Key findings
- 1.76 ± 1.71 cmmean height gain over three months in 79 Thai boys aged 12–17 years; 57 of them (72.2%) grew
- 2.08 vs 0.90 cmthree-month height gain with exercise on at least four days a week versus three days or less (p = 0.0018)
- 0.84 cmgreater gain still associated with frequent exercise after adjustment for age, milk intake and baseline target gap (95% CI 0.11–1.57)
- 36 of 79boys (45.6%) were already taller than their mid-parental predicted midpoint; among 17-year-olds, 20 of 37 (54.1%)
Abstract
Mid-parental height is widely used to estimate expected adult stature, although the predicted value represents a central estimate rather than a fixed biological limit. This three-month observational study examined short-term height gain and its associations with parental-height prediction, exercise frequency, and milk intake among Thai male adolescents. Seventy-nine boys aged 12–17 years were included. Baseline and follow-up height, parental height, weekly exercise frequency, and daily milk intake were recorded. Mid-parental predicted height was calculated as (father’s height + mother’s height + 13 cm)/2. The primary outcome was three-month height gain, while a positive height gap was defined as current height exceeding the predicted midpoint.
Mean height gain was 1.76 ± 1.71 cm, and 57 participants (72.2%) showed an increase in height. Overall, 36 of 79 boys (45.6%) had a current height above the mid-parental predicted midpoint. Participants reporting exercise on at least four days per week showed greater three-month height gain than those exercising three days per week or less (2.08 ± 1.75 vs. 0.90 ± 1.26 cm; mean difference 1.18 cm, 95% CI 0.46–1.90; p=0.0018). After adjustment for age, milk intake, and baseline target-height gap, higher exercise frequency remained associated with greater height gain (adjusted β=0.84 cm, 95% CI 0.11–1.57; p=0.024). Higher milk intake was associated with greater height gain in unadjusted analysis, but the association was attenuated after adjustment (β=0.68 cm, 95% CI −0.22 to 1.58; p=0.141). A larger baseline gap below the mid-parental predicted height was also associated with subsequent height gain (β=0.098 cm per 1-cm gap, 95% CI 0.047–0.148; p<0.001).
These findings should be interpreted as observational associations rather than causal effects. Mid-parental height appeared more appropriately interpreted as an expected midpoint than as a strict biological ceiling. Longer longitudinal follow-up with standardized height measurement, assessment of pubertal maturation, and more detailed behavioral and nutritional data is needed to determine whether these associations persist toward final adult height.
Introduction
Human height reflects the combined effects of genetic potential and the environment in which growth occurs. Parental stature is an important predictor of adult height, but nutrition, health, pubertal timing, socioeconomic conditions, and other environmental factors also influence how growth potential is expressed. Adolescence is especially important because substantial linear growth can still occur during this period.
Mid-parental height is widely used to estimate expected adult stature. For boys, a commonly used equation is:
Mid-parental height = (Father’s height + Mother’s height + 13) / 2
This value is useful for clinical growth assessment, but it should not be interpreted as a fixed genetic ceiling. Prediction accuracy may vary across populations and generations. In Japan, for example, target-height equations have been revised in response to secular changes in population height, and contemporary estimates are interpreted within a range rather than as a single absolute value.
Secular trends provide further evidence that attained height can change substantially across generations. Mean height among Japanese late-adolescent males increased by more than 7 cm between 1955 and 1995 before reaching a plateau. Thailand has shown a similar pattern. Data from the Thai Cohort Study indicated an increase in adult height of approximately 1 cm per decade among people born between 1940 and 1990, while another study in southern Thailand reported an increase in final male height of about 1.32 cm per decade. These changes are unlikely to result from genetic change alone and highlight the importance of nutrition, health, living conditions, and development.
Recent studies have also questioned the accuracy of traditional parental-height equations. A 2024 Korean study found that Tanner-based target height underestimated final adult height in many participants, with observed height exceeding the predicted value by several centimeters on average. Such findings suggest that mid-parental height is best viewed as a population-based prediction tool rather than a biological limit.
Modifiable behaviors may also contribute to the growth environment during adolescence. Milk provides protein, calcium, and other nutrients relevant to skeletal development, while regular physical activity supports bone and musculoskeletal health. However, evidence that either milk intake or exercise directly increases final adult height remains inconsistent. Their relationship with short-term adolescent growth therefore requires empirical evaluation.
The present study was designed as an observational analysis rather than an intervention study. Its purpose was not to establish whether exercise or milk intake causes an increase in height, but to examine whether these naturally occurring behavioral characteristics were associated with short-term height change after accounting for age and parental-height prediction. Specifically, this study examined three-month height change among Thai male adolescents aged 12–17 years, evaluated current height relative to the mid-parental predicted midpoint, and explored associations between short-term height change, exercise frequency, milk intake, age, and baseline target-height gap. These variables were evaluated within a multivariable framework to reduce, although not eliminate, the influence of measured confounding.
Materials and Methods
Study design and participants
This was a three-month observational study examining short-term height change among Thai male adolescents aged 12–17 years. Exercise frequency and milk intake were treated as naturally occurring behavioral exposures rather than assigned interventions. The study was designed to evaluate statistical associations between these exposures, parental-height prediction, age, and observed height change; it was not designed to establish causal effects of exercise or milk consumption on linear growth. The original dataset contained 83 responses. Because the present study focused on male adolescent growth, two female participants were excluded. Seven male students whose age had initially been recorded as 18 years were corrected to 17 years after verification of the source data, as the discrepancy resulted from the way age had been counted at the time of data entry. One record contained an invalid age value of 171.7 and was excluded from age-dependent analyses. One probable duplicate response was also removed to avoid counting the same participant twice. The final primary analytic sample therefore included 79 male adolescents aged 12–17 years.
