Head circumference values among Inuit children in Nunavut, Canada: a retrospective cohort study

Abstract

Background

Inuit children from Nunavut have been observed to have high rates of macrocephaly, which sometimes leads to burdensome travel for medical evaluation, often with no pathology identified upon assessment. Given reports that World Health Organization (WHO) growth charts may not reflect all populations, we sought to compare head circumference measurements in a cohort of Inuit children with the WHO charts.

Methods

We extracted head circumference data from a previous retrospective cohort study where, with Inuit partnership, we reviewed medical records of Inuit children (from birth to age 5 yr) born between Jan. 1, 2010, and Dec. 31, 2013, and residing in Nunavut. To create a cohort of Inuit children similar to the cohorts used in the development of the WHO growth charts, we excluded children with preterm birth, documented neurologic or genetic disease, and most congenital anomalies. We compared head circumference values with the 2007 WHO charts using centiles estimated with a generalized additive model.

Results

We analyzed records of 1960 Inuit children (8866 data points), of whom 993 (50.8%) were female. Most data were from ages 0 to 36 months. At all age points, we found that the study cohort had statistically significantly larger head circumferences than WHO medians, with most z scores for head circumference measurements among Inuit children falling 0.5–1 standard deviations above the WHO reference (p < 0.001). At age 12 months, median head circumferences were 1.3 cm and 1.5 cm larger for male and female Inuit children, respectively. Using WHO growth curves, macrocephaly was significantly overdiagnosed and microcephaly was underdiagnosed (p < 0.001).

Interpretation

Our results support the observation that Inuit children from Nunavut have larger head circumferences than other populations, and use of the WHO charts may thus lead to overdiagnosis of macrocephaly and underdiagnosis of microcephaly. Population-specific growth curves for Inuit children should be considered to provide timely and appropriate diagnoses of microcephaly and avoid overinvestigation of macrocephaly.

Growth curves are important tools used by health care professionals to monitor growth in children by comparing their height, weight, and head circumference to normative data created from a large cohort of children. Monitoring with growth curves may allow early identification of potential medical concerns. In Canada, the 2007 World Health Organization (WHO) Growth References are currently used.¹ These were created using measurements from children in 6 different countries who generally had no health, environmental, or economic constraints on growth.²

Standardized head circumference growth charts may not reflect local populations.^3–5 For instance, in a Cree population in northern Quebec, Willow and colleagues⁶ reported that the average newborn’s head circumference was 1.5 cm and 1.6 cm larger than the WHO references for males and females, respectively. This may lead to errors in recognizing microcephaly or macrocephaly, defined as a head circumference that is 2 standard deviations (SDs) below (< 3rd percentile) or above (> 97th percentile) the mean, respectively. ^7,8 A 2018 report from Greenland found increased rates of macrocephaly (7.4%–11.3%) and decreased rates of microcephaly (0%–1.5%) among local children.⁹ This study did not exclude non-Inuit children; around 89% of Greenland inhabitants are Inuit.¹⁰ Anecdotally, pediatric neurologists (S.V., K.J.) working at the Children’s Hospital of Eastern Ontario, which provides medical services for many children from northern Canada, have observed similar trends among Inuit children in their clinical practices. Clinicians (L.A., A.M., S.V., K.J.) with experience providing medical care for Inuit children have noted that those with syndromes typically associated with microcephaly may still be regarded to have normal head circumferences according to current reference ranges.

Microcephaly has many potential causes. Diagnostic considerations include craniosynostosis, exposure to prenatal infections or toxins, brain injury such as a stroke or hypoxic–ischemic encephalopathy, or genetic or metabolic conditions.⁷ Because microcephaly is associated with poor neurodevelopmental outcomes — around 90% of those affected have an intellectual disability⁷ — it is important to identify and investigate microcephaly appropriately. The reported incidence of microcephaly varies between 1.3 and 150 cases per 100 000 births.⁷ Similarly, macrocephaly has a variety of causes, including vascular abnormalities, hydrocephalus, intracranial masses, skeletal dysplasias, autism, and genetic disorders.⁸ Macrocephaly has been reported to affect 2%–5% of children,⁸ with pathology found in 17%–41% of children with macrocephaly.^11,12

Because of the clinical implications of diagnoses associated with micro- or macrocephaly, affected children often require specialist assessment and advanced neuroimaging with magnetic resonance imaging (MRI), which, in Canada, is primarily available in major centres. For Inuit families living in Nunavut, this means multi-day travel to unfamiliar settings, often involving multiple flights, disruptions, and delays, as well as substantial childcare and financial burdens (e.g., arranging childcare for other children, taking time off work). Furthermore, since Inuktitut is the primary language spoken in Nunavut,¹³ children referred for assessment and their families may not speak the language (English or French) spoken in the assessment centres, resulting in additional stress.

