Association between anthropometric indices and electrocardiographic variables among healthy black adolescents

¹Federal Teaching Hospital, Ido-Ekiti, Ekiti State, Nigeria, ²Obafemi Awolowo University Teaching Hospitals Complex, Ile-Ife, Osun State, Nigeria, ³Child and Adolescent Mental Health Services, Lincolnshire Partnership NHS Foundation Trust, England, UK

^&Corresponding author
Olukemi Tolulope Bamigboye-Taiwo, Obafemi Awolowo University Teaching Hospitals Complex, Ile-Ife, Osun State, Nigeria

Introduction    

The electrocardiogram (ECG) is a non-invasive, safe, quick and relatively inexpensive test useful in obtaining information on the heart's electrical activity and evaluating cardiovascular abnormalities [1,2] Anthropometric indices can serve as important indicators of health and fitness in adolescents. ECG variables are influenced by anthropometric indices that change throughout adolescence [1,2]. In addition, the nutrition, dietary and weight regulation habits of black adolescents differ from those of Caucasian adolescents. Adolescence represents a dynamic period of growth and physiological transition during which substantial changes occur in body composition and cardiovascular function that may influence electrocardiographic (ECG) parameters. Anthropometric indices such as height, weight, and body mass index (BMI) are key indicators of growth and nutritional status, and they influence cardiovascular performance through their effects on cardiac structure and function [3,4]. Changes in ECG parameters can be associated with an increased risk of adverse cardiovascular events [5]. For instance, prolonged PR interval has been associated with a risk of heart failure and incident atrial fibrillation [6,7]; while prolonged QRS duration is a potential marker of cardiac structural and functional abnormality [8,9]. Prolonged QTc intervals predict the risk of sudden death in patients without evidence of cardiac dysfunction [10]; while electrocardiographic axes are known important markers in cardiovascular prognosis.

Height is a strong determinant of left atrial size which by itself, is a risk factor for incident atrial fibrillation. Left atrial size and left ventricular hypertrophy are correlated in adolescents, suggesting a symmetrical increase in cardiac chamber size and volumes in healthy adults [11]. Kofler et al. [12], in their study of young adults, found significant associations of measured and genetically determined height with PR interval and QRS duration. Tall individuals had significantly longer PR intervals compared with short individuals implying that adult height may be a marker of altered cardiac conduction. In a study by Santini et al. [13] of 24,062 adolescents, right bundle branch block (RBBB) was found to be associated with height irrespective of age, sex and BMI suggesting that tall stature is associated with anomalous conduction in the right bundle branch.

Body mass index (BMI) or Quetelet index which is calculated as weight (kilogrammes)/ height² (metres²) [14] is a surrogate marker for body fat, which does not directly measure adiposity but affects ECG variables and can be used as a measure in clinical practice to predict and assess cardiovascular morbidity. Swamy et al. [15] observed that PR interval and QRS duration were positively correlated with body mass index (BMI). Longer PR intervals and wider QRS durations were observed with higher BMI categories. A study conducted on 51 healthy male individuals aged 18 to 23 years concluded that body weight is positively correlated to QRS duration. It also showed that greater BMI shifted the electrical activity of repolarisation of the ventricles to the left [16]. Mirahmadizadeh et al. demonstrated that QRS, QT, and R in aVL intervals increased with elevated BMI [2]. Eugen et al. [17] also reported that increasing weight and BMI correlated with the widening of the QRS duration, increasing R voltage and leftward QRS axis. Suthar et al. [18] also observed that a significant BMI increase causes statistically significant changes in the mean QRS axis and P wave duration. Ardissino et al. [6], in their genome-wide association studies (GWAS) on BMI, waist-hip ratio adjusted for BMI, height, weight and fat-free mass, reported that higher genetically predicted BMI was associated p wave duration, as was higher fat mass, fat-free. Genetically predicted BMI, height and weight were associated with longer QTc.

