Widening inequality in extreme macrosomia between Indigenous and non-Indigenous populations of Québec, Canada

Abstract

Objective : To evaluate trends in macrosomia by severity in Indigenous vs. non-Indigenous populations of Québec, Canada.

Methods: We used a retrospective cohort of 2,298,332 singleton live births in the province of Québec, 1981–2008. Indigenous births were identified by community of residence (First Nations, Inuit, non-Indigenous) and language spoken (First Nations, Inuit, French/English). High birth weight (HBW) and large-for-gestational-age (LGA) births were categorised by severity (moderate, very, extreme). Time trends in HBW and LGA, by severity, were estimated using odds ratios (OR) and rate differences for Indigenous vs. non-Indigenous births, adjusting for maternal characteristics.

Results: Relative to non-Indigenous, First Nations (but not Inuit) had higher rates of extreme HBW (1.3% vs. 0.1%) and extreme LGA birth (12.6% vs. 2.2%), and rates increased over time. First Nations had progressively elevated ORs with greater severity of macrosomia, and associations were strongest for extreme HBW >5,000 g (OR=12.4) and LGA >97th percentile (OR=7.2).

Conclusion: Inequalities in extreme macrosomia between First Nations and non-Indigenous Quebecers are pronounced and widened between 1981 and 2008.

Implications: Studies are needed to determine why macrosomia rates are increasing in Québec's First Nations, and how they compare with Indigenous sub-groups of demographically similar countries, including Australia and New Zealand.

Indigenous populations in Canada, Australia, New Zealand and the US are disproportionately disadvantaged relative to the non-Indigenous majority. Perinatal health disparities are particularly pronounced. A unique phenomenon for Canadian Indigenous populations compared with other countries is the high prevalence of macrosomia,¹ or excessive fetal growth, defined by birth weight alone (commonly, >4000 g) or birth weight for gestational age (>90^th percentile).² The prevalence of macrosomia appears particularly high among First Nations in the provinces of Québec and Ontario, with a third of infants born >4,000 g and a tenth >4,500 g,^3–5 compared with the general North American population where approximately 10% of births are >4,000 g and only 2% are >4,500 g.^6,7 In contrast, Indigenous groups of Australia,⁸ New Zealand,⁹ and the US¹⁰ have lower rates than the non-Indigenous population, though there is evidence that Torres Strait Islanders in Australia,¹¹ Pacific Islanders in New Zealand⁹ and the US,¹² as well as US Pima Indians¹³ have more births >4,500 g compared with non-Indigenous.

Few studies have sought to determine trends in extreme macrosomia over time, which is concerning because extreme fetal macrosomia is associated with much higher risk of complications, including perinatal mortality,² birth injuries,^7,14,15 obesity or diabetes later in life,¹⁶ and maternal morbidities.^7,14,15 Extreme macrosomia in Indigenous populations has been under-investigated compared with other perinatal outcomes. Data on temporal trends may be useful for planning health service delivery, and for advancing knowledge on why Indigenous groups such as Canadian First Nations, Australian Torres Strait Islanders, and New Zealand Pacific Islanders have higher rates of macrosomia than others. We sought to determine trends in extreme macrosomia for Indigenous versus non-Indigenous populations over the last three decades in Québec, the only Canadian province where both First Nations and Inuit births can be identified by mother tongue and language spoken at home on birth registrations.

Methods

We conducted a retrospective cohort study of singleton live births in the province of Québec, Canada, from 1981 through 2008 (n=2,310,466). The birth file is compiled from live birth certificates in Québec, and provides presumed complete coverage of all births to Québec residents. Births with missing weight (N=11,963) or maternal age (N=196) were excluded, leaving 2,298,332 cases for analyses. Missing birth weight data were more common for First Nations (0.8%) and Inuit (1.3%) than non-Indigenous (0.5%).

Macrosomia was defined by two indicators, high birth weight (HBW) and large-for-gestational-age (LGA) birth. Three categories of HBW were evaluated: moderate (4001–4500 g), very (4501–5000 g), and severe (>5000 g), based on birth weight cut-offs incrementally associated with perinatal morbidity and mortality.⁷ LGA birth was defined as birth weight greater than the 90^th percentile for gestational age and sex based on population reference curves, including moderate (>90^th to 95^th percentile), very (>95^th to 97^th percentile), and extreme (>97^th percentile), according to Canadian sex- and gestational age-specific fetal growth reference values for births at 22–43 gestational weeks.¹⁷ Gestational age was available for 2,279,858 births.