The study focused on boys because sex-specific differences in pubertal timing and growth velocity may affect the interpretation of short-term changes in height. Restricting the analysis to one sex also allowed the same mid-parental height equation to be applied consistently across the study population.
Height assessment and three-month height gain
Height was recorded in centimeters. Two values were used for each participant: height at the beginning of the three-month observation period and current height at the end of follow-up.
Three-month height gain was calculated as:
Height gain (cm) = Current height − Baseline height
A positive value indicated an increase in height during the observation period. The main continuous outcome of the study was height gain in centimeters.
Because short-term changes of only a few centimeters may be sensitive to measurement technique, height change was interpreted at the group level rather than as evidence of a precise biological growth rate for each individual.
Parental height and mid-parental height prediction
Participants reported the heights of their biological father and mother in centimeters. Expected height based on parental stature was calculated using the conventional mid-parental height equation for boys:
Mid-parental predicted height = (Father’s height + Mother’s height + 13) / 2
The calculated value was treated as a predicted midpoint, rather than as an absolute genetic maximum.
To examine whether participants had reached or exceeded this prediction, a height gap was calculated as:
Height gap = Current height − Mid-parental predicted height
A value above zero was defined as a positive height gap, indicating that the participant's current height exceeded the predicted midpoint. A negative value indicated that current height remained below the prediction.
A second variable, the remaining target-height gap at baseline, was calculated as:
Target gap = Mid-parental predicted height − Baseline height
This variable was used to account for the amount of predicted growth potential that had not yet been reached at the beginning of the observation period.
Exercise exposure
Participants reported the number of days per week in which they performed regular physical activity. The main exercise program consisted of approximately two hours of physical activity per day, with activities centered on running, jumping-jack exercises, and stretching.
Exercise frequency was analyzed both as a continuous measure, expressed as days per week, and as a categorical variable. For the primary comparison, participants were divided into:
- Lower-frequency exercise: 0–3 days/week
- Higher-frequency exercise: 4–7 days/week
The cut-off was selected to distinguish adolescents with more consistent weekly participation from those with less frequent exercise.
The study did not assume that exercise directly lengthened bone. Instead, exercise frequency was evaluated as a potentially modifiable lifestyle factor associated with short-term growth.
Milk intake
Daily milk consumption was reported as the approximate number of bottles consumed per day. Milk intake was initially examined as an ordinal exposure and was also categorized for the main analysis as:
- <2 bottles/day
- ≥2 bottles/day
This grouping was selected because some individual intake categories contained only small numbers of participants. Combining them provided more stable estimates for statistical comparison.
Milk intake was treated as a proxy for one aspect of nutritional behavior. The study did not assess total energy intake, dietary protein, calcium intake from other foods, vitamin D status, or overall dietary quality.
Study outcomes
Two related outcomes were examined. The primary outcome was three-month height gain in centimeters. The secondary outcome was whether the participant's current height was above the mid-parental predicted midpoint. Participants were therefore classified as having either a positive or non-positive height gap.
Additional exploratory analyses compared current height with paternal height and assessed the proportion of adolescents who crossed from below the predicted midpoint at baseline to above the predicted midpoint during the three-month period.
Table 1. Operational definitions of the main study variables
| Variable | Definition | Analysis |
|---|---|---|
| 3-month height gain | Current height − baseline height | Continuous, cm |
| Mid-parental predicted height | (Father + Mother + 13)/2 | Continuous, cm |
| Positive height gap | Current height − predicted height >0 | Yes/No |
| Target gap at baseline | Predicted height − baseline height | Continuous, cm |
| Exercise frequency | Days of exercise per week | ≤3 vs ≥4 days/week |
| Milk intake | Reported bottles per day | <2 vs ≥2 bottles/day |
| Age | Age at study period | 12–14, 15–16, 17 years |
Data quality and cleaning
The dataset was screened before statistical analysis. Records outside the intended sex criteria were removed. Age values were checked for implausible entries, and probable duplicate responses were reviewed.
One clearly invalid age entry was excluded from the primary analysis. One probable duplicate record was also removed. A sensitivity analysis retaining the probable duplicate produced materially similar findings, supporting the robustness of the main results.
All derived variables, including height gain, mid-parental predicted height, height gap, and target gap, were recalculated from the source variables before analysis.
Statistical analysis
Continuous variables were summarized using means and standard deviations when appropriate, together with medians and interquartile ranges. Categorical variables were presented as frequencies and percentages.
The relationship between age and height gain was examined using correlation analysis. Because the distribution of height gain differed across some age categories, the Kruskal–Wallis test was used to compare growth among age groups.
Exercise groups and milk-intake groups were compared using Welch's independent-samples t-test, which does not require equal variances. Mean differences were reported with 95% confidence intervals. Effect sizes were estimated using Hedges' g.
Associations between weekly exercise frequency and height gain, and between milk-intake level and height gain, were also examined using Spearman's rank correlation coefficient. Pearson correlation was used for continuous variables for which a linear relationship was appropriate, including age and the target-height gap.
