When monitoring the growth and development of neonatal HC, the primary concern is to use the best tools¹⁰. There is a lot of discussion recently concerning the choice of whether regional or global, age and sex-specific growth references or growth standards should be used for different populations^11,12,13. Our study revealed that in late preterm and term periods, with a typically smallest neonatal head circumference (HC) variability within a population, the differences between populations are the most pronounced (Table 2). The differences between the findings of the studies examined increase with the increasing GA, and particularly starting from the late preterm period, and especially, in the term newborns. In the present Lithuanian study, the variation of the mean HC in extremely, moderate to late preterm newborns HC was < 1 cm, in term newborns—> 1 cm compared to IG-21 (Table 1). In extremely preterm gestations, the means of HC varies within most studies^25,26,27,28 less than 0.5 cm compared to Lithuanian. However, according to some studies, in later gestations and in term newborns, the differences between populations in HC increase to more than 1 cm. Most of the similarities were found between Lithuanian and Finnish neonatal HC, the biggest differences – between Lithuania and Indonesia (Table 2).

Analysing the variability of HC with regard to gestational age within and between populations, the coefficient of variation (CV) was examined. According to different studies^25,26,27,28, the CV of HC in every population varies within a very narrow range, but is the highest in extremely preterm gestations, however, within the population, it decreases together with the increasing gestational age, same as the standard deviation (SD) (Tables 1 and 2). Hence, the closer to term, the narrower was the variability of the population’s neonatal HC. The CV of HC is higher in extremely preterm periods, but HC means and extremes appear to be very similar in different populations. We presume that in early gestation there is no need to strictly set head parameters according to the mother’s pelvis size, hence, greater biological variation is allowed, which is similar in most populations. On the other hand, the CV decreases with the increasing gestational age, but the means and marginal HC variants move according to a population-specific direction which is highly dependent on maternal size, particularly height and pelvic size²⁹. Here, the size of the neonatal head seems to be maximally adapted to maternal pelvic size. These considerations support the idea that head circumference is strongly anthropometrically limited by the maternal bony pelvis—“evolutionary bottleneck”, as named by M. Odent²⁰.

As the shape of the human pelvis is often interpreted as an evolutionary compromise between bipedal locomotion and childbirth of a highly encephalized neonate³⁰, HC is expected to be more strongly genetically determined and anthropometrically limited by the indices of the bony birth canal. Even though the newborn HC should be less influenced by internal or external factors than birth weight or length, many studies have raised the discussion on the complex interaction between the intrinsic and extrinsic factors in the development of neonatal HC^31,32. Furthermore, females with a large head, who are likely to give birth to neonates with a large head, were found to possess birth canals that are shaped to better accommodate large-headed neonates²⁹. Moreover, it is already known that variation in the shape of the female pelvis is significantly geographically structured³³.

What is more, the pelvis shape was found to be significantly associated with the stature for taller women having a more oval pelvic inlet and better accommodating a larger foetal head²⁹. In the study of R. G. Tague³⁴ femoral length/stature in females showed a significant, positive partial correlation with the anteroposterior diameter and shape of the pelvic inlet. A recent Swedish study³⁵ proved this relationship from the clinical point of view reporting decreasing risk of caesarean section (CS) with increasing maternal height after adjustment for maternal age, BMI, gestational age, parity, high birth weight and country of birth. With average Swedish women’s height of 166.1 cm, maternal height of 178–179 cm was associated with the lowest risk of CS (OR = 0.76, 95% CI 0.71–0.81), whereas height below 160 cm explained 7% of CS cases³⁵. It is worth mentioning that according to the NCD Risk Factor Collaboration³⁶, Lithuanian women are among the tallest women in the world with an average height of 167.6 cm. The comparative results of the female average height reflect the differences found between the mean neonatal HC from different populations, as shown in Table 2. Finnish women with an average height of 166.5 cm are closest to Lithuanians, followed by Canadians (164.7 cm), Chinese (163.5 cm), and finally Indonesians (154.4 cm)³⁶. This supports the previous study’s findings²⁹ that perhaps maternal height is linked to pelvic size, particularly the size of the birth canal, and through that to the neonatal HC. This possible relationship between neonatal HC (cm) at term (40 weeks of gestation) and average women’s height across compared countries is presented in Fig. 3 and compiled after^{26,27,28,29,35}.

With regard to these findings, scientists debate the appropriateness of growth standards vs. references, regional vs. global for proper evaluation of growth and development of neonatal HC. In our study, most of similarities with global study of IG-21 were disclosed in cut-off points for the lowest percentiles in extremely preterm newborns, apart from that, extremely preterm newborns (especially girls) had more similarities not only in the third, but also in the 50th and 97th percentiles of HC. The Brazilian study³⁷ revealed a similar pattern and found the trajectory of the third percentile parallel with the IG-21 study until the term period. The sample size of the IG-21¹⁹ reference was only modest for < 37 weeks gestation, and the later study on very preterm neonates ’should be interpreted with caution given the small sample size’²⁴. This may explain a “wave” at 33 gestational weeks observed in the percentile curves of IG-21 (Figs. 1 and 2). Thus, although IG-21 facilitates the evaluation of the main HC percentiles for extremely preterm newborns and might serve as cut-off points for the pathological microcephaly in preterm newborns of different populations, it should be considered with caution to be confidently used as a global standard at early gestations.

As for the other extreme, the 97th percentile, above which infants would be diagnosed with macrocephaly, a large gap between the curves of both studies of more than 1 cm from late preterm to post-term was detected which could lead to an overestimation of macrocephaly in our cohort. If we compared our results with the HC curves provided by the Centers for Disease Control and Prevention (CDC)³⁸, the gap would be smaller. In line with other studies³⁹ evaluating the influence of growth curves used for the distribution of HC, our study also claims that the important consequences could have been triggered by the percentile misclassification. Therefore, from the standpoint of clinical practice, to predict the course of HC higher percentiles in moderate or late preterm periods and, especially, in Lithuanian term newborns, regional standards should be used.

Accordingly, the question has been raised by scientists whether children’s growth references should be global, or specific to different populations: ’it has become apparent that a single “global” reference fails adequately to mirror the diversity in human growth’¹⁷. Human growth is determined by inherited factors, and the significant variability of foetal growth in utero between ethnic groups supports this statement³¹. Therefore, the IG-21 project charts based on the idea that foetuses, infants, and children grow similarly all over the world under ideal nutritional, environmental, psychological living conditions have been widely discussed. A number of studies^{32,37,40,41,42,43,44,45,46} recently have compared their foetal and neonatal national growth references with the IG-21 study that was recently published, and obtained diverse results. Some studies did not find appreciable differences with IG-21 for newborn HC⁴⁰ or a statistically significant difference was observed only of female HC in the 97th percentile³². While others determined that IG-21 standards for foetuses⁴⁷ were found to be unrepresentative for regional populations leading to considerable overdiagnosis of foetal microcephaly or misclassification of infant birth size^{37,41,42,43,44,45,46} and a conclusion that regional validation was needed prior to the implementation of IG-21. We found that global standards like INTERGROWTH-21^st might facilitate the evaluation of neonatal head circumference in early gestations, while in later gestations, the specific features of neonatal head circumference of a particular population tend to be more precisely represented by regional standards.

Therefore, we suggest taking into consideration the regional standards for neonatal head circumference in order to better evaluate a possible clinical pathology. It is important to stress that over the process of evolution, neonatal body size and head circumference have adapted to the mother’s body size, especially her pelvis, as a result of diverse adaptation mechanisms common to different populations in different geographical areas and under different living conditions.