Dental biorhythm is associated with adolescent weight gain.

Background: Evidence of a long-period biological rhythm present in mammalian hard tissue relates to species average body mass. Studies have just begun to investigate the role of this biorhythm in human physiology.
Methods: The biorhythm is calculated from naturally exfoliated primary molars for 61 adolescents. We determine if the timing relates to longitudinal measures of their weight, height, lower leg length and body mass collected over 14 months between September 2019 to October 2020. We use univariate and multivariate statistical analyses to isolate and identify relationships with the biorhythm.
Results: Participants with a faster biorhythm typically weigh less each month and gain significantly less weight and mass over 14-months, relative to those with a slower biorhythm. The biorhythm relates to sex differences in weight gain. Conclusions: We identify a previously unknown factor that associates with the rapid change in body size that accompanies human adolescence. Our findings provide a basis from which to explore novel relationships between the biorhythm and weight-related health risks.

Introduction

Human adolescence is a period of rapid change in body size following the onset of puberty[1]. Sex-specific increases in lean muscle, bone mass, stature, and the amount and distribution of subcutaneous and total body fat[2-4] contribute to extensive gains in body size[2,5-8]. These shifts vary by the stage of puberty for males and females[9,10]. Adolescents can gain 8.3–9.0 kg a year[2,6] depending upon genetic[11-14] and environmental factors such as dietary habits[6] and activity levels[15,16].

The hypothalamus plays a pivotal role in the pubertal transition. It is a region of the brain that stimulates the release of hormones and regulates food intake and energy expenditure. Under the influence of growth hormone and insulin-like growth factor-I in early adolescence, the steroid hormone oestradiol creates the main growth spurt responsible for body size changes in both sexes (testosterone is converted in males)[17,18]. The change in body size is mediated via the hypothalamic-pituitary-gonadal axis[17,18].

Life on earth is regulated by biological rhythms. Some are daily rhythms linked to the light-related circadian cycle[19,20]. Others are longer than 24-h with an infradian cycle. Evidence of the infradian cycle is present in a range of organisms (such as tree rings) and mammalian physiological systems[20-23]. For humans, a near 7-day rhythm has been identified in adult heart rate, core body temperature, excretion of metabolites and salt and blood pressure during pregnancy[24-29].

Accumulating evidence suggests an infradian biorhythm may act upon the mammalian hypothalamus to regulate cell growth and body mass[30,31]. Microscopic-layered structures of mammalian teeth retain evidence of this rhythm. In human tooth enamel, the rhythm is referred to as Retzius periodicity (RP)[32] (Fig. 1). RP forms through a circadian-like process, occurring with a repeat interval that can be measured through histology with a resolution of days. The rhythm is consistent within the permanent molars of individuals[33,34] that do not retain evidence of developmental stress[35]. RP relates to the period in which tooth enamel forms. For human primary molars, this is the two-year period following birth[36]. The human modal RP has a near 7-day cycle[33,34,37,38] but varies from five to 12 days[38,39] when compared between individuals. Higher RP values occurring over more days suggest a slow underlying biorhythm. Lower RP values suggest a faster biorhythm.

Fig. 1

Calculating RP-biorhythm in human primary molars.

a Sectioned primary (deciduous, ‘milk’) molar. An arrow pointing to lateral enamel with Retzius lines on the far right. b Thin section through enamel with Retzus lines to the right (lower white circle). The Upper white circle overlays tubular enamel rods, which formed as groups of cells (named ameloblasts) lay down new enamel as a tooth crown develops. c A record of ameloblast pathways are preserved in teeth as enamel rods. d Daily cross striations. Enamel deposition by ameloblasts is interrupted every 24-h producing regions along rods that have relatively less mineral. When prepared and examined under a microscope, these differences in mineralisation along rods appear as cross striations. This occurs because variation in mineralisation alters the refractive index of light transmitted by a microscope, producing the striations. Cross striations are used to calculate Retzius periodicity. e Black arrows point to Retzius lines in primary molar enamel. f White arrows point to cross striations and 6 days of enamel formation between two adjacent Retzius lines giving a Retzius periodicity of 6 days. Parts of Fig. 1 (part of the panel a and all of panel c) were created using a template from BioRender.com (2022).

Researchers during the 1990s[40,41] suggested variation in RP might relate to species-specific average body mass. Interspecific studies (meaning studies comparing different species) confirmed these observations revealing that, with exceptions[42,43], RP-biorhythm was higher (‘slower’) in larger-bodied living species, including anthropoids[30,31,44-46]. In these studies, biological pathways connecting RP and interspecific variation in body size were proposed. Larger-bodied species were suggested to attain their greater adult size through a slower biorhythm that produces slow growth rates over long periods of time, relative to the faster biorhythm of smaller-bodied species[30,31]. This pathway has emerged as a key hypothesis for advancing understanding of the evolution of primate life history[31].