A multivariable linear regression model was then constructed with three-month height gain as the dependent variable. The principal model included:
Height gain = β0 + β1(Age) + β2(Exercise) + β3(Milk intake) + β4(Target gap) + ε
Exercise was entered as ≥4 versus ≤3 days/week, while milk intake was entered as ≥2 versus <2 bottles/day. The target-height gap represented the difference between mid-parental predicted height and baseline height.
Heteroskedasticity-consistent HC3 robust standard errors were used for the regression model. Adjusted regression coefficients were reported with 95% confidence intervals and p-values. A two-sided p-value <0.05 was considered statistically significant.
Interpretation of mid-parental height
An important methodological principle of this study was that mid-parental height was not treated as a fixed genetic ceiling. A participant who exceeded the predicted midpoint was therefore not described as having “overcome genetics.” Rather, the difference between attained height and the parental prediction was used to assess how closely a conventional prediction equation reflected observed stature in contemporary Thai adolescents.
This distinction was especially important because parental-height equations estimate a central expected value and are subject to population differences, secular changes in height, pubertal timing, environmental influences, and prediction error.
Ethical considerations
This study was conducted in accordance with the ethical principles of the Declaration of Helsinki and was approved by the Institutional Review Board (IRB) of CADET-TUTOR SCHOOL (Approval No. 06004/2026).
Results
Participant characteristics
The primary analysis included 79 male adolescents aged 12–17 years. Mean age was 15.97 ± 1.24 years. Mean height increased from 170.77 ± 8.05 cm at baseline to 172.53 ± 7.53 cm after three months. The mean mid-parental predicted height was 173.14 ± 4.88 cm.
Participants exercised an average of 4.61 ± 1.99 days per week. Fifty-eight boys (73.4%) exercised at least four days per week, while 21 (26.6%) exercised three days per week or less. Twenty-one participants (26.6%) reported drinking at least two bottles of milk per day.
Table 2. Characteristics of the Study Participants
| Characteristic | Value |
|---|---|
| Participants, n | 79 |
| Age, years | 15.97 ± 1.24 |
| Age 12–14 years, n (%) | 9 (11.4) |
| Age 15–16 years, n (%) | 33 (41.8) |
| Age 17 years, n (%) | 37 (46.8) |
| Baseline height, cm | 170.77 ± 8.05 |
| Current height, cm | 172.53 ± 7.53 |
| 3-month height gain, cm | 1.76 ± 1.71 |
| Mid-parental predicted height, cm | 173.14 ± 4.88 |
| Target gap at baseline, cm | 2.37 ± 8.38 |
| Exercise, days/week | 4.61 ± 1.99 |
| Exercise ≥4 days/week, n (%) | 58 (73.4) |
| Exercise ≤3 days/week, n (%) | 21 (26.6) |
| Milk ≥2 bottles/day, n (%) | 21 (26.6) |
| Milk <2 bottles/day, n (%) | 58 (73.4) |
Three-Month Height Gain
During the three-month observation period, 57 of 79 participants (72.2%) increased in height. Mean height gain was 1.76 ± 1.71 cm (95% CI 1.38–2.15 cm), with a median gain of 1.0 cm. Individual changes ranged from 0 to 8 cm.
Height gain differed substantially by age. Boys aged 12–14 years gained an average of 3.62 ± 1.68 cm, compared with 1.70 ± 1.38 cm among those aged 15–16 years and 1.37 ± 1.72 cm among 17-year-olds. The difference across age groups was statistically significant (Kruskal–Wallis p=0.0018).
Age was inversely correlated with three-month height gain (r=−0.332, p=0.0028), indicating that younger participants generally showed greater short-term linear growth.
Exercise Frequency and Height Gain
A clear difference was observed according to exercise frequency. Participants who exercised four or more days per week gained an average of 2.08 ± 1.75 cm, whereas those exercising three days per week or less gained 0.90 ± 1.26 cm.
The unadjusted mean difference was 1.18 cm (95% CI 0.46–1.90 cm; Welch's p=0.0018). The standardized effect was moderately large (Hedges' g=0.72).
Weekly exercise frequency also showed a positive correlation with height gain (Spearman ρ=0.297, p=0.008).
Table 3. Three-Month Height Gain According to Age, Exercise Frequency, and Milk Intake
| Factor | n | Height gain, cm | Difference | 95% CI | p-value |
|---|---|---|---|---|---|
| Age | 0.0018† | ||||
| 12–14 years | 9 | 3.62 ± 1.68 | — | — | |
| 15–16 years | 33 | 1.70 ± 1.38 | — | — | |
| 17 years | 37 | 1.37 ± 1.72 | — | — | |
| Exercise frequency | |||||
| ≤3 days/week | 21 | 0.90 ± 1.26 | Reference | — | |
| ≥4 days/week | 58 | 2.08 ± 1.75 | +1.18 | 0.46 to 1.90 | 0.0018 |
| Milk intake | |||||
| <2 bottles/day | 58 | 1.47 ± 1.50 | Reference | — | |
| ≥2 bottles/day | 21 | 2.57 ± 2.01 | +1.10 | 0.11 to 2.08 | 0.030 |
Milk intake and height gain
Participants who reported consuming at least two bottles of milk per day gained 2.57 ± 2.01 cm, compared with 1.47 ± 1.50 cm among those consuming less than two bottles per day.
The unadjusted difference was 1.10 cm (95% CI 0.11–2.08 cm; p=0.030), with a Hedges' g of 0.66. Milk intake also showed a weak positive rank correlation with height gain (Spearman ρ=0.224, p=0.048). The association, however, became weaker after other growth-related factors were considered in the multivariable model.
Independent factors associated with height gain
A multivariable linear regression model was constructed to distinguish the effects of exercise and milk intake from age and the participant's remaining distance from the mid-parental predicted height. The overall model was statistically significant (p<0.001) and explained 36.3% of the variance in three-month height gain (R2=0.363; adjusted R2=0.329).
After adjustment, exercising at least four days per week remained independently associated with greater height gain. These participants gained an estimated 0.84 cm more over three months than those exercising three days or fewer (adjusted β=0.84 cm, 95% CI 0.11–1.57; p=0.024).
The association with milk consumption was attenuated after adjustment. Drinking at least two bottles per day was associated with an estimated 0.68 cm greater gain, but the confidence interval included zero (β=0.68 cm, 95% CI −0.22 to 1.58; p=0.141).
The strongest independent association was observed for the target gap at baseline. Each additional centimeter between the mid-parental predicted height and baseline height was associated with approximately 0.10 cm greater height gain during the following three months (β=0.098, 95% CI 0.047–0.148; p<0.001).
Table 4. Multivariable Linear Regression for Three-Month Height Gain
| Predictor | Adjusted β (cm) | Robust SE | 95% CI | p-value |
|---|---|---|---|---|
| Age, per year | 0.02 | 0.21 | −0.40 to 0.43 | 0.936 |
| Exercise ≥4 days/week | 0.84 | 0.37 | 0.11 to 1.57 | 0.024 |
| Milk ≥2 bottles/day | 0.68 | 0.46 | −0.22 to 1.58 | 0.141 |
| Target gap, per 1 cm | 0.098 | 0.026 | 0.047 to 0.148 | <0.001 |
HC3 heteroskedasticity-robust standard errors. Model R2=0.363; adjusted R2=0.329; n=79.
Observed height relative to mid-parental prediction
The comparison with parental height produced one of the most relevant findings for the central question of this study. Of the 79 participants, 36 (45.6%) had a current height above their mid-parental predicted midpoint, four (5.1%) were exactly at the predicted value, and 39 (49.4%) remained below it.
The cohort as a whole was, on average, only 0.61 cm below the predicted midpoint. However, this overall value concealed a strong age-dependent pattern.
None of the nine boys aged 12–14 years had yet exceeded their predicted midpoint. Their current height remained an average of 12.21 cm below prediction, which is consistent with substantial remaining adolescent growth.
Among boys aged 15–16 years, 16 of 33 (48.5%) were already above the predicted midpoint. Their mean height gap was +0.41 cm.
Among 17-year-olds, 20 of 37 (54.1%) had exceeded the mid-parental predicted midpoint, and their average current height was 1.31 cm above prediction.
Table 5. Current Height Relative to Mid-Parental Predicted Height
| Age group | n | Above prediction, n (%) | Mean current − predicted height |
|---|---|---|---|
| 12–14 years | 9 | 0 (0.0%) | −12.21 cm |
| 15–16 years | 33 | 16 (48.5%) | +0.41 cm |
| 17 years | 37 | 20 (54.1%) | +1.31 cm |
| Overall | 79 | 36 (45.6%) | −0.61 cm |
The proportion above the predicted midpoint differed across age groups (p=0.013). Four participants who were at or below their predicted midpoint at baseline moved above it during the three-month observation period. Among the 47 boys who had not exceeded the midpoint at baseline, this represented 8.5% crossing the predicted midpoint during follow-up.
In addition, 44 of 79 participants (55.7%) were already taller than their fathers, further illustrating that parental stature should not be interpreted as an individual upper boundary for attained height.
Modifiable factors and positive height gap
Exercise and milk intake were clearly associated with recent height gain, but they were not significantly associated with the cross-sectional probability of already being above the mid-parental predicted midpoint.
A positive height gap was observed in 26 of 58 boys (44.8%) who exercised at least four days per week and 10 of 21 boys (47.6%) who exercised less frequently. Similarly, 9 of 21 boys (42.9%) consuming at least two bottles of milk per day had a positive height gap compared with 27 of 58 (46.6%) consuming less milk.
After adjustment for age, neither exercise nor milk intake independently predicted a positive height gap. Age remained the only significant variable in this exploratory model.
Table 6. Logistic Regression for a Positive Height Gap
| Predictor | Adjusted OR | 95% CI | p-value |
|---|---|---|---|
| Age, per additional year | 1.75 | 1.10–2.79 | 0.019 |
| Exercise ≥4 days/week | 1.47 | 0.48–4.46 | 0.496 |
| Milk ≥2 bottles/day | 1.12 | 0.37–3.37 | 0.840 |
This distinction is important. Exercise was associated with the amount of growth observed during the study, but not with whether a participant had already exceeded his parental-height midpoint. The latter reflects growth accumulated over many years and is strongly influenced by developmental stage.
Actual height versus mid-parental predicted height
Mid-parental height remains useful as a reference for growth potential, but it does not operate as a fixed ceiling. Nearly half of the cohort—and more than half of the 17-year-olds—had already exceeded its predicted midpoint. At the same time, the modifiable behaviors measured in this study were more closely related to recent growth than to the accumulated positive height gap.
Discussion
This study examined short-term height gain in Thai male adolescents and asked a second, broader question: how closely does a conventional mid-parental height estimate reflect the height actually attained by contemporary adolescents? Two findings stood out. First, boys who exercised more frequently showed greater height gain over the three-month observation period, even after adjustment for age, milk intake, and the remaining gap from predicted height. Second, almost half of the participants had already reached a height above the midpoint predicted from parental stature. Among 17-year-olds, more than half were above that predicted midpoint. These findings do not imply that lifestyle can replace or overcome genetic influences. Rather, they support a more cautious interpretation of parental-height equations as population-based estimates that operate within a wider biological and environmental context.
Mid-parental height is a prediction, not a fixed ceiling
Mid-parental height remains useful in pediatric growth assessment because parental stature carries important genetic information. However, the value produced by the equation is a central estimate, not a maximum possible height. This distinction is important because the formula is sometimes interpreted too rigidly in non-clinical settings.
In the present cohort, 36 of 79 boys (45.6%) were already taller than the mid-parental predicted midpoint. The pattern became more evident with age. None of the boys aged 12–14 years had exceeded the midpoint, which is expected because many were still in active growth. In contrast, 48.5% of those aged 15–16 years and 54.1% of 17-year-olds were already above prediction. The mean height gap among 17-year-olds was approximately +1.31 cm.
This finding should not be interpreted as evidence that the conventional formula is invalid. A target-height equation has an expected range around its midpoint, and individuals can fall above or below that midpoint while still remaining fully consistent with familial growth potential. The present data instead show why the predicted midpoint should not be used as a rigid biological boundary.
The concept is consistent with recent work from Korea. In a contemporary Korean sample, Tanner-based target-height equations underestimated final adult stature, with observed height exceeding the conventional estimate by several centimeters on average. That study emphasized the need for population-specific reassessment of prediction equations when secular changes in height occur. The Thai data presented here point in the same direction, although the present sample is smaller and includes adolescents who have not all reached final adult height.
Secular growth trends support reassessment of parental-height formulas
The relevance of these findings becomes clearer when viewed against long-term changes in population height. Japan provides a useful example. Mean height among late-adolescent Japanese males increased by more than 7 cm over several decades before reaching a plateau. Such a change occurred far too rapidly to be explained by genetic evolution alone. Improvements in nutrition, healthcare, living conditions, infection control, and childhood development are more plausible explanations for much of the secular increase.
Thailand has also shown a positive secular trend in adult height. Large cohort data have demonstrated increases of approximately 1 cm per decade across several birth generations, while studies in southern Thai boys have shown increases in final height alongside changes in health, nutrition, and pubertal timing.
These population trends have an important implication for parental-height prediction. A formula based on the heights of one generation may systematically underestimate the next generation if environmental conditions improve and average attained height rises. This does not weaken the role of genetics. Instead, it highlights the difference between inherited potential and realized stature.
The present study extends this concept to individual adolescents. The parental midpoint remained relevant, but it did not act as a strict boundary. A substantial proportion of older adolescents had already exceeded it.
Exercise frequency and short-term height gain
Among the behavioral variables examined, exercise frequency showed the clearest statistical association with short-term height gain. Boys who exercised at least four days per week gained an average of 2.08 cm over three months, compared with 0.90 cm among those exercising three days per week or less. The crude difference was 1.18 cm, and the association remained significant after adjustment. In the multivariable model, frequent exercise was associated with approximately 0.84 cm greater three-month height gain.
This result should be interpreted carefully. Exercise does not directly elongate long bones in the way a pharmacological growth treatment might. Linear growth depends on activity at the epiphyseal growth plates, endocrine regulation, nutritional sufficiency, maturation, and genetic background. Exercise may instead contribute indirectly by supporting musculoskeletal health, appetite, body composition, sleep quality, and overall metabolic health.
The exercise program reported by participants included running, jumping jacks, and stretching for approximately two hours per day. These activities involve repeated weight-bearing and mechanical loading. Such loading is well established as beneficial for bone health, although evidence that it increases final adult height remains limited.
The present findings therefore support an association between frequent exercise and short-term growth, but they do not establish causality. It is also possible that adolescents in a more active stage of pubertal growth were more likely to exercise frequently, or that active adolescents differed in nutrition, sleep, or general health in ways that were not fully measured.
Why the age effect became weaker after adjustment
Age showed a clear inverse relationship with height gain in the unadjusted analysis. Boys aged 12–14 years gained substantially more height than those aged 17 years. This is biologically expected because younger adolescents are more likely to be closer to the pubertal growth spurt, whereas many older adolescents are approaching the end of linear growth.
An interesting finding was that chronological age was no longer significant in the multivariable model after the target-height gap was included.
This may reflect overlap between age and remaining growth potential. A younger boy is often further below his eventual height than an older boy. In this dataset, the difference between baseline height and mid-parental predicted height may therefore have captured part of the biological information that chronological age otherwise represented.
The finding should not be interpreted to mean that age is unimportant. Rather, it suggests that age alone is an incomplete marker of maturation. Future work should include Tanner stage or bone age, both of which would provide a more direct estimate of remaining growth potential.
Milk intake: positive signal, but weaker after adjustment
Milk intake showed a positive association with height gain in the unadjusted analysis. Boys consuming at least two bottles per day gained approximately 1.10 cm more than those consuming less than two bottles per day.
However, this association was no longer statistically significant after adjustment.
This attenuation is important because it suggests that the crude milk-height relationship may partly reflect other factors. Boys who drink more milk may differ in age, total energy intake, protein intake, family habits, socioeconomic status, or general nutritional quality. Milk itself provides protein, calcium, and other nutrients relevant to skeletal development, but the present study did not assess the total diet.
Therefore, the data support milk intake as a possible marker of a growth-supportive nutritional pattern rather than as an independent determinant of height.
Future studies should measure total protein, calcium, vitamin D, energy intake, and overall dietary quality instead of relying on milk intake alone.
The target gap was the strongest predictor of subsequent growth
The strongest independent predictor in the regression model was the difference between predicted height and baseline height.
For every additional centimeter between baseline height and the mid-parental predicted value, subsequent three-month height gain increased by approximately 0.10 cm.
This finding is biologically plausible. Adolescents who remain further below their expected familial stature may simply have more growth remaining. The target gap may therefore function as a rough indicator of residual growth potential.
At the same time, this result also highlights a limitation of the parental-height approach. The target gap is partly constructed from the same prediction formula being evaluated. It should therefore be interpreted as a descriptive growth marker rather than as proof that the formula accurately defines each participant's final potential.
Positive height gap and modifiable behaviors are not the same outcome
A useful distinction emerged between recent growth and having already exceeded the parental midpoint. Exercise was associated with three-month height gain, but it was not significantly associated with the probability of already having a positive height gap. Milk showed the same pattern.
This makes biological sense. A positive height gap reflects growth accumulated over many years, while the present exercise and milk variables describe current or recent behavior. A participant may have exceeded the parental midpoint because of many earlier influences that were not measured in the study.
The lack of association with positive height gap should therefore not be viewed as a contradiction. Instead, it suggests that the two outcomes capture different aspects of growth.
Implications for the interpretation of genetic height potential
The central message of this study is not that genetics can be defeated.
A more accurate interpretation is that genetic potential is expressed through a changing biological and environmental context.
Parental height remains informative because stature is strongly heritable. However, inheritance does not operate in isolation. Growth is influenced by nutrition, disease burden, endocrine status, physical activity, sleep, socioeconomic conditions, and pubertal timing.
For this reason, an adolescent who grows above a parental-height midpoint has not necessarily exceeded his genetic potential. He may simply be expressing a phenotype that the simplified parental formula did not predict precisely.
This distinction is especially important when communicating with families. The mid-parental equation should be presented as an estimate with uncertainty, not as a guaranteed destination or limit.
Strengths of the study
This study has several strengths. It combined short-term longitudinal height change with parental-height prediction rather than examining only one of these outcomes. It also evaluated modifiable behaviors alongside familial height information.
Restricting the study to males reduced variation related to sex-specific pubertal timing and allowed a consistent target-height equation to be applied. The use of multivariable regression also helped separate the crude effects of age, exercise, milk intake, and predicted growth potential.
The concept of a positive height gap provides an intuitive way to examine how observed adolescent stature relates to conventional parental expectations.
Limitations
Several limitations require careful consideration. First, the sample size was modest and was drawn from a specific group of Thai male adolescents. The results should not be generalized to all Thai adolescents.
Second, not all participants had reached final adult height. This is particularly important for boys aged 12–16 years. The present study therefore evaluates attained adolescent height, not definitive adult stature.
Third, pubertal stage and bone age were not available. These variables are central determinants of growth velocity and would substantially improve interpretation of short-term height gain.
Fourth, height values and behavioral variables may be affected by measurement or recall error. A 1–2 cm difference can arise from posture, time of day, measurement equipment, or technique. Future studies should use a calibrated stadiometer and standardized repeated measurements.
Fifth, milk intake was used as a simple nutritional indicator. Total dietary intake was not measured. Sixth, sleep duration could not be analyzed because individual bedtime or total sleep-hours data were not available in the current dataset.
Finally, the observational design does not establish causality. Frequent exercise was associated with greater height gain, but the study cannot determine whether exercise itself caused the difference.
Future research
The most important next step is to follow these participants until near-final or final adult height. Serial measurements every three to six months would allow calculation of annualized growth velocity. Adding Tanner stage, bone age, BMI, dietary protein, calcium intake, vitamin D status, sleep duration, and socioeconomic variables would strengthen causal interpretation.
A larger study could also directly compare conventional mid-parental prediction with observed final adult height using prediction-error statistics such as mean absolute error, root mean square error, and Bland–Altman analysis. Such work could determine whether the conventional equation requires recalibration for contemporary Thai adolescents.
Main interpretation
Taken together, the findings support a more flexible view of adolescent height prediction. The mid-parental equation remains useful as a clinical reference, but it should not be interpreted as a fixed genetic ceiling. Almost half of the boys in this cohort had already exceeded the predicted midpoint, and this proportion increased with age. Frequent exercise was independently associated with greater short-term height gain, while milk intake showed a positive but weaker association after adjustment.
The data therefore support a model in which parental stature defines part of the expected growth framework, while developmental stage and modifiable environmental factors contribute to how that potential is expressed.
Conclusion
This three-month observational study identified associations between short-term height change, exercise frequency, age-related growth characteristics, and parental-height prediction among Thai male adolescents. Participants reporting exercise on at least four days per week showed greater short-term height gain than those exercising less frequently, and the association remained statistically significant after adjustment for age, milk intake, and baseline target-height gap. Milk intake showed a positive association in unadjusted analysis, but this association was attenuated after adjustment. These findings represent observational associations and should not be interpreted as evidence that exercise or milk consumption directly causes increased linear growth.
Nearly half of the participants had a current height above their mid-parental predicted midpoint, including more than half of the 17-year-old participants. This finding supports interpreting mid-parental height as an expected reference value rather than a fixed biological ceiling. However, because many participants had not reached final adult height and pubertal stage and bone age were unavailable, the present findings cannot determine the accuracy of mid-parental height for predicting final stature.
Longer longitudinal studies incorporating standardized repeated height measurements, pubertal stage, bone age, detailed nutritional assessment, and more precisely characterized physical activity are needed to determine whether the observed associations persist and whether conventional mid-parental height equations require recalibration for contemporary Thai adolescents.
Declarations
- Ethics Approval and Consent to Participate
- This study was conducted in accordance with the ethical principles of the Declaration of Helsinki and was approved by the Institutional Review Board (IRB) of CADET-TUTOR SCHOOL (Approval No. 06004/2026).
- Written informed consent was obtained from the participants’ parents or legal guardians, and assent was obtained from the adolescent participants before data collection.
- Consent for publication
- No individually identifiable information, photographs, or personal clinical data are reported in this manuscript. Consent for publication was therefore not applicable beyond the consent obtained for participation and use of anonymized research data.
- Data availability
- The dataset generated and analyzed during the present study contains information obtained from adolescent participants and is not publicly available because of privacy considerations. De-identified data may be made available by the corresponding author upon reasonable request and subject to applicable ethical and institutional requirements.
- Conflict of interest
- The authors declare that they have no financial or non-financial conflicts of interest related to this study.
- Funding
- This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
- Author contributions
- Kritsanat Klakhaeng1 contributed to participant coordination, data collection, data verification, and preparation of the study dataset.
- Teeravee Akrarath-amorn2 contributed to participant coordination, data collection, data checking, and review of the manuscript.
- Pongkit Ekvitayavetchanukul3 conceived and designed the study, developed the research methodology, supervised the project, performed the statistical analysis, interpreted the findings, prepared the tables and figures, drafted the manuscript, critically revised the intellectual content, and approved the final version of the manuscript.
- Use of artificial intelligence
- Artificial intelligence–assisted language tools were used only to support language editing, organization, and improvement of readability. The authors independently verified all scientific content, statistical results, interpretations, citations, and conclusions and take full responsibility for the final manuscript. AI tools were not used to generate, alter, or fabricate research data.
- Acknowledgments
- The authors thank the participating students and their families for providing the information required for this study. The authors also acknowledge the staff involved in coordinating data collection and supporting the three-month follow-up.
References
- Alves, J. G. B., & Alves, G. V. (2019). Effects of physical activity on children's growth. Jornal de Pediatria, 95(Suppl. 1), 72–78. https://doi.org/10.1016/j.jped.2018.11.003
- Bents, D., Groth, D., & Satake, T. (2018). The secular trend and network effects on height of male Japanese students from 1955 to 2015. Anthropologischer Anzeiger, 74(5), 423–429. https://doi.org/10.1127/anthranz/2018/0838
- Bland, V. L., Heatherington-Rauth, M., Howe, C., Going, S. B., & Bea, J. W. (2020). Association of objectively measured physical activity and bone health in children and adolescents: A systematic review and narrative synthesis. Osteoporosis International, 31(10), 1865–1894. https://doi.org/10.1007/s00198-020-05485-y
- Bogin, B. (2013). Secular changes in childhood, adolescent and adult stature. Nestlé Nutrition Institute Workshop Series, 71, 115–126. https://doi.org/10.1159/000342581
- Brailey, G., Metcalf, B., Lear, R., Price, L., Cumming, S., & Stiles, V. H. (2022). A comparison of the associations between bone health and three different intensities of accelerometer-derived habitual physical activity in children and adolescents: A systematic review. Osteoporosis International, 33, 1191–1222. https://doi.org/10.1007/s00198-021-06218-5
- Ciancia, S., Ribes Cajas, P., & Cools, M. (2026). How accurate is Tanner’s formula in estimating target height? BMC Pediatrics, 26, 30. https://doi.org/10.1186/s12887-025-05551-6
- Cole, T. J. (2003). The secular trend in human physical growth: A biological view. Economics & Human Biology, 1(2), 161–168. https://doi.org/10.1016/S1570-677X(02)00033-3
- Ekvitayavetchanukul, P., Bhavani, C., Nath, N., Sharma, L., Aggarwal, G., Singh, R. (2024). Revolutionizing Healthcare: Telemedicine and Remote Diagnostics in the Era of Digital Health. In: Kumar, P., Singh, P., Diwakar, M., Garg, D. (eds) Healthcare Industry Assessment: Analyzing Risks, Security, and Reliability. Engineering Cyber-Physical Systems and Critical Infrastructures, vol 11. Springer, Cham. https://doi.org/10.1007/978-3-031-65434-3_11