Indigenous status (First Nations, Inuit) was captured with parental mother tongue or home language (First Nations, Inuit, French/English, foreign, unknown), and community of residence (First Nations, Inuit, non-Indigenous).^18,19 In Québec, an Indigenous community consists of a territory or reserve identified through municipality codes (N=80). Since rates of macrosomia among speakers of Indigenous languages vary by community,¹⁸ a joint variable with a category for each language group by community was constructed. The referent was defined as French/English births residing in non-Indigenous communities (i.e., areas in the rest of Québec).

Period was assessed in three intervals (1981–1989, 1990–1999, 2000–2008). Other covariates included maternal age (<20, 20–34, ≥35 years), education (no high school diploma, high school diploma, some post-secondary, some university or more, unknown), marital status (legally married, yes/no), and parity (0, 1, ≥2 previous deliveries).

Statistical analysis

Prevalence rates of extreme, very, and moderate macrosomia were calculated by Indigenous status and period. Generalised estimating equations accounting for area-level clustering (N=1,838 municipalities) were used to compute odds ratios (OR), rate differences (RD) and 95% confidence intervals (CI) of extreme, very, and moderate HBW or LGA birth in models that were unadjusted and adjusted for maternal age, education, marital status, parity, and period. Time trends in ORs and RDs were examined, adjusting for maternal age, education, marital status, and parity. In sensitivity analyses, models were run excluding birth records with implausible birth weight for gestational age using the algorithm proposed by Alexander.²⁰ SAS version 9.2 was used to undertake analyses (SAS Institute Inc., Cary, North Carolina). This project was performed under the population health surveillance mandate of the Institut national de santé publique du Québec. The institutional review board of the University of Montréal Hospital Centre waived the requirement for ethics review, as the study was based on anonymous administrative data and conformed to the 2012 Tri-Council Policy Statement for ethical conduct of research involving humans in Canada (additional ethics approval was not required).

Results

Prevalence rates of macrosomia overall were highest for First Nations language births in First Nations communities (HBW 27.7%; LGA 28.8%) and lowest for French/English births in non-Indigenous areas (HBW 9.4%; LGA 8.1%) (Supplemental Table 1). Extreme HBW was present in more than 1% of First Nations language births in both First Nations and non-Indigenous communities, but only 0.1% of French/English births in non-Indigenous areas (Table 1). Similarly, rates of extreme LGA were higher for First Nations language births in First Nations (12.6%) and non-Indigenous communities (8.5%) compared with French/English births in non-Indigenous areas (2.2%). French/English births in First Nations communities also had high rates of extreme HBW (0.9%) and LGA (9.2%). Rates of extreme HBW were not high for Inuit language births in either Inuit or non-Indigenous communities, although elevated rates of extreme LGA birth were observed.

Table 1.  Prevalence rates of macrosomia by severity and Indigenous status, singleton live births, Québec, 1981–2008.

	High birth weight				Large-for-gestational-age birth
	Moderate n (%)	Very n (%)	Extreme n (%)	Total births	Moderate n (%)	Very n (%)	Extreme n (%)	Total births
Notes: a Not calculated due to insufficient counts.
First Nations language First Nations community Non-Indigenous area	2,407 (20.6) 626 (16.3)	675 (5.8) 139 (3.6)	149 (1.3) 42 (1.1)	11,664 3,845	1,246 (10.8) 362 (9.5)	621 (5.4) 156 (4.1)	1,458 (12.6) 322 (8.5)	11,587 3,795
Inuit language Inuit community Non-Indigenous area	495 (10.2) 48 (9.2)	81 (1.7) 12 (2.3)	6 (0.1) NC^a	4,838 522	360 (7.5) 30 (5.9)	141 (2.9) 17 (3.3)	214 (4.5) 29 (5.7)	4,784 511
French/English First Nations community Inuit community Non-Indigenous area	1,518 (16.9) 78 (11.5) 163,339 (8.2)	396 (4.4) 17 (2.5) 21,867 (1.1)	77 (0.9) NC^a 2,336 (0.1)	8,988 678 1,987,136	833 (9.3) 51 (7.6) 85,226 (4.3)	378 (4.2) 20 (3.0) 31,593 (1.6)	817 (9.2) 38 (5.7) 42,833 (2.2)	8,919 672 1,972,644

ORs of HBW for First Nations language births in First Nations communities relative to French/English births in non-Indigenous communities rose progressively with greater severity of HBW (OR_Moderate=3.45, OR_Very=6.81, OR_Extreme=12.4) (Table 2). ORs for First Nations language births in non-Indigenous communities and French/English births in First Nations communities were also elevated. ORs for LGA birth were lower in magnitude than ORs for HBW, but followed a similar trend. Inuit language births in contrast did not have higher odds of extreme HBW, although odds of very and moderate HBW were elevated. Inuit births in every area, including French/English births in Inuit communities, all had higher likelihoods of LGA birth, but a gradient with severity of LGA status was less apparent.

