical differences. Despite caveats around some data, we have provided a baseline of the geographical burden of myeloma in New Zealand and have suggested several directions for further epidemiological analysis.

Appendix

Appendix Table 1: Myeloma incidence rates and rate ratios by TLA adjusted for age, ethnicity and sex, and urban/rural designation.

*Coloured by incidence rate ratio. Blue <=1.0, green 1-<1.5, black 1.5-<2.0, orange 2.0-<2.5, red>=2.5
**1=urban, 2=small urban (pop density 15-80 /km2), 3 =rural (pop density <15/km2)

Summary

Abstract

AIMS: To investigate regional variation in myeloma incidence in New Zealand in order to inform aetiological investigations. METHODS: All new registrations of myeloma (1991–2016) were extracted from the New Zealand Cancer Registry. Ethnic classifications used prioritised ethnicity. For geographical groupings, 74 Territorial Local Authority (TLA) categories for 2006 and population densities were used. Negative binomial regression was used to estimate incidence rate ratios, 95% confidence intervals and p-values. RESULTS: Between 1 January 1991 and 31 December 2016, 7,083 myelomas were registered. The Clutha TLA had a significantly lower incidence than the New Zealand average. Compared to Clutha, many regions had a significantly higher incidence, but there was no clear spatial pattern. The highest incidence rate was for Māori men in the North Island. Women had significantly lower incidence than men of the same ethnic group and in the same area. CONCLUSIONS: As both extremes of myeloma incidence occurred in rural areas, and as all TLAs (except one, Horowhenua) in the two lowest risk categories were rural, it seems unlikely that farming confers an increased risk. Results suggest that some other factor is driving the differences in myeloma incidence by ethnic group. We have provided a baseline of the geographical burden of myeloma in New Zealand.

Aim

Method

Results

Conclusion

Author Information

Mary Jane Sneyd: Senior Research Fellow and Deputy Director of the Hugh Adam Cancer Epidemiology Unit, Dept of Preventive and Social Medicine, University of Otago. Andrew Gray: Biostatistician, Biostatistics Centre, University of Otago. Ian Morison: Haematologist and Professor of Pathology, Department of Pathology, University of Otago.

Acknowledgements

Financial support: The Richdale Trust.

Correspondence

Mary Jane Sneyd, Senior Research Fellow and Deputy Director of the Hugh Adam Cancer Epidemiology Unit, Dept of Preventive and Social Medicine, University of Otago, PO Box 56, Dunedin 9054, New Zealand. +64 3 4797226

Correspondence Email

mary-jane.sneyd@otago.ac.nz

Competing Interests

Dr Sneyd reports grants from the University of Otago, during the conduct of the study.

1. GLOBOCAN 2012 v1.0 Cancer Incidence and Mortality Worldwide [database on the Internet]. IARC. 2013 [cited 29/10/14]. Available from: http://globocan.iarc.fr.

2. Renshaw C, Ketley N, Moller H, Davies E. Trends in the incidence and survival of multiple myeloma in South East England 1985-2004. BMC Cancer. 2010;10:74-81.

3. Samy E, Ross J, Bolton E, et al. Variation in incidence and survival by ethnicity for patients with myeloma in England (2002-2008). Leukemia & Lymphoma. 2015;56(9):2660-7.

4. Sneyd M, Cox B, Morison I. Trends in myeloma incidence, mortality and survival in New Zealand (1985-2016). Cancer Epidemiol. 2019;60:55-9.

5. Alexander D, Mink P, Adami H, et al. Multiple myeloma: a review of the epidemiologic literature. Int J Cancer. 2007;120(Suppl 12):40-61.

6. Sergentanis T, Zagouri F, Tsilimidos G, et al. Risk factors for multiple myeloma: a systematic review of meta-analyses. Clinical lymphoma, myeloma and leukemia. 2015;15(10):563-77.

7. Wallin A, Larsson S. Body mass index and risk of multiple myeloma: a meta-analysis of prospective studies. Eur J Cancer. 2011;47:1606-15.

8. Psaltopoulou T, Sergentanis T, Kanellias N, et al. Tobacco smoking and risk of multiple myeloma: a meta-analysis of 40 observational studies. Int J Cancer. 2013;132:2413-31.

9. Psaltopoulou T, Sergentanis T, Sergentanis I, et al. Alcohol intake, alcoholic beverage type and multiple myeloma risk: a meta-analysis of 26 observational studies. Leukemia and Lymphoma. 2015;56(5):1484-501.

10. Statistics New Zealand. Territorial Authority 2015.

11. StataCorp. Stata Statistical Software: Release 15. 2017.

12. R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2019.

13. Statistics New Zealand. Ethnic groups in New Zealand. Wellington, New Zealand: Statistics New Zealans; 2014 [cited 2019].

14. Kristensen P, Svendsen K, Grimsrud T. Incidence of lymphohaematopoietic cancer at a university laboratory: a cluster investigation. Eur J Epidemiol. 2008;23:11-5.

15. Ministry of Health. Wai 2575 Maori Health Trends Report. Wellington: Ministry of Health 2019.

16. Atkinson J, Salmond C, Crampton P. NZDep2013 Index of Deprivation. Wellington: Department of Public Health, University of Otago, Wellington 2014.

17. Data tables. 2013 Census [database on the Internet]. Statistics New Zealand. 2015 [cited June, 2018].

18. Te Ropu Rangahau Hauora a Eru Pomare. DHB Maori health profiles and profile summaries. Wellington: University of Otago 2018.

19. Tsang M, Le M, Ghazawi F, et al. Multiple myeloma epidemiolgy and patient geographic distribution in Canada: a population study. Cancer. 2019;125:2435-44.

20. Fritschi L, Johnson K, Kliewer E, Fry R. Canadian Cancer Registries Epidemiology Research Group. Animal-related occupations and the risk of leukemia, myeloma, and non-Hodgkin's lymphoma in Canada. Cancer Causes Control. 2002;13:563-71.

21. Rajabli N, Naeimi-Tabeie M, Jahangirrad A, et al. Epidemiology of leukemia and multiple myeloma in Golestan, Iran. Asian Pac J Cancer Prev. 2013;14(4):2333-6.

Contact diana@nzma.org.nz
for the PDF of this article

Subscriber Content

The full contents of this pages only available to subscribers.

LOGINSUBSCRIBE