The result of the study supports a causal role of BMI on multiple ECG indices that have previously been associated with atrial and ventricular arrhythmic risk. Joyarani et al. [19] found, in their study of healthy young women, that the P wave duration and the P wave dispersion were higher in obese women. The P wave dispersion which is the difference between maximum and minimum P wave duration has been recently defined as a new electrographic marker of atrial fibrillation [19]. Sun G et al. [5] who studied 5,556 students aged 5 to 18 years reported that in children and adolescents, general and abdominal obesity is associated with longer PR intervals, wider QRS duration and a leftward shift of frontal P wave, QRS, and T wave axis independent of age, sex, ethnicity and blood pressure. Sadoh et al. [20] compared ECG findings in obese and normal Nigerian adolescents and found that the obese subjects had a higher resting heart rate.

As highlighted above, studies have demonstrated associations between anthropometric indices and ECG intervals or amplitudes. However, data on these associations remain limited in African adolescents whose growth patterns and cardiovascular profiles may differ. Understanding how anthropometric indices relate to ECG parameters during adolescence may improve interpretation and help in differentiating normal physiological variations from early signs of cardiac pathology. This knowledge is particularly relevant in sub-Saharan Africa, where resources for advanced cardiac imaging are limited and the ECG remains the first-line screening tool for cardiovascular evaluation. The objective of this study was to determine the association between anthropometric indices-specifically height, weight, and body mass index-and electrocardiographic variables among healthy Black adolescents in Ido-Ekiti, Nigeria.

Methods     

Study design: this study employed a cross-sectional descriptive design to determine the association between anthropometric indices and electrocardiographic (ECG) variables among apparently healthy Black adolescents from 12 public secondary schools in Ido-Osi LGA, Ekiti State, in southwest Nigeria.

Study setting and population: the study was conducted in secondary schools within Ido-Ekiti, southwestern Nigeria. It is a semi-urban population, and its residents are predominantly of Yoruba ethnicity. Data collection was performed on school premises under standardized conditions. Participants were 1,194 apparently healthy adolescents aged 10-19 years recruited from selected secondary schools in Ido-Ekiti. Inclusion criteria required participants to be of Black African descent, free from known cardiovascular or systemic illness, and not on medications affecting cardiac function. Exclusion criteria included blood pressure profiles higher above the 95^th percentile for age, sex, and height using the blood pressure tables in the fourth report on the diagnosis, evaluation, treatment of high blood pressure in children adolescents [21]; obesity or underweight as determined by the body mass index (BMI) using the CDC Chart, acute febrile or non-febrile illness, use of any medications that could adversely affect cardiovascular function, dysmorphic features; facial anomalies, chest wall and other musculoskeletal deformities; and history and/ or physical examination signs that suggest cardiac disease or any other chronic condition that may affect cardiovascular function.

Sample size determination: sample size was determined using Fisher's formula for descriptive cross-sectional studies assuming a prevalence of 50% (since no prior estimate was available), a 95% confidence level, and a precision of 5%. A finite population correction was applied and the sample size adjusted to account for anticipated non-response. The calculated sample size was adjusted for the multistage sampling design using a design effect of 3. A total of 1194 adolescents were recruited for the study [22].

Sampling method: a multistage sampling technique was employed to select participants from the different classes in various schools. Stage 1: selection of 12 schools from across the communities via simple random sampling. One school was selected per community. Stage 2 and 3: the number of participants per school and from each class was determined using a proportionate sampling technique. Stage 4: selection of arms of classes by simple random sampling. Stage 5: selection of number of eligible participants per arm of a class by systematic sampling method using a sampling frame.

Data collection and measurements: data were collected directly from study participants using standardised study proforma, anthropometric measurements, physical examination, and resting 12-lead electrocardiography. Anthropometric measurements (height, weight) were obtained using standard equipment. Body mass index was calculated as weight (kg)/height (m²). Following verbal assent and written parental informed consent, the procedure was explained in full detail to the participants. Information such as age, sex, history of smoking, consumption of alcohol, involvement in athletic activities, drug use, and presence of any medical complaints and symptoms were sought and obtained. Privacy was ensured for all participants. A thorough physical examination was conducted and height (cm), weight (kg), temperature (°C), blood pressure (mmHg) and pulse rate (beats per minute) were recorded. Body weight and height measurements were performed by well-trained personnel. Participants were weighed barefoot, to the nearest kilogram, with light clothing on. This was done using an analogue weighing scale. Height was measured with participants barefoot to the nearest 0.1 cm, using a stadiometer. Body mass index was calculated as weight (kg) divided by height squared (m²).