Interspecific relationships are not always found within species[47]. But when the underlying cause is similar across different taxonomic levels, then similar biological relationships can be present within and between species[48]. The hypothalamus has a central role in human growth[17,18]. Our previous studies suggest aspects of human growth may relate to RP-biorhythm. Specifically, we have shown that the size of microscopic canals that house blood vessels in human adolescent ribs relates to RP[49]. Larger canals facilitate greater blood flow and nutrient transfer[50,51]. We observed higher RP values correspond with increased deposition of primary bone in humeri of young children[52]. These studies hint at a biorhythm underlying RP that influences rates of cell proliferation during the childhood growth years.

Studies of adult humans indicate taller adults tend to have lower RP values compared to shorter individuals[53-55]. The biorhythm appears to have a limited association with adult human weight[55]. Researchers utilised the height data from adults to hypothesise a biological pathway for human growth that differs from the interspecific pathway[30,31]. As the duration of human growth is constrained, relative to interspecific variation in growth periods, the biorhythm might accelerate to increase cell proliferation to achieve greater body size[30]. Thus, in contrast to the interspecific positive correlation between the duration of growth periods and body size and RP, the idea is that stature and RP should correlate negatively in humans. Currently, however, evidence of the biorhythm in relation to human growth[49,52] is limited.

Here, we calculate the biorhythm from primary molars in relation to weight gain for 61 children (average starting age = 10.33 years) from Dunedin, southern New Zealand, over a period of 14 months between September 2019 to October 2020. Adolescent weight is of particular interest because of the substantial gains during puberty that are driven by the hypothalamus. We demonstrate that adolescents with a faster biorhythm gain less weight over 14 months and have the smallest change in their body mass index (BMI) compared to adolescents with a slower biorhythm.

Methods

Participants, dental samples, study design and ethics

The 61 participants (n = 34 females and n = 27 males) were selected from a larger cohort that were part of the Biorhythm of Childhood Growth project. The BCG is an ongoing prospective cohort study that investigates childhood development in middle-income children from Southern New Zealand[56]. Participants attended primary schools at the start of the project and then intermediate schools (see acknowledgements) within Dunedin city, New Zealand. 49 participants were of New Zealand European ethnicity. Six participants were of mixed heritage, either New Zealand European/Māori or New Zealand European/Pasifika. Six participants were either Māori, Pasifika, Iranian or of mixed Swiss/Korean heritage.

Naturally exfoliated primary molars were collected from all BCG participants (n = 125 children) and n = 61 were randomly selected from these for the current analyses based upon histology criteria (see methods). RP was calculated for each participant directly from their naturally exfoliated primary molars, which was compared to measures of that individual’s weight and BMI. RP was calculated by one of us, GM, in the United Kingdom independently and blind to the weight and height data recorded in New Zealand by another author (SW).

We focused on primary molars only, as RP is a sequence for some individuals that can change between tooth types along the tooth row[33]. All deciduous molars, both maxillary and mandibular, were naturally exfoliated during the project. They were collected once a month during the monthly measurement of the growth variables. Molars with accentuated markings (also known as stress lines) were excluded as RP can sometimes change on either side of a stress marking[35].

Additional measures were incorporated into our study design so we could identify their effect on potential relationships between RP and weight gain. Adolescence typically commences in females (age 9 to 12 years) before males (age 11 to 14 years)[1,2]. Peak growth in height is greater for males but occurs sooner for females[57]. Because of these sex differences in the timing of adolescence, we expected females to gain more weight and height than males over the course of 14 months. If the biorhythm relates to adolescent weight/BMI gains, then there should be sex differences in these relationships.

Many factors influence body size during puberty. Body composition has a genetic component[11-14] and can be influenced by dietary habits[6], social environment, and variation in activity levels[15,16] related to seasons[5]. A recent study reports the effect of a Covid-19 national lockdown on adolescent BMI[57]. We, therefore, recorded the timing of maturation stages for participants in our study, modelled from longitudinal measurements of height and lower-leg length, and variation in these parameters and weight gain related to ancestry, seasons of the year and a Covid-19 lockdown that occurred unexpectedly between the end of March 2020 until the beginning of June when New Zealand returned to Level 1 (https://covid19.govt.nz/assets/resources/tables/COVID-19-Alert-Levels-summary-table.pdf).