Given these considerations, it is important to determine whether standard reference growth curves may be underdiagnosing microcephaly and overdiagnosing macrocephaly among Inuit children. Thus, we sought to compare head circumference measurements in a cohort of healthy Inuit children from Nunavut with the WHO growth charts and assess differences in diagnosis of micro- or macrocephaly.

Methods

Study design

We conducted a retrospective cohort study to compare the head circumferences of Inuit children in Nunavut with WHO head circumference charts. We obtained relevant data via a data sharing agreement from a larger study, in which researchers conducted an extensive chart review to assess the health outcomes of Inuit children, with a particular focus on the effect of a common genetic variant (CPT1A p.P479L).¹⁴ Several authors of this study (S.C., S.E., G.H.A., A.M., L.A.) participated in the original study.

We used the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement checklist to report this study.¹⁵

Setting

In Canada, Inuit live primarily in Inuit Nunangat (Inuit homeland), which includes the Inuvialuit Settlement Region (Northwest Territories), Nunavik (Quebec), Nunatsiavut (Labrador), and Nunavut. According to the 2021 Census around 70 540 Inuit live in Canada, which is 0.19% of the country’s population.¹⁶ We focused on Inuit children living in Nunavut, where around 43.7% of Inuit in Canada live.¹³ Nunavummiut (residents of Nunavut) live in 25 communities along the arctic coastline, and around 84% identify as Inuit.¹³ Health care resources are limited in Nunavut because of its remote northern location and patients must be flown to other Canadian centres for specialized or tertiary care, including for investigations and assessment by medical specialists.

Cultural context

The original study¹⁴ was developed and conducted in partnership with Nunavut Tunngavik Inc., an organization that represents Inuit under the Nunavut Agreement and ensures that the federal and territorial governments fulfill their obligations;¹⁷ the Qaujigiartiit Health Research Centre; and the Government of Nunavut Department of Health. These entities contributed to the design of the original study (including collection of relevant variables), ensured that Inuit self-determination in research was respected, and were in agreement that the current project would support optimal outcomes for Inuit families and communities, principles that were later defined in the National Inuit Strategy on Research.¹⁸

The authors involved in the current study have clinical experience providing medical care for Inuit or Nunavummiut (K.J., S.V., A.M., L.A.), have participated in research involving Inuit or Nunavummiut (S.C., L.A., A.M., G.H.A., A.C., J.A., M.D., S.K.), or are Inuit (S.E., J.A.). Several authors (K.J., S.C., A.M., M.D., S.K.) have completed training on Inuit cultural competency or Indigenous cultural safety. The results of this study and the manuscript have been shared and approved by our Nunavut and Inuit partners. Further details on our social positions as it pertains to this study are included in Appendix 1, available at www.cmaj.ca/lookup/doi/10.1503/cmaj.230905/tab-related-content.

Study population

The original retrospective cohort study used data from a review of clinical charts for Inuit children (from birth to 5 years of age), born between Jan. 1, 2010, and Dec. 31, 2013, in Nunavut, as well as their birthing parent.¹⁴ All communities with more than 20 births per year were included, representing 18 of the 25 communities in the territory. We classified children as Inuit if either parent was identified as Inuit on the prenatal record (Appendix 1, Supplemental Figure 1).