Standard resting 12-lead ECGs were recorded for participants in a supine position based on the American Heart Association [23] specification using a portable electrocardiograph Zoncare™ ZQ 1203G (25 mm/s, 10 mm/mV calibration) with a frequency range of 1 - 150 Hz and a sampling rate of 8,000 Hz. The trained research assistant served as a chaperone throughout the procedure. ECG measurements were interpreted and read manually using an ECG ruler and digital ECG calliper. Visual inspection and interpretation were performed using a magnifying glass (x10). The P, Q, R, S, and T wave amplitudes per lead of 12 leads were measured in millivolts. The duration of the waves, intervals and segments namely P wave duration, QRS duration, T wave duration, PR interval, RR interval, and QT interval were measured in milliseconds. Corrected QT interval was calculated using the Bazett formula [24].

Values for the P wave axis, T wave axis and QRS axis were the only computer-generated values recorded. Measurements were verified by two independent investigators blinded to the anthropometric data.

Variables: exposure (independent) variables were anthropometric indices: height, weight, and body mass index (BMI). Outcome (dependent) variables were ECG parameters including heart rate, P-wave duration and amplitude, P-wave axis, PR interval, QRS duration, QTc, QRS axis, and wave amplitudes. Covariates were age and sex.

Data analysis/statistical methods: data analysis was conducted using SPSS version 25.0. Descriptive statistics (means, standard deviations) summarized anthropometric and ECG parameters. An independent t-test was used for group comparisons of males and females. Correlations between ECG variables and anthropometric indices were assessed using Pearson's correlation coefficient for normally distributed continuous variables with linear relationships and Spearman's rank correlation coefficient when the assumptions for Pearson correlation coefficient were not met. A test of normality was performed with the Kolmogorov-Smirnov test. To identify independent associations between anthropometric indices and ECG parameters, multivariate linear regression models were fitted with each ECG parameter as the outcome variable, anthropometric indices (height, weight and BMI) as the primary exposure variables, and age and sex included as covariates.

Bias: selection bias was minimized through the use of a multistage probability sampling technique. Schools were selected by simple random sampling, while participants were recruited using proportionate and systematic sampling methods to ensure adequate representation of eligible students across the selected schools and class levels. Measurement bias was reduced by standardizing data collection and calibration of instruments. Standardized protocols were used for anthropometric and ECG measurements. Interpretation of ECG was verified independently by two qualified physicians who were blinded to the participants' anthropometric data.

Results     

Descriptive data: the ages of the participants ranged from 10 to 19 years, with a mean (SD) of 14.3 ± 2.0 years. There were 530 males and 664 females, giving a male-to-female ratio of 1: 1.3.

Outcome data: the overall mean (SD) weight was 45.0 (8.8) kg; 44 (9.6) kg for males and 46 (8.1) kg for females (range 25 to 78 kg). The overall mean (SD) of their heights was 154.8 (9.8) cm; 154.7 (11.7) cm and 154.9 (8.0) cm for males and females respectively (range 128.3 to 182.0 cm) while the overall mean (SD) BMI was 18.7 (2.1) kg/m²; 18.1 (1.7) kg/m² for the boys and 19.1 (2.2) kg/m² for the girls with a range of 14.7 to 30.0 kg/m². The anthropometric parameters (mean values for height, weight and BMI) of male and female participants were categorized by age groups and shown in Table 1. The BMI values were higher in females across all age groups, and these differences were statistically significant (p < 0.001). In the age groups 10 - 13 and 14 - 16 years, females were heavier, while the males became heavier in the 17-19-year age group. These differences were statistically significant. In early adolescence (10-13 years), the female participants were taller, but by late adolescence (17-19 years), the males had become taller. Mean anthropometric indices were within normal limits for age and sex. Table 2 shows the mean values of heart rate, QRS axis, P wave duration, PR intervals, QRS duration and QTc of males and females across the adolescent age groups. These values were within physiological limits. Older adolescents showed lower mean heart rates. The values for ECG (P, Q, R, S, T) wave amplitudes were also obtained, with males having higher values.