Table 2.  Adjusted odds ratios (OR) and 95% confidence intervals (CI) of macrosomia for Indigenous status, singleton live births, Québec, 1981–2008.^a

	High birth weight			Large-for-gestational-age birth
	Moderate	Very	Extreme	Moderate	Very	Extreme
	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)
Notes: a Adjusted for maternal age, education, marital status, previous deliveries and period, relative to French/English in non-Indigenous areas. b Not calculated due to insufficient counts.
First Nations language
First Nations community	3.45 (2.90–4.11)	6.81 (5.26–8.82)	12.4 (9.10–17.0)	3.37 (2.89–3.93)	4.37 (3.59–5.31)	7.16 (5.66–9.06)
Non-Indigenous area	2.57 (2.23–2.97)	4.11 (3.35–5.04)	10.2 (7.00–14.9)	2.89 (2.53–3.29)	3.32 (2.76–3.99)	4.81 (4.13–5.59)
Inuit language
Inuit community	1.34 (1.17–1.53)	1.55 (1.21–1.99)	0.97 (0.42–2.25)	1.84 (1.64–2.06)	1.88 (1.62–2.17)	2.00 (1.72–2.32)
Non-Indigenous area	1.25 (0.90–1.75)	2.25 (1.37–3.72)	NC^b	1.57 (1.16–2.13)	2.36 (1.43–3.88)	2.81 (1.96–4.02)
French/English
First Nations community	2.49 (2.00–3.10)	4.66 (3.16–6.87)	7.83 (4.73–13.0)	2.57 (2.04–3.23)	3.06 (2.19–4.27)	4.69 (3.10–7.08)
Inuit community	1.44 (1.16–1.79)	2.25 (1.49–3.38)	NC^b	1.80 (1.44–2.24)	1.83 (1.26–2.66)	2.47 (1.71–3.59)
Non-Indigenous area	Referent	Referent	Referent	Referent	Referent	Referent

Over time, prevalence rates of extreme HBW and LGA increased for First Nations by as much as 1.5 and 7.9 percentage points, respectively, between 1981–1989 and 2000–2008 (Table 3). In contrast, rates of extreme macrosomia for French/English births in non-Indigenous areas were stable (HBW) or increased less (LGA 0.9 percentage points). Furthermore, the increase in rates over time in First Nations was greater for extreme than very or moderate LGA birth. Adjusted rate differences confirmed these patterns and suggested that the increase in rates of extreme LGA birth was largest for First Nations language births in First Nations communities (Table 4). Inuit rates were stable over time. In sensitivity analyses, similar results were observed for both sexes, and exclusion of births with potentially implausible birth weights did not influence results (data not shown).

Table 3.  Prevalence rates of extreme, very, and moderate macrosomia for First Nations and non-Indigenous over time, singleton live births, Québec, 1981–2008.^a

Table 4.  Prevalence rate differences (%) in macrosomia by severity over time for First Nations vs. non-Indigenous, singleton live births, Québec, 1981–2008.^a

Discussion

This study documented trends in extreme macrosomia for Indigenous populations of Canada. We found a marked increase in rates of extreme macrosomia for First Nations over time. Odds of extreme HBW and extreme LGA birth for First Nations language births and residents of First Nations communities were very high relative to the non-Indigenous population. Relative inequalities in macrosomic birth rates between First Nations and the non-Indigenous population were particularly pronounced for extreme HBW and LGA. For Inuit, however, inequalities relative to the non-Indigenous population were smaller, with no increase in rates of extreme macrosomia over time.