- Ekvitayavetchanukul, P., & Ekvitayavetchanukul, P. (2025). AI-Driven Design Thinking: Transforming Learning Efficiency in Pre-Medical Education. Medical Research Archives, 13(4). https://doi.org/10.18103/mra.v13i4.6410
- Ekvitayavetchanukul, P., & Ekvitayavetchanukul, P. (2023). Comparing the effectiveness of distance learning and onsite learning in pre-medical courses. Recent Educational Research, 1(2), 141-147. https://doi.org/10.59762/rer904105361220231220143511
- Fudvoye, J., & Parent, A.-S. (2017). Secular trends in growth. Annales d'Endocrinologie, 78(2), 88–91. https://doi.org/10.1016/j.ando.2017.04.003
- Jaruratanasirikul, S., & Sriplung, H. (2015). Secular trends of growth and pubertal maturation of school children in Southern Thailand. Annals of Human Biology, 42(5), 447–454. https://doi.org/10.3109/03014460.2014.955057
- Jordan, S., Lim, L., Seubsman, S.-A., Bain, C., Sleigh, A., & Thai Cohort Study Team. (2012). Secular changes and predictors of adult height for 86,105 male and female members of the Thai Cohort Study born between 1940 and 1990. Journal of Epidemiology and Community Health, 66(1), 75–80. https://doi.org/10.1136/jech.2010.113043
- Kang, K., Sotunde, O. F., & Weiler, H. A. (2019). Effects of milk and milk-product consumption on growth among children and adolescents aged 6–18 years: A meta-analysis of randomized controlled trials. Advances in Nutrition, 10(2), 250–261. https://doi.org/10.1093/advances/nmy081
- Kim, S., Yoo, J. H., & Chueh, H. W. (2024). Tanner’s target height formula underestimates final adult height in Korean adolescents and young adults: Reassessment of target height based on the Korean National Health and Nutrition Examination Survey 2010–2019. BMJ Paediatrics Open, 8(1), e002653. https://doi.org/10.1136/bmjpo-2024-002653
- Luo, Z. C., Albertsson-Wikland, K., & Karlberg, J. (1998). Target height as predicted by parental heights in a population-based study. Pediatric Research, 44, 563–571. https://doi.org/10.1203/00006450-199810000-00016
- Marshall, T. A., Curtis, A. M., Cavanaugh, J. E., Warren, J. J., & Levy, S. M. (2018). Higher longitudinal milk intakes are associated with increased height in a birth cohort followed for 17 years. The Journal of Nutrition, 148(7), 1144–1149. https://doi.org/10.1093/jn/nxy071
- Mathasuriyapong, P., Korcharlermsonthi, N., Ekvitayavetchanukul, P., & Ekvitayavetchanukul, P. (2025). Modeling the health burden of PM2.5: Forecasting hospital admissions and medical demand in Bangkok and neighboring regions. Journal of Posthumanism, 5(6), 862–873. https://doi.org/10.63332/joph.v5i6.2155
- Matsumoto, K. (1982). Secular acceleration of growth in height in Japanese and its social background. Annals of Human Biology, 9(5), 399–410. https://doi.org/10.1080/03014468200005941
- Ogata, T. (2006). Genetics of human growth. Clinical Pediatric Endocrinology, 15(2), 45–53. https://doi.org/10.1297/cpe.15.45
- Ogata, T., Tanaka, T., & Kagami, M. (2007). Target height and target range for Japanese children: Revisited. Clinical Pediatric Endocrinology, 16(4), 85–87. https://doi.org/10.1297/cpe.16.85
- Silventoinen, K., Kaprio, J., Lahelma, E., & Koskenvuo, M. (2000). Relative effect of genetic and environmental factors on body height: Differences across birth cohorts among Finnish men and women. American Journal of Public Health, 90(4), 627–630. https://doi.org/10.2105/AJPH.90.4.627
- Sutabutra, T., Rujachan, P., Manasakorn, K., Sripetchnai, M., & Ekvitayavetchanukul, P. (2025). The impact of design thinking vs rote learning on secondary student achievement: An experimental study in Bangkok schools. Asian Journal of Education and Social Studies, 51(2), 411–422. https://doi.org/10.9734/ajess/2025/v51i21794
- Tan, V. P. S., Macdonald, H. M., Kim, S., Nettlefold, L., Gabel, L., Ashe, M. C., & McKay, H. A. (2014). Influence of physical activity on bone strength in children and adolescents: A systematic review and narrative synthesis. Journal of Bone and Mineral Research, 29(10), 2161–2181. https://doi.org/10.1002/jbmr.2254
- Zeevi, D., Ben Yehuda, A., Nathan, D., Zangen, D., & Kruglyak, L. (2024). Accurate prediction of children's target height from their mid-parental height. Children, 11(8), 916. https://doi.org/10.3390/children11080916
- Zhang, W., Wang, X., Liu, Y., He, Q., Ding, Q., Mei, J., & Li, X. (2024). Effects of exercise on bone mass and bone metabolism in adolescents: A systematic review and meta-analysis. Frontiers in Physiology, 15, 1512822. https://doi.org/10.3389/fphys.2024.1512822
Article information
Authors
- Kritsanat KlakhaengPholwitaya schoolCorresponding author · [email protected]
- Teeravee Akrarath-amornPholwitaya school
- Pongkit EkvitayavetchanukulThe Board of Khon Kaen University Affairs, Khon Kaen University
How to cite
Klakhaeng, K., Akrarath-amorn, T., et al., 2026. Reassessing Mid-Parental Height in Thai Male Adolescents: A Three-Month Observational Study of Short-Term Height Change and Associated Factors. Medical Research Archives, [online] 14(10). https://doi.org/10.18103/mra.2026.0671
Record
- Journal
- Medical Research Archives
- Issue
- Vol 14 No 10 (2026)
- Section
- Research Articles
- Published
- 31 October 2026
- ISSN
- 2375-1924
- Downloads
- 71
Copyright and licence
Published by the European Society of Medicine under a Creative Commons Attribution 4.0 International licence (CC BY 4.0), which permits unrestricted use, distribution and reproduction in any medium, provided the original authors and source are credited.