The details of the relationship between height, weight and BMI with ECG lead-independent parameters are shown in Table 3. Height significantly correlated negatively with heart rate and T wave duration but positively correlated with P wave axis, P wave duration, RR interval, QT interval and QRS duration. Weight significantly correlated negatively with HR and positively correlated with the P axis, P wave duration, PR interval, RR interval, QT interval, and QRS duration. Body mass index (BMI) significantly correlated negatively with heart rate and QRS axis and positively correlated with P axis, P wave duration, PR interval, RR interval, QT interval and QRS duration (p < 0.05). The correlation of height, weight, and BMI with corrected QT interval (QTc) was not statistically significant (p > 0.05). The noted positive or negative correlations stated above, though weak, are statistically significant.

The details of the relationship between height, weight and BMI with the ECG wave (P, Q, R, S, T, U) wave amplitude across all the leads are shown in Table 4. Height, weight and BMI were weakly but significantly negatively correlated with P wave amplitude in the precordial leads except for height in V1 and V3. Height, weight and BMI were significantly negatively correlated to Q wave in leads I, II and V6 and significantly negatively correlated to R wave amplitude in the precordial leads (Table 4). Height and weight were weakly but significantly negatively correlated with S wave depth in the left precordial leads, while BMI was significantly negatively correlated with S wave depth in all the precordial leads.

The correlation of height with T wave amplitude was significant in V1 and V2, while the negative correlation of weight with T wave amplitude was significant in the limb leads, V1 and left precordial leads. BMI was significantly negatively correlated with T wave amplitude in all leads except aVL. Details of the multiple regression analysis of height, weight and BMI as predictors of ECG parameters are shown in Table 5 and Table 5.1. The standardized beta coefficients for the anthropometric indices (height, weight, and BMI) indicated the contribution of each independent variable to the dependent ECG parameters.

This analysis showed that weight had a negative predictive effect (β = -0.297) on heart rate, i.e for every unit (kg) increase in weight, there is a 0.297 times reduction in the heart rate, and this was found to be statistically significant (β = 0.297, p-value 0.044). Height and BMI had positive predictive and statistically significant effects on the T wave axis (β = 0.338, p < 0.001; β = 0.245, p = 0.006, respectively). Also, weight had a strong negative predictive effect on the T wave axis, which was also significant (β = -0.509, p = 0.001). All the anthropometric parameters were found to be significant independent predictors of the T axis and P wave duration, while only BMI significantly predicted QTc (β = 0.194, p = 0.030). In contrast, the anthropometric variables were found not to be predictors of T wave duration, PR interval, QTI, RRI and QRS duration.

As shown in Table 6, height, weight and BMI were found to be significant predictors of R wave amplitude in left precordial leads (p <0.001) and Q wave amplitude in V5. Weight and BMI had a significant predictive effect on S wave amplitude in V2. Only BMI had a significant predictive effect on T wave amplitude in V5. When the analysis was adjusted for height and weight, the predictive effect of the anthropometric indices on the electrocardiographic variables differed slightly. There was an increased predictive effect on the ECG axes but decreased effect on heart rate and ECG durations. Similarly, the effect on the ECG wave amplitudes decreased except for BMI, whose predictive effect increased significantly.

Discussion     

There is a paucity of data on the association between anthropometric indices and ECG parameters in adolescents. Most studies have focused on relationships between ECG characteristics and obesity or its indices rather than attempting to define the relationship between anthropometric indices and ECG variables in healthy persons [5,13,15]. This study focused on describing the relationship between anthropometric indices - height, weight, and BMI and ECG variables in healthy black adolescents. This study demonstrated significant associations between anthropometric indices and ECG variables among adolescents. Height, weight, and BMI correlated positively with ECG intervals and negatively with heart rate and wave amplitudes.

Height, weight, and BMI were positively correlated with P-wave duration, P-wave axis, QRS duration, QT interval, and RR interval but negatively correlated with heart rate and precordial wave amplitudes. Larger body size was thus linked to slower conduction, lower heart rate, and longer depolarization and repolarization intervals. In this study, height, weight, and BMI correlated negatively with heart rate. This is inconsistent with the findings of Sun et al. [5], who reported a higher heart rate with an average increase of 3 to 4 beats per minute in the obese group compared with those with a normal weight. However, Sadoh et al. [20] also observed a higher heart rate in obese and overweight adolescents in Benin City, Nigeria. In this study, height, weight, and BMI had a significant negative correlation with R wave amplitude in the precordial leads.

Similar to the findings by Mandic et al. [25], this study showed that height had a significant positive correlation with QRS duration. In addition, height also correlated positively with the P wave axis, P wave duration, QT interval, and QRS duration. The study by Santini et al. [13] among adolescents found a relationship between height and the cardiac conduction system, which might explain the positive correlation between height and P wave and QRS duration. A higher frequency of the right bundle branch block (RBBB) was observed in tall subjects. Kofler et al. [12] reported that tall individuals had significantly longer PR intervals as compared with short individuals, suggesting a relationship between the heights of adults and an altered cardiac conduction system. The present study, done in black adolescents, however, found a weak positive correlation between height and PR interval, which was not statistically significant. This may imply racial differences in the association between height and cardiac conduction or suggest that such a finding is not observed in adolescents.

In this study, weight correlated negatively with heart rate but positively with the P wave axis, the P wave duration, the PR interval, the RR interval, the QT interval and the QRS duration. This is similar to findings from earlier reports. Tan et al. [17] reported that weight had a positive correlation with P wave duration and QRS duration, and Mandic et al. [25] also found that weight correlated positively with QRS duration. However, the finding of a negative correlation of weight with R wave amplitude in this study is at variance with reports by Tan et al. [17] who stated that increasing weight correlated with increasing R wave voltage. The reason for this is not clear but may be related to racial differences or the systematic differences in precordial lead placement. The correlation between height, weight, and BMI with QTc in this study was not statistically significant.

In our study, BMI correlated negatively with heart rate and QRS axis but positively with the P wave axis, the P wave duration, the PR interval, the RR interval, the QT interval and the QRS duration. BMI also correlated negatively with amplitudes of the waves (Q, R, S and T). These findings are congruent with other studies [1,5,15]. Sun et al. [3] observed that BMI correlated negatively with the QRS axis and positively with PR interval and QRS duration. Swamy et al. [15] also found that an increase in BMI was associated with a longer PR interval and wider QRS duration. Maruyama et al. [1] also observed that an increase in BMI was associated with the lengthening of PR interval and QRS duration and a leftward shift of the QRS axis. Although the correlations observed in this index study were weak, they were found to be statistically significant. Mandic et al. [25] had earlier reported that ECG measurements correlated poorly with body dimensions.

In the study by Sadoh et al. [20], the differences between mean values of the PR interval, QRS duration, QTc, and QRS axis of 49 obese and overweight subjects and 49 controls were not statistically significant. This might have been due to the small sample size studied. PR interval indicates intra-atrial and atrioventricular conduction time. PR prolongation observed in the high BMI group in some of the earlier studies is probably due to increased intra-atrial conduction rather than atrioventricular conduction delay because an increase in left atrial size is one of the subclinical cardiac remodelling observed in obese youth. The leftward shift of the QRS axis in the obese is possibly explained by an upward shift of the diaphragm and a more horizontal anatomic position of the heart due to abdominal adiposity [1,26,27]. Recorded voltages are determined by the balance between original ECG voltages from the left ventricle and attenuation through the chest wall. The thickened chest wall laden with adipose tissue in obese individuals would attenuate original ECG voltages, markedly resulting in lower ECG voltage recording than in normal weight individuals [28].

From the multiple regression analysis, weight was a significant negative predictor of heart rate. For every unit (kg) increase in weight, there was a 0.297-times reduction in the heart rate. All the anthropometric parameters were found to be significant independent predictors of the T wave axis and P wave duration, while BMI significantly positively predicted QTc. This is at variance with the study by Mandic et al. [25] in which height was a predictor of QRS duration, and QTc showed a significant relationship with height.

In the present study, none of the anthropometric parameters significantly predicted the P wave axis, QRS axis, T wave duration, PR interval, RR interval, QT interval, and QRS duration. However, Sun et al. [5] reported that BMI had a positive predictive effect on PR interval and QRS duration but a negative predictive effect on the QRS axis. The disparity noticed with the findings of other studies might be related to racial differences and peculiar characteristics of the population studied. For instance, Mandic et al. [25] studied student-athletes. The exclusion of adolescents who were obese and underweight in this study may have also contributed to some of the associations with BMI and weight not being statistically significant. However, Hassing et al. [27] stated that BMI-related discrete electrocardiographic changes can be observed in normal individuals. In their study of individuals with normal BMI, a higher BMI was independently related to increased P wave duration, P wave dispersion, PR interval, and a leftward shift of the heart axis and decreased Sokolow-Lyon voltage. These results suggest that atrial size may be related to BMI in healthy individuals with normal BMI. Hypothetically, the volume of epicardial and pericardial fat is dependent on BMI in young non-obese individuals. Cardiac fat depositions were found to have metabolic and inflammatory functions which can contribute to the fibrotic remodeling of atrial tissue believed to induce the above electrocardiographic changes that could also be found in obese individuals [27,29-31]. Height and BMI were significant negative predictors, and weight was a positive predictor of R wave amplitude in left precordial leads (LPL) and Q wave in V5. BMI was a significant negative predictor of Q, R, S and T wave amplitude. Mandic et al. [25] also observed that BMI had a significantly negative predictive effect on the amplitude of P, R and T waves.

Interpretation

These findings suggest that body size and composition influence cardiac electrophysiology during adolescence. The positive association between body size and ECG intervals is consistent with reports in other populations, while the inverse relationship with heart rate reflects normal physiological adaptation. There is a clear association between body size and cardiac mass. The association between cardiac mass and surface potential is more complex, primarily because thoracic impedance varies with body composition and body size. Thus, although larger participants have a larger heart mass, they also have greater impedance because of the larger amount of tissue between the myocardial surface and skin surface. This impedance also varies with body morphology, which is a possible reason why BMI was negatively correlated with amplitudes of the P, the QRS, and the T waves [25]. Skin resistance delays the conduction of the impulse from the site of generation to the surface of the skin, leading to an increase in PR interval and/or widening of the QRS complex [32]. The observed prolongation of QRS and QT intervals with higher BMI is clinically relevant as it may predispose to arrhythmic risk [29]. The findings in this study suggest that increased body size prolongs electrical conduction time and reduces ECG amplitudes, possibly due to increased cardiac mass and thoracic impedance.

Limitations

The cross-sectional design limits causal inference. The exclusion of underweight and obese adolescents from the study may have accounted for weak correlations, although significant.

Generalisability

Although the study was regionally based, its relatively large sample and standardized methods enhance external validity for similar African settings. Findings may not generalize to adults or non-Black populations due to ethnic and environmental differences. Broader multicentre studies are needed to confirm these relationships and establish normative ECG standards for African adolescents.

Conclusion     

Anthropometric indices (height, weight and BMI) are significantly associated with electrocardiographic variables in healthy Black adolescents. Increases in height, weight, and BMI correspond with slower cardiac conduction, longer ECG intervals, and lower wave amplitudes. Since changes in anthropometric indices affect ECG parameters and abnormalities in ECG variables are associated with an increased risk of adverse cardiovascular outcomes, modifiable anthropometric indices like weight and BMI must be maintained within the normal range. The findings in this study highlight the importance of considering anthropometric characteristics when interpreting adolescent ECG and support the development of population-specific reference standards for African adolescents.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Conceptualisation - Julia Chigozie Okolugbo and Oluwagbemiga Oyewole Adeodu. Data curation - Okolugbo Julia Chigozie and Gbenga Popoola. Methodology - Julia Okolugbo, Olukemi Tolu Bamigboye-Taiwo, John Akintunde Okeniyi and Oluwagbemiga Oyewole Adeodu. Investigation - Julia Okolugbo, Olukemi Tolu Bamigboye-Taiwo, John Akintunde Okeniyi and Oluwadare Ogunlade. Formal Analysis - Okolugbo Julia Chigozie and Gbenga Popoola. Writing (original draft) - Julia Okolugbo. Writing (Review and editing), Olukemi Tolu Bamigboye-Taiwo, John Akintunde Okeniyi, Oluwadare Ogunlade and Oluwagbemiga Oyewole Adeodu. Supervision - Olukemi Tolu Bamigboye-Taiwo, John Akintunde Okeniyi and Oluwagbemiga Oyewole Adeodu. All the authors have read and aprroved the final version of this manuscript.

Tables     

References