Our findings confirm that rates of macrosomia are elevated in First Nations and Inuit populations of Canada,^3–5 but extend these observations by demonstrating the extent to which extreme macrosomia affects First Nations in Québec, including a rapid increase over three decades. In contrast, Australian Aboriginals and New Zealand Māori have lower rates of macrosomia compared to Whites, whereas Australian Torres Strait Islanders and New Zealand Pacific Islanders have higher rates, but data on trends over time are lacking.^9,11 In the US, rates of macrosomia vary depending on Indigenous origin, with a greater prevalence in Pima Indians¹³ and Hawaiian Pacific Islanders,¹² although the problem is not as pronounced as in Canadian First Nations. In general, US Indigenous groups have lower rates of macrosomia.¹⁰

Why macrosomia is more prevalent in Canadian First Nations than Indigenous populations of other countries is unclear, but genetic²¹ and environmental differences may be implicated. Indigenous populations transitioned rapidly from traditional hunting and gathering to predominantly sedentary lifestyles with carbohydrate-rich diets, and Indigenous subgroups vary in genetic susceptibility to diabetes and obesity,^22,23 two important risk factors for macrosomia.^6,24,25 Such factors are prevalent in First Nations,^3–5,26 with as many as 53% of pregnancies complicated by obesity and 12% by gestational diabetes (versus 11% and 5% of non-Indigenous pregnancies, respectively).^3,4 Inuit women of childbearing age also have high rates of obesity (24%),²⁷ which contrasts with their relatively low prevalence of macrosomia. Evidence suggests that increases in maternal weight contributed to larger fetal size over time,⁶ but temporal trends for Indigenous groups are poorly understood, including how they relate to patterns in macrosomia. Furthermore, Australian Torres Strait Islanders (33%) and US Pima Indians (38%) have higher diabetes rates than Canadian First Nations (26%), yet macrosomia is much more prevalent in the latter,²² and New Zealand Pacific Islanders have lower rates of diabetes than Māori,²² but higher rates of macrosomia.⁹ Trends in smoking, a well-established risk factor for low birth weight, may partly account for the increase in macrosomia,⁶ as rates have fallen over time in Québec.²⁸ Smoking during pregnancy is more prevalent in Inuit (82%) than First Nations (36%) and Inuit have lower rates of diabetes,²⁷ which may help explain why extreme macrosomia is more common in First Nations. More accurate measurement of gestational age over time may have increased rates of LGA birth,²⁹ but systematic differences between Indigenous and non-Indigenous Quebecers are unlikely, and the increase was also observed for HBW. Higher rates of preterm birth over time, especially among Inuit and non-Indigenous,³⁰ may have slowed the increase in HBW, but the extent is unclear.

Another interesting finding was that disparities between Indigenous and the non-Indigenous population varied by residential community. Compared with French/English births in non-Indigenous areas, disparities were larger for First Nations language births in First Nations communities than First Nations language births in non-Indigenous areas. This finding suggests that living on reserves may be a stronger risk factor for extreme macrosomia than living in non-Indigenous areas, as high rates were even observed for French/English language births in Indigenous areas (who likely represent a mix of Indigenous and non-Indigenous births). Many First Nations communities are located in remote northern regions, whereas First Nations in non-Indigenous areas are more likely to live in the more densely populated areas of southern Québec, where access to health care or nutritious foods may be easier.¹⁸ The extent to which such differences account for our findings is unknown.

Limitations included lack of data on gestational or pre-existing diabetes, maternal weight, income and behaviours such as smoking and alcohol use, which are potential intermediates in the pathway between Indigenous status and macrosomia. Language and area of residence were proxies for Indigenous status and may not have captured all Indigenous births, resulting in possible non-differential misclassification of some Indigenous births as non-Indigenous (i.e., French/English in non-Indigenous areas). Misclassifications may have attenuated the disparities. Changes in recording of variables over time and the impact on study findings could not be determined.

Conclusion

In summary, we found progressively greater disparity in prevalence of extreme macrosomia over time for First Nations versus non-Indigenous populations of Québec. Greater awareness of the high and increasing risk of extreme macrosomia in Canadian First Nations is warranted. Further studies are needed to determine the underlying causes and to understand how these trends compare with Indigenous populations of other developed countries such as Australia and New Zealand.

Acknowledgments

This work was supported by the Establishment of Young Researchers Junior 1 of the Fonds de recherche du Québec (FRQ). NA acknowledges a career award from the FRQ.

Supporting Information

Additional supporting information may be found in the online version of this article:

Supplementary Table 1: Macrosomia prevalence rates according to maternal characteristics, singleton live births, Québec, 1981–2008.