For the current study, we constructed a data set derived from the original study, excluding preterm births (< 37 wk gestation), and children with documented neurologic or genetic conditions, as well as most major congenital anomalies (Appendix 1). We included children with congenital anomalies if these would not be expected to influence head circumference (e.g., isolated imperforate anus, hypospadias). Table 1 compares our eligibility criteria with those of the WHO Multicentre Growth Reference Study.¹⁹

Table 1:

Eligibility criteria for World Health Organization (WHO) Multicentre Growth Reference Study²⁰ versus Inuit cohort

Study	Criteria
WHO Multicentre Growth Reference Study 20	No health, environmental, or economic constraints on growth Mother willing to follow feeding recommendations Term birth: gestational age ≥ 37 completed weeks (259 d) and < 42 completed weeks (294 d) Single birth Absence of significant morbidity Nonsmoking mother (before and after delivery)
Inuit cohort of current study	Living in Nunavut, Canada, with accessible health and social service supports, living in communities with more than 20 births a year Term birth: ≥ 37 completed weeks Singleton or twin birth No neurologic or genetic conditions, or congenital anomalies that could affect head circumference At least 1 parent self-identified as Inuit at prenatal assessment

Data sources and collection

For the original study, data were collected from review of clinical charts located at community health centres, the Qikiqtani General Hospital in Iqaluit, and the Iqaluit Public Health office.¹⁴ The charts were primarily paper, with the exception of some charts in Iqaluit, which were electronic health records. Data collected included birth data (gestational age, birth weight, length, and head circumference), and all health information up to 5 years of age, including well-baby and child visits, visits to health centres, and admissions to hospital for congenital anomalies, illnesses, and other major medical diagnoses. This information was documented in the child’s chart primarily by community and public health nurses, but sometimes by physicians.

Within the health charts were well-baby forms that were originally developed for the Nutaqqavut (Our Children) Health Information system, which was in effect from 2011 to 2014.²⁰ These forms and other well-baby forms were derived from the Rourke Baby Record.²¹ Head circumference was recorded at birth and at visits at ages 2 months, 6 months, 12 months, 2–3 years, and 4–5 years; additional head circumference measurements were collected if available. Head circumference measurements were completed by trained public health and community health nurses within the communities of residence. Data on food insecurity and breastfeeding were obtained from well-baby forms (Appendix 1).

Maternal characteristics such as education, smoking, and other exposures and illnesses in pregnancy were collected via standardized prenatal forms.^22,23 Information regarding smoking was collected on the first prenatal visit.

Duplicate records were identified and removed by use of unique study numbers and birth date of the child. Data collection was carried out by a team of chart reviewers and, although it was not carried out in duplicate, intermittent quality checks were carried out (S.C.).¹⁴

Variables

Individual variables included in the original data sharing agreement and transferred for this current study included head circumference, sex, gestational age at birth, and age at measurements (by days of life). These were consistent with those used in the WHO head circumference growth charts.¹ Variables used for sensitivity analyses were food insecurity and smoking recorded at first prenatal visit.

We used aggregate descriptive information on maternal age, pregnancy complications (such as hypertension, diabetes, preeclampsia), singleton and twin gestations, birth weight, and breastfeeding status.

Statistical analysis

We summarized demographic data using descriptive statistics. For categorical variables, we reported frequencies and percentages, and for continuous variables, medians and SDs. We performed all analyses separately for males and females. If head circumference, sex, or age data were missing, the data point was not included.

We removed implausible head circumference measurements for age by first fitting generalized additive models for location, scale, and shape (GAMLSS), using Box–Cox t distribution and cubic smoothing splines for years of age, and then removing measurements that fell above the 75th percentile plus 5 times the interquartile range (IQR), and those that fell below the 25th percentile minus 5 times the IQR (Appendix 1).²⁴ We then refit the same model to obtain estimated centiles of head circumference for age.²⁴ We compared these centiles for head circumference for age with those from the WHO.¹ Model checking is further described in Appendix 1.

To compare head circumferences between the Nunavut Inuit cohort and WHO sample, we calculated z scores using the LMS method,²⁵ assuming that the head circumference measurements came from the same distribution. If this assumption was met, the distribution of the z scores for Inuit population in each age interval would be centred near 0. Therefore, the z score represented the distance and direction of observed Inuit head circumference from the WHO median. We used a 1-sample t test to determine whether the mean z score for Inuit children differed significantly from 0. We separated medians from WHO data into time intervals (i.e., birth, 0–2 months, and so on); we compared the median Inuit head circumference in each time interval with the WHO median value at the midpoint of the interval (e.g., at 1 month for data from the 0–2-month interval).

We compared proportions of Inuit children measured to have macrocephaly (i.e., above the 97th percentile) from our model and those above the 97th percentile according to WHO tables. Similarly, for microcephaly, we compared proportions of Inuit children below the 3rd percentile from our model and those below the 3rd percentile from WHO tables. For the purposes of this analysis, children would be considered to have microcephaly or macrocephaly if any one of their measurements were below the 3rd percentile or above the 97th percentile, respectively. We calculated adjusted Wald confidence intervals for the difference in dependent proportions and associated p values.²⁶

We tested several smoothing parameters before selecting the final model (Appendix 1).

Sensitivity analysis

We conducted a sensitivity analysis including children without exposure to maternal smoking, food insecurity, or both. We compared z scores for these subpopulations to the WHO references.

Ethics approval

Ethics approval for the original study was granted by the University of British Columbia (UBC), University of Victoria, and University of Manitoba Research Ethics Boards (REBs), and a research license was granted by the Nunavut Research Institute. For the current study, ethics approval was granted by the Children’s Hospital of Eastern Ontario REB and as an amendment to the UBC REB approval. The Nunavut Research Institute provided a license waiver. All research partners agreed to the use of the data for the current study.

Results

Of 2522 charts reviewed, we excluded 562 from analysis (Appendix 1, Supplemental Table 1). We analyzed health care information from 1960 Inuit children, comprising 8866 data points (Figure 1). Maternal and newborn characteristics for the study population are presented in Table 2. After removing outliers for head circumference based on estimated quartiles, 1956 children, comprising 8789 (n = 4414 female, n = 4375 male) data points remained (Figure 1 and Table 3). The median number of data points per individual was 4 (range 1–11). Scatterplots of data points including outliers are shown in Appendix 1, Supplemental Figure 5.

Figure 1:

Study flowchart. *May be excluded for more than 1 reason. Details regarding exclusions can be found in Appendix 1, Supplemental Table . See Related Content tab for accessible version. †Deaths within the original cohort of live births documented in community charts. Note: HC = head circumference.

Table 2:

Maternal and newborn characteristics for the study cohort of Inuit children born in Nunavut, 2010–2013

Characteristic	No. (%) of children* n = 1960
Maternal characteristics
Mean maternal age, yr	24.03
Prenatal anemia†	645 (40.6)
Maternal diabetes (pre-existing or gestational)	42 (2.4)
Maternal hypertension (without preeclampsia)	82 (4.6)
Preeclampsia or eclampsia	26 (1.3)
Prenatal smoking, no. of cigarettes/d‡	1379 (83.6)
Light (1–4)	311 (18.8)
Moderate (5–10)	646 (39.2)
Heavy (> 10)	210 (12.7)
Amount unknown	212 (12.8)
Pregnancy characteristics
Twin births	6§ (0.3)
Post-term births (≥ 42 wk)	12 (0.6)
Neonatal characteristics
Females	993 (50.8)
Males	963 (48.2)
Breastfed¶
Initiated	1391 (75)
Continued at 6 mo	703 (38)
Low birth weight** (< 2500 g)	32 (1.7)
Females, n = 979	19 (1.9)
Males, n = 954	13 (1.4)
Birth weight**, g, mean ± SD
Females	3436 ± 462
Males	3538 ± 489
Congenital anomalies††	19 (0.1)
Food insecurity‡‡	951 (53.5)

^*Unless indicated otherwise.

^†Data available for 1588 (81.0%) children.

^‡Data available for 1650 (84.2%) children.

^§Number of children born from twin gestational pregnancies (i.e., 3 twin pregnancies with 6 births).

^¶Data available for 1850 (94.4%) children.

^**Data available for 1933 (98.6%) children (979 females and 954 males).

^††Congenital anomalies such as imperforate anus, branchial cleft cyst, hypospadias, polydactyly, etc., conditions not expected to affect head circumference (Appendix 1).

^‡‡Data available for 1777 (97.7%) children. As noted on well-baby forms with the question “since your baby was born/your last visit, were there times when the food for you and your family just did not last and there was no money to buy enough food?” Answers of “Often” or “Sometimes” were coded as Yes.

Table 3:

Data points for head circumference, by time point, for the study cohort of Inuit children born in Nunavut, 2010–2013

Variable	No. of data points*
	Total n = 8789	Male n = 4375	Female n = 4414
Time point, mo
0†	1848	905	943
> 0 to < 2	499	251	248
2 to < 4	982	488	494
4 to < 6	598	302	296
6 to < 8	953	468	485
8 to < 10	518	246	272
10 to < 12	171	92	79
12 to < 18	1230	621	609
18 to < 24	585	277	308
24 to < 30	542	278	264
30 to < 36	99	57	42
36 to < 48	42	25	17
≥ 48	102	64	38
No. of measurements per person, median (IQR)	4 (3–5)	4 (3–5)	4 (3–5)
Range	1–11	1–10	1–11

Note: IQR = interquartile range.

^*Unless indicated otherwise.

^†As documented on birth record.

Median head circumference

Median head circumferences in the study cohort were significantly different than the WHO reference values at all time points for both males and females (Table 4). Most of the z scores for Inuit head circumference measurements fell between 0.5 and 1 SD above the WHO reference (p < 0.001) (Figure 2). At birth, males and females had absolute differences of 0.5 cm and 0.6 cm from reference values, respectively. The maximum absolute difference in head circumference was 1.9 cm for males and 2.0 cm for females. Growth charts using our data were overlaid on WHO growth charts to visually analyze the differences (Figure 3).

Table 4:

Median head circumference for the study cohort of Inuit children born in Nunavut, 2010–2013, and World Health Organization (WHO) reference data^28* by age

Age, mo	Males			Females
	Inuit median, cm	WHO median, cm	Difference*, cm	Inuit median, cm	WHO median, cm	Difference*, cm
0†	35.0	34.5	0.5	34.5	33.9	0.6
0–2	38.0	37.3	0.7	37.5	36.5	1.0
2–4	40.5	39.8	0.7	39.5	38.9	0.6
4–6	43.0	42.1	0.9	42.2	41.0	1.2
6–8	45.0	43.7	1.3	44.0	42.5	1.5
8–10	46.5	44.8	1.8	45.5	43.6	1.9
10–12	47.0	45.6	1.4	46.0	44.4	1.6
12–18	48.0	46.7	1.3	47.0	45.5	1.5
18–24	49.5	47.8	1.7	48.5	46.6	1.8
24–30	50.5	48.6	1.9	49.5	47.5	2.0
30–36	51.0	49.2	1.8	49.7	48.2	1.5
36–48	51.5	49.8	1.7	50.3	48.9	1.4

^*p < 0.001 for all comparisons.

^†As documented on birth record.

Figure 2:

Distribution of z scores of head circumference for a cohort of (A) male and (B) female Inuit children born in Nunavut, 2010–2013, compared with World Health Organization (WHO) reference data.²⁸ The blue line represents a reference line and indicates no difference between Inuit and WHO measurements of head circumference. Each rectangle represents the middle 50% of z scores (25th to 75th percentile), with the bolded line indicating the median. Width of box represents number of data points. Supporting data are presented in Table 4. Note: IQR = interquartile range.

Figure 3:

Percentile curves for head circumference in a cohort of (A) male and (B) female Inuit children born in Nunavut, 2010–2013, compared with the 3rd, 50th, and 97th percentiles in World Health Organization (WHO) reference data.²⁸ Dots represent data points. Dashed and solid lines indicate 97th (top), 50th (middle), and 3rd (bottom) percentiles in the WHO reference data and Inuit data, respectively. Supporting data are presented in Table 4 and Appendix 1, Supplemental Table 3.

Macrocephaly and microcephaly

Rates of macrocephaly and microcephaly were both statistically different than expected (p < 0.001). Proportions of those with diagnoses of macro- or microcephaly based on Inuit versus WHO data were significantly different (p < 0.001) for males and females (Figure 4 and Appendix 1, Supplemental Table 4). For macrocephaly, around 25%–30% fewer males and 15%–20% fewer females had diagnoses when using Inuit-specific percentiles (Figure 4); in our cohort, this would mean 651 fewer children with macrocephaly, with 245 cases diagnosed by both charts. Appendix 1, Supplemental Table 3, provides the absolute values for 3rd and 97th percentile measurements for head circumference at each time point.

Figure 4:

Difference between proportion of children with diagnoses of microcephaly (below the 3rd percentile) among (A) males and (B) females, and macrocephaly (above the 97th percentile) among (C) males and (D) females, based on World Health Organization (WHO) reference data²⁸ or Inuit data. The blue line indicates no difference when using WHO or Inuit-based percentiles. Black circles with 95% confidence intervals indicate the difference in percentage of children with diagnoses of microcephaly or macrocephaly from Inuit-based data instead of WHO percentiles (above the blue line indicates that more Inuit children would have diagnoses; below the blue line indicates that fewer Inuit patients would have diagnoses, compared with WHO standards). Confidence intervals have been corrected for dependent data (modified Wald). Supporting data are presented in Appendix 1, Supplemental Table 2.

Around 2% more children were considered microcephalic when using percentiles derived from the Inuit population (Figure 4). Of the total sample of 1956 children, this translated to an additional 79 children with diagnoses of microcephaly, in addition to the 145 children with diagnoses according to both charts. Four children had diagnoses of microcephaly using the WHO data but not the Inuit-specific chart, occurring at 0–2 months among males, for whom the 3rd percentile was 35.2 cm in WHO data and 35.1 cm in Inuit-specific data (Appendix 1, Supplemental Table 2).

Sensitivity analysis

Of 1960 children, 271 were not exposed to maternal smoke and 950 did not report food insecurity. After excluding those with either exposure (n = 1824), 136 children with 595 data points remained. The overall pattern of results for those without exposure to maternal smoke, without food insecurity, or both, was the same as in the main analyses (Appendix 1, Supplemental Figures 3, 4, and 2, respectively).

Interpretation

Our data provide evidence that head circumference measurements among Inuit children in Nunavut are larger than WHO standards. We observed larger median head circumference values, distinct patterns of growth curves, higher rates of children with head circumferences greater than the 97th percentile (macrocephaly), and lower rates of children with measurements below the 3rd percentile (microcephaly).

Clinicians must be able to identify children with potential medical issues appropriately, without underdiagnosis or overdiagnosis at the extremes of head circumference measurements. We have shown that the normal growth curves for Inuit children are not the same as the WHO standards. When using WHO standards for head circumference in children, clinicians should be aware that 20%–30% of the general Inuit population of young children may fall into the category of having macrocephaly without actual pathology. Conversely, when using WHO standards, fewer Inuit children may be recognized to be affected with microcephaly, and medical evaluation of underlying medical conditions such as a congenital infection or brain malformation may not be pursued or may be delayed.

Overdiagnosis of macrocephaly may lead to unnecessary investigations, such as computed tomography or MRI brain scans, or consultations with specialists, requiring costly and burdensome travel for families to tertiary care centres. Given the remote location of Nunavut, travelling for investigations and consultations results in major family and financial upheaval involving multi-day travel without family and community supports, away from culture and language (as Inuktitut is the most common language spoken in Nunavut).¹³ This unnecessary overinvestigation perpetuates a system that continues to bring harm to Inuit people, given the historical context of racism, mistreatment, and experimentation by settler health care workers.

Variance from WHO reference charts has been shown in other populations. In a 2014 systematic review, significant global variability in head circumference ranges was noted, with outlying groups including European, Turkish, Asian Indian, Australian Aboriginal, Canadian Cree, Japanese, and Pacific Island populations.²⁷ The authors noted that 18 of the 26 studies reporting local head circumference measurements showed means that were more than 0.5 SD higher than the published WHO means, as well as higher rates of macrocephaly and lower rates of microcephaly.

By contrast, the Nellhaus reference chart of head circumference measurements for children from birth to age 18 years reported no racial or ethnic differences.²⁸ However, the mean head circumference for Alaskan Inuit and Yupik children at age 2 years was slightly larger than the other cohorts. With the data available at the time, this difference was reported as “considerably less than one standard deviation.”²⁸ The United States Centers for Disease Control and Prevention (CDC) produce a head circumference growth chart, but ethnic or racial data were too limited to make comparisons.²⁹ It was not reported in these studies as to how ethnicity was determined (Appendix 1, Supplemental Table 4).

Given that average growth parameters can change over time, France has recently updated its head circumference reference chart with local data, using a big-data approach enabled by electronic medical records that capture routine clinical data.³⁰ Going forward, Canada may benefit from taking a similar approach by producing local growth charts, either on a national scale or with Inuit-specific charts. We provide Inuit-specific charts based on our data in Appendix 1, Supplemental Figures 6A and 6B. Further evaluation is also warranted for the Cree people in northern Quebec, given similar data on head circumference measurements for this population.⁶

The strengths of this study include the large data set for population size that was available for analysis, accounting for nearly 25% of Inuit children in Canada at the time of the study.³¹ This allowed clear statistical and clinical evidence for differences in mean head circumference values. In addition, this study relied entirely on self-declaration of ethnicity.

We acknowledge controversy related to including race and ethnicity as a variable in medical calculations and assessments, given changing definitions for race and ethnicity and the current understanding that race is a social construct rather than a biologic construct.³² There has also been a move toward reducing or eliminating racial qualifiers in medical calculations. ³³ However, given the substantial burden on Inuit families in Canada for evaluation of macrocephaly, the risk of not diagnosing microcephaly, and the desire for more precise, population-specific data arising from Inuit associations and affiliated organizations, the results of our study are likely to have an important, positive effect for young Inuit children and their families.

Although this study addressed population-specific norms for head circumference, given the clinical actionability of patients with head circumference measurements outside of population norms, future research on height and weight are also indicated. It is of note that although there is some correlation between head circumference and other growth parameters, head circumference has been shown to follow a growth pattern independent of height and weight.³⁴

Limitations

The data were collected from charts, as recorded from well-baby visits, with measurements carried out by trained community health and public health nurses. Data from chart reviews have inherent limitations, such as the potential for aberrant measurements, data entry error, or the development of relevant medical conditions later in life. We are not aware of specific chart audits carried out within the health care delivery system to determine the accuracy of the data entered in the chart. Each chart was reviewed by 1 individual; however, intermittent data checks were carried out by the supervising manager (S.C.).

Although the chart review was extensive, around 10% of the total number of charts of children expected to be living within Nunavut during that time period were not available for review, including those from the 7 communities where chart review did not occur. They were excluded because of logistical constraints and the low number of births in these communities (< 20 births per year). There were also rare circumstances where charts were not available for review while the abstractors were working within the communities.

Our data were from a cohort of children born in 2010–2013, meaning that the data are more than 10 years old; however, these data are more recent than the data from the cohorts used to develop the WHO charts (published 2007), CDC charts (published 2000), and Nellhaus charts (published 1968).^1,28,35

We compared our sampling criteria with the WHO eligibility criteria. ¹⁹ The WHO included only singleton births, whereas we did not exclude multiple births if they met the other criteria. This accounted for only 0.3% of children. The WHO also excluded those born after 42 weeks’ gestation, while we did not (accounting for 0.6% of our cohort). We did not exclude children with exposure to maternal smoke or food insecurity, but we did a sensitivity analysis excluding these exposures, and head circumference results were similar. Given that our results showed larger head circumferences than reported in WHO cohorts, we reviewed rates of maternal diabetes, as this condition could result in macrosomia, which could confound our results. Rates of maternal diabetes were lower in the Nunavut population than the Canadian average; however, this was not discussed in the WHO Multicentre Growth Reference Study. Thus, although we had fewer exclusion criteria than the WHO, our population represented the generally healthy local population of interest. Furthermore, smoking has been, if anything, associated with smaller head size.³⁶ Therefore, these factors were unlikely to falsely elevate the number of children with clinically larger head sizes in this population. Food insecurity has not been clearly associated with smaller head circumference.³⁷ The WHO head circumference charts were made with a total of 27 339 data points (13 541 male, 13 798 female), while our data set included 8789 data points (4414 female, 4375 male). Although the larger WHO data set may allow more precise determination of head circumference percentiles, considering it included data from 6 different countries, our sample of 1 small subpopulation within Canada offers a sizeable comparison.

Most data in our cohort were for children younger than 3 years, so our most robust results are limited to this age group. We used all data points available (between 1 and 11 per child). More data points could suggest a child was being monitored for a medical concern; however, given our exclusion criteria (exclusion of children with known neurologic or genetic conditions and most major congenital anomalies), we are confident that the reason for increased monitoring would not be for conditions associated with alterations in head circumference. In addition, our exclusion criteria likely reduced the number of data points at the extremes of percentiles (> 97th percentile and < 3rd percentile). We may not have identified other confounders, as growth parameters are affected by complex contributing factors. We could not determine whether macrocephaly or microcephaly were associated with pathology in this study.

Conclusion

We have provided strong evidence that head circumference measurements among Inuit children in Nunavut diverge from the standard WHO curves. This is an important clinical concern, as there is risk of misclassification of macrocephaly and microcephaly, resulting in over- or underinvestigation of potential underlying pathology with substantial medical and social consequences. Our findings likely have implications for other Inuit populations in Canada and the circumpolar regions. The implementation of growth curves relevant to this population would necessitate ongoing discussion with Inuit organizations, health care professionals, and public health officials in these regions.