Ethical approval for monthly measurements from participants and collection of primary molars was obtained from the University of Otago Human Ethics Committee (approval number H19/030). Research consultation with Māori was obtained from the Ngāi Tahu Research Consultation Committee. In New Zealand, research consultation with Māori is mandated in all areas of research that involve people of Māori descent. Informed consent was obtained from all participants and their parents or guardians. A list of participating schools in Dunedin is given in Acknowledgments.

Histology

Thin sections were created following standard procedures[39]. Teeth were embedded in resin (Buehler EpoxiCure®) and sectioned through the tip of the mesial cusp and dentin horn using a Buehler Isomet 1000 precision saw. Sections were fixed to glass microscope slides (Evo Stick® resin), ground (grit P400, P600, P1200) (Buehler® EcoMet 300), polished with a 0.3 µm aluminium oxide powder (Buehler® Micro-Polish II), cleaned in an ultrasonic bath, dehydrated in 95–100% ethanol, cleared (Histoclear®), and mounted with a coverslip (DPX®). Thin section thickness is determined by the visibility of incremental lines. Lines can become visible at different depths in thin sections of a primary molar from different individuals. Sections were examined using a high-resolution microscope (Olympus® BX53) and microscope camera (Olympus® DP25). Images were obtained and analysed in CELL® Live Biology imaging software.

Retzius periodicity data was recorded by GM in the United Kingdom, independently and blind to the New Zealand growth data. Each participant was selected for inclusion in the study if we were able to produce two matching RPs for their primary molars, either: (a) from the outer lateral enamel of each participant’s first and second primary molars or (b) from one single primary molar. Lateral enamel commences as the first Retzius line emerges on the outer enamel surface as perikymata (meaning, growth lines on the exterior rather than the interior of the tooth enamel).

We found no evidence that RP changed within an individual when compared between their primary molars, either in comparisons between mandibular and maxillary molars or first and second molars (Supplementary Table 1). This is consistent with findings for permanent molars[33]. Oblique thin sections were identified and removed from the study. Oblique sections can be easily identified from the morphology of the dentin horn together with the slope of the enamel buccal and lingual surfaces of the functional and guiding cusps.

RP was calculated in two standard ways. The number of daily cross striations was counted along a prism between two adjacent Retzius lines in lateral enamel at 200-400x magnifications (including the ocular magnification). When consecutive cross striations were not clearly visible between two Retzius lines, RP was calculated from local daily enamel secretion rates (DSRs) divided by prism lengths[45].

We had a good understanding of DSRs in primary molars of these New Zealand children[56]. Variation in DSRs was not a confounding factor in our calculation of RP as DSRs vary only slightly in the outer lateral enamel of primary molars of New Zealand European children[56]. DSRs were calculated by measuring along a prism across the span of six cross striations, which corresponds to 5 days of enamel formation (two adjacent cross striations = 24 h of enamel secretion) and dividing this measurement by five to get a daily mean DSR. This was repeated six times within the local enamel so that a grand mean DSR could be calculated. Following this first calculation, the distance between four to six adjacent Retzius lines was also measured, corresponding to three to five repeat intervals respectively, and divided by three or five. This distance between two adjacent Retzius lines was then divided by the grand mean DSR to yield an RP value.

Measurements of weight, height, maturation and body mass index

These were recorded by SW in Dunedin, independently and blind to the Retzius periodicity data that was generated in the United Kingdom. Height, weight and lower-leg length measurements were recorded from each child over a 14-month period between September 2019 to October 2020 during visits to the schools. Most measurements were taken about 4 weeks apart, excluding January 2020 during the school holiday and between March to early June 2020 during the national lockdown due to the onset of the COVID-19 pandemic. Standing height measurements were taken using a Seca 213 Stadiometer. Lower-leg length measurements were recorded three times per participant per visit, using a custom-made laser measuring device with the children in a standardised seated position. Weight was recorded on calibrated scales.

The maturity status of each participant was primarily estimated by modelling longitudinal measurements of their heights taken approximately once per month. Measurements were modelled using fixed bandwidth kernel weighted robust third-degree polynomial regression smoothing of heights on measurement dates[58]. Each individual was assigned one of four maturity scores based upon criteria involving the shape of individually modelled curves along with their sex and age-specific heights. Individuals who were relatively short for age and had not reached pre-spurt minimum height growth velocity were assigned a maturity score of 1 (‘pre’ in Table 1). Individuals who had reached pre-spurt minimum height growth velocity but who were not near peak height velocity were assigned a maturity score of 2 (‘early’). Those individuals who were very close to or who had just exceeded peak height growth velocity were assigned a maturity score of 3 (‘peak’). Individuals who had clearly exceeded peak height velocity and were approaching an upper asymptote were assigned a maturity score of 4 (‘late’). Individual maturity status was also estimated using the same approach but with longitudinal measures of lower-leg length. Results were very similar and are available from the authors.

Table 1

Descriptive statistics for RP-biorhythm and growth measures.

Participants		Retzius periodicity		Starting age	Maturation stagea				Gains			Ending	Gain
	n	mode	mean	yrs	1.pre	2.early	3.peak	4.late	Weightkg	Heightcm	Legcm	BMIpercentile	BMIkg/m2
All	61	6	7.26 ±1.31	10.33 ± 0.57	9	20	27	3	6.33 ± 2.79	6.92 ± 1.39	2.37 ± 0.57	69.06 ± 25.44	1.13 ± 1.04
Femaleb	34	8	7.50 ± 1.33	10.30  ± 0.59	1	2	26	3	6.97 ± 2.82	7.39 ± 1.55	2.41 ± 0.47	69.19 ± 25.64	1.31 ± 1.06
Male	27	6	6.96 ± 1.25	10.36  ± 0.56	8	18	1	0	5.56 ± 2.60	6.37 ± 0.92	2.32 ± 0.67	68.91 ± 25.69	0.95 ± 1.01

aDetermined from longitudinal measurements of height and weight modelled using fixed bandwidth kernel weighted robust third-degree polynomial regression smoothing.

bWe were unable to assign a maturation stage to two females.

BMI and BMI percentiles for a given age and sex were calculated using each participant’s birth date, sex, height, weight and date that the measurements were taken. These measurements were entered into the online calculator for New Zealand children provided by the New Zealand Ministry of Health.

Statistical analyses

Data were log-transformed. Pearson correlation coefficient was used to measure the strength of association between gains in weight gain and height, lower-leg length, starting age and maturation stage. The influence of starting age on the relationship between RP and weight gained over 14 months was assessed through partial correlations. Height and weight were compared between males and females with a two-tailed t-test. Weight was compared between females grouped by RP using a Kruskal–Wallis H with multiple comparisons. The relationship between RP and weight/gained over 14 months was modelled using quadratic regression with p values adjusted using a Bonferroni correction. We conducted further analyses using a Kruskal–Wallis H-test with multiple comparisons to analyse the rank order of RPs and weight/BMI gained when grouped by those with 6, 7 and 8 days, which were the largest sample sizes. A Chi-square test was used to determine if there was a relationship between participants with RPs of 5 or 6 days and a BMI of or greater than the 95th percentile compared to those with RPs of 7 or 8 days. Multivariate regression was undertaken to assess the relative strength of the effect of log-transformed weight, leg length and stature on the predictor variable RP, using standardised beta coefficients. We also examined the relative relationship of RP to total gains in log-transformed weight, height, and leg length, using standardised beta coefficients, just for females with maturation scores of three.

Table 2

Regression analyses of log-transformed gains in weight and body mass index and their association with log-transformed Retzius periodicity.

	Quadratic curve
RP in days vs:	Intercept	Slope	r	r2	p
Total weight gained in kg
Sept 2019 to Aug 2020	−5.530	7.593	0.492	0.243	0.012*
to Sep 2020	−2.614	6.796	0.498	0.248	0.007*
to Oct 2020a	−2.876	7.526	0.476	0.227	0.001*
to Nov 2020	−3.390	8.861	0.524	0.275	0.002*
Total adjusted maximum weight gained in kgb
Sept 2019 to Oct 2020c	−3.126	7.849	0.483	0.233	0.000*
Total change in body mass indexd in kg/m2
Sept 2019 to Oct 2020	−2.864	7.351	0.441	0.190	0.005*

aExcludes one extreme outlier.

bLast minus first measurement/time interval.

cExcludes one extreme outlier.

dExcludes one outlier. Variable reflected and then log-transformed. Excludes RP of 10 (n = 2).

*Statistically significant with p <  0.05.

Table 3

Regression analyses of log-transformed average total weight and associations with log-transformed Retzius periodicity.

	Quadratic curve
RP in days vs:	Intercept	Slope	r	r2	p
Average weight over 14 months in kg
	−0.375	4.431	0.333	0.111	0.035*
Average monthly weight in kga
Aug 2020	−2.132	8.829	0.401	0.161	0.026*
Sep 2020	−1.205	6.507	0.404	0.163	0.022*
Oct 2020	−1.379	6.908	0.397	0.157	0.012*

aRetzius periodicities of 5 to 9.

*Statistically significant with p  < 0.05.