Computed Tomography Measurement of Rib Cage Morphometry in Emphysema

Background

Factors determining the shape of the human rib cage are not completely understood. We aimed to quantify the contribution of anthropometric and COPD-related changes to rib cage variability in adult cigarette smokers.

Methods

Rib cage diameters and areas (calculated from the inner surface of the rib cage) in 816 smokers with or without COPD, were evaluated at three anatomical levels using computed tomography (CT). CTs were analyzed with software, which allows quantification of total emphysema (emphysema%). The relationship between rib cage measurements and anthropometric factors, lung function indices, and %emphysema were tested using linear regression models.

Results

A model that included gender, age, BMI, emphysema%, forced expiratory volume in one second (FEV₁)%, and forced vital capacity (FVC)% fit best with the rib cage measurements (R² = 64% for the rib cage area variation at the lower anatomical level). Gender had the biggest impact on rib cage diameter and area (105.3 cm²; 95% CI: 111.7 to 98.8 for male lower area). Emphysema% was responsible for an increase in size of upper and middle CT areas (up to 5.4 cm²; 95% CI: 3.0 to 7.8 for an emphysema increase of 5%). Lower rib cage areas decreased as FVC% decreased (5.1 cm²; 95% CI: 2.5 to 7.6 for 10 percentage points of FVC variation).

Conclusions

This study demonstrates that simple CT measurements can predict rib cage morphometric variability and also highlight relationships between rib cage morphometry and emphysema.     

Respiratory cross-sectional area-flux measurements of the human chest wall

A new device that utilizes the voltages induced in separate coils encircling the rib cage and abdomen by a magnetic field is described for measurement of cross-sectional areas of the human chest wall (rib cage and abdomen) and their variation during breathing. A uniform magnetic field (1.4 X 10(-7) Tesla at 100 kHz) is produced by generating an alternating current at 100 kHz in two square coils, 1.98 m on each side, parallel to the planes of the areas to be measured and placed symmetrically cephalad and caudad to these planes at a mean distance of 0.53 m. We demonstrated that the accuracy of the device on well-defined surfaces (squares, circles, rectangles, ellipses) was within 1% in all cases. Observed errors are due primarily to small inhomogeneities of the magnetic field and variation of the orientation of the coil relative to the field. Using a second magnetic field (80 kHz) perpendicular to the first, we measured the errors due to nonparallel orientation during quiet breathing and inspiratory capacity maneuvers. In 10 normal subjects, orientation effects were less than 2% for the rib cage and less than 0.7% for the abdomen. In five of these subjects, orientation effects at functional residual capacity in lateral and seated postures were generally less than or equal to 5%, but estimated tidal volume during spontaneous breathing was comparable to measurements in the supine posture. In five curarized patients, we assessed the linearity of volume-motion relationships of the rib cage and abdomen, comparing cross-sectional area and circumference measurements. Departures from linearity using cross-sectional areas were only one-third of those using circumferences. In seven normal subjects we compared cross-sectional area measurements with respiratory inductive plethysmography (RIP) and found comparable estimates of lung volume change over a wide range of relative rib cage contributions to tidal volume (-5 to 105%), with slightly higher standard deviations for the RIP (SD = 10% for RIP; SD = 4% for cross-sectional area).     

Potential effects of ethnicity in genetic and environmental sources of variability in the stature, mass, and body mass index of children

We examined sources of variability in stature, body mass, and body mass index (BMI) in families of black and white elementary schoolchildren from Philadelphia, Pennsylvania. The sample consisted of 445 black and 379 white children, 7-13 years old, and their parents (total n = 2016). The sample was distributed among 596 nuclear families, each representing an independent pedigree. Maximum-likelihood-based variance decomposition methods were used to simultaneously estimate ethnic group-specific effects of genes, sex, age by sex, and unmeasured environmental factors on stature, body mass, and BMI. Likelihood ratio tests were performed to assess the significance of h2 estimates and differences in sigma g and sigma e between black and white families. Genes account for moderate proportions of the phenotypic variance (h2) of these traits in black and white children. In black and white children, respectively, h2 estimates were 0.37 and 0.53 for stature, 0.37 and 0.31 for body mass, and 0.38 and 0.24 for BMI (p < 0.0005). Although the differences in h2 between ethnic groups were not significant (stature, p = 0.23; body mass, p = 0.49; BMI, p = 0.14), black children exhibited a significantly greater total residual phenotypic standard deviation (sigma e and sigma g) in body mass and BMI and a significantly greater sigma e for stature compared with white children. The larger residual phenotypic variance in the black sample is likely due to exposure to unmeasured environmental factors that are not accounted for in this model. Given that sigma g for stature is not significantly different between ethnic groups, the slightly lower estimates in black children are due to the increased contribution of the environment to the phenotypic variance in this trait.     

A model for studies of the deformable rib cage.

An earlier model for the study of rib cage mechanics was modified so that rib deformity in scoliosis could be better represented. The rigid ribs of that model were replaced by five-segment deformable ribs. Literature data on cadaver rib mechanical behavior were used to assign stiffnesses to the new individual model ribs so that experimental and model rib deflections agreed. Shear and tension/compression stiffnesses had little effect on individual rib deformation, but bending stiffnesses had a major effect. Level-to-level differences in mechanical behavior could be explained almost exclusively by level to level differences in the rib shape. The model ribs were then assembled into a whole rib cage. Computer simulations of whole rib cage behaviors, both in vivo and in vitro, showed a reasonable agreement with the measured behaviors. The model was used to study rib cage mechanics in two scolioses, one with a 43 degrees and the other with a 70 degrees Cobb angle. Scoliotic rib cage deformities were quantified by parameters measuring the rib cage lateral offset, rib cage axial rotation, rib cage volume and rib distortion. Rib distortion was quantified both in best-fit and simulated computer tomography (CT) scan planes. Model rib distortion was much smaller in best-fit planes than in CT planes. The total rib cage volume changed little in the presence of the scolioses, but it became asymmetrically distributed.     

A model for studies of mechanical interactions between the human spine and rib cage

A three-dimensional mathematical model useful for studies of the mechanics of the human skeletal thorax is described. To construct this model, rib cage elements are incorporated into a previously reported model of the thoracolumbar spine. The vertebrae and bony portions of the ribs and sternum are idealized as rigid bodies. The behavior of the discs, ligaments and costal cartilages are modelled by deformable elements. Appropriate geometric and stiffness property data are assigned to the elements of the model. In constructing the model, it was found that the mechanical response of the costo-vertebral joint is strongly influenced by articulation geometry. Although rigid bodies were used to model calcified portions of the ribs, the model predicted rib cage deformations in close agreement with those measured experimentally. These studies indicate that the rigid body motion of calcified portions of the rib makes a major contribution to the deformation of the rib cage in response to certain types of loadings. Quantitative results are also reported on the roles the rib cage plays in bending responses of the spine, the lateral stability of the spine, and the production and correction of several scoliotic deformities.     

Paradoxical inward rib cage motion during rapid eye movement sleep in infants and young children

In neonates, rib cage motion on inspiration during rapid eye movement sleep is almost exclusively paradoxical. We wondered whether or not duration of paradoxical inward rib cage motion on inspiration during rapid eye movement sleep decreases in infancy and early childhood. Thirteen healthy infants from 7 to 31 months of age were tested during natural afternoon naps. Electroencephalogram, electrooculogram and electromyogram were all recorded. Airflow was measured by nasal and buccal thermistors, abdominal and rib cage anteroposterior diameters by magnetometers. Transcutaneous partial pressure of O2 was monitored. Diaphragmatic electromyographic activity was recorded using surface electrodes. The average total sleep time was 138 min ranging from 107 to 186 and rapid eye movement sleep time amounted to 15% of total sleep time ranging from 6 to 25. During rapid eye movement sleep, the total duration of paradoxical inward rib cage motion was measured and expressed as a percentage of rapid eye movement sleep time. We found that duration of paradoxical inward rib cage motion during rapid eye movement sleep decreased significantly with age (r = -0.66, P less than 0.02) which may be explained by the changes in chest wall compliance and geometry of the rib cage occurring with growth. We observed no decrease in transcutaneous partial pressure in O2 during paradoxical inward rib cage motion during rapid eye movement sleep in infants in contrast to that reported in neonates.     

Quantification of age-related shape change of the human rib cage through geometric morphometrics

The aim of this study is to quantify patterns of age-related shape change in the human thorax using Procrustes superimposition. Landmarks (n=106) selected from anonymized computed tomography (CT) scans of 63 adult males free of skeletal pathology were used to describe the form of the rib cage. Multivariate linear regression was used to determine a relationship between landmark location and age. Linear and quadratic models were also investigated. A permutation test employing 1×10⁵ random trials was used to assess the model significance for both model formulations. Linear relationships between the centroid size (CS) of a landmark set and the corresponding individual's height, weight, and BMI were conducted to enable scaling of the dimensionless results from the Procrustes analysis. A significance level of =0.05 was used for all tests. The average age of the study subjects was 57.0±17.3 years. Complete landmark sets were obtained from most of the scans (44 of 63). The quadratic relationship between the age and landmark location was found to be significant (p=0.037), thereby establishing a relationship between the age and thoracic shape change. The linear relationship was mildly significant as well (p=0.073). Significant relationships between the centroid size of the dataset and subject weight, height and BMI were determined, with the best-correlated value being weight (p=0.002, R²=0.22). Landmark datasets calculated using the quadratic model exhibited shape change consistent with the clinical observations (increasing kyphosis and rounding of the thoracic cage). Procrustes superimposition represents a potential improvement in the approach used to generate computational models for injury biomechanics studies. The coefficients from the quadratic model are provided and can be used to generate the complete set of model landmark data points at a given age.     

Development and validation of subject-specific finite element models for blunt trauma study

This study developed and validated finite element (FE) models of swine and human thoraxes and abdomens that had subject-specific anatomies and could accurately and efficiently predict body responses to blunt impacts. Anatomies of the rib cage, torso walls, thoracic, and abdominal organs were reconstructed from X-ray computed tomography (CT) images and extracted into geometries to build FE meshes. The rib cage was modeled as an inhomogeneous beam structure with geometry and bone material parameters determined directly from CT images. Meshes of soft components were generated by mapping structured mesh templates representative of organ topologies onto the geometries. The swine models were developed from and validated by 30 animal tests in which blunt insults were applied to swine subjects and CT images, chest wall motions, lung pressures, and pathological data were acquired. A comparison of the FE calculations of animal responses and experimental measurements showed a good agreement. The errors in calculated response time traces were within 10% for most tests. Calculated peak responses showed strong correlations with the experimental values. The stress concentration inside the ribs, lungs, and livers produced by FE simulations also compared favorably to the injury locations. A human FE model was developed from CT images from the Visible Human project and was scaled to simulate historical frontal and side post mortem human subject (PMHS) impact tests. The calculated chest deformation also showed a good agreement with the measurements. The models developed in this study can be of great value for studying blunt thoracic and abdominal trauma and for designing injury prevention techniques, equipments, and devices.     

Mechanical coupling of upper and lower canine rib cages and its functional significance

We studied rib cage distortability and reexamined the mechanical action of the diaphragm and the rib cage muscles in six supine anesthetized dogs by measuring changes in upper rib cage cross-sectional area (Aurc) and changes in lower rib cage cross-sectional area (Alrc) and the respective pressures acting on them. During quiet breathing in the intact animal the rib cage behaved as a unit (Aurc: 14.6 +/- 7.9 vs. Alrc: 15.1 +/- 9.6%), whereas considerable distortions of the rib cage occurred during breathing after bilateral phrenicotomy (Aurc: 21.0 +/- 5.1 vs. Alrc: 7.0 +/- 4.8%). These distortions were even more pronounced during phrenic nerve stimulation and separate stimulation of the costal and crural parts of the diaphragm (e.g., phrenic nerve stimulation; Aurc: -7.1 +/- 5.1 vs. Alrc: 6.9 +/- 3.5%). During the latter maneuvers the upper rib cage deflated along the relationship between upper rib cage dimensions and pleural pressure obtained during passive deflation, whereas the lower rib cage inflated close to the relationship between lower rib cage dimensions and abdominal pressure obtained during passive inflation. The latter relationship is expected to differ between costal and crural stimulation, since costal action has both an appositional and insertional component and crural action only has an appositional component. The difference between costal and crural stimulation, however, was relatively small, and the slopes were only slightly steeper for the costal than for the crural stimulation (2.9 +/- 1.2 vs. 2.2 +/- 1.0%.(ABSTRACT TRUNCATED AT 250 WORDS)     

In vitro analysis of kinematics and elastostatics of the human rib cage during thoracic spinal movement for the validation of numerical models

Neither kinematic nor stiffness properties of the rib cage during thoracic spinal motion were investigated in previous studies, while being essential for the accurate validation of numerical models of the whole thorax. The aim of this in vitro study therefore was to quantify the kinematics and elastostatics of the human rib cage under defined boundary conditions. Eight fresh frozen human thoracic spine specimens (C7-L1, median age 55 years, ranging from 40 to 60 years) including entire rib cages were loaded quasi-statically in flexion/extension, lateral bending, and axial rotation using pure moments of 5 Nm. Relative motions of ribs, thoracic vertebrae, and sternal structures as well as strains on the ribs were measured using optical motion tracking of 150 reflective markers per specimen, while specimens were loaded displacement-controlled with a constant rate of 1°/s for 3.5 cycles. The third full cycle was used to determine relative angles and strains at full loading of the spine for all motion directions. Largest relative angles were found in the main loading directions with only small motions at the mid-thoracic levels. Highest strains of the intercostal spaces were detected in the anterior section of the lowest fourth of the rib cage, showing compressions and elongations of more than 10% in all spinal motion planes. Elastostatic rib deformation was generally less than 1%. Rib-sternum relative motions exhibited complex motion patterns, overall showing relative angles below 2°. The results indicate that rib cage structures are not macroscopically deformed during spinal motion, but exhibit characteristic reproducible kinematics patterns.     

An analysis of action of intercostal muscles in human upper rib cage

The actions of the intercostal and paraspinal muscles in stabilizing the human upper rib cage have been analyzed using a geometrically realistic mathematical model of the first six ribs, vertebrae, and associated musculature. The model suggests roles of the deep layers of erector spinae in stabilizing the vertebral column so that it can support the loads placed upon it by the ribs under physiological load. If we assume that the tension exerted by an intercostal muscle is proportional to its local thickness, the model predicts that the observed distribution of intercostal thickness is close to that which minimizes the stresses in ribs when the model is subjected to peak physiological load. The observed shape of the ribs are optimal to withstand the calculated pattern of loading along their length. These calculations raise the hypothesis that the arrangement of intercostal musculature and rib geometry result in an optimally light rib cage, which is capable of withstanding the loads placed upon it. The analysis of the mechanics of the entire model indicates that the geometrical simplifications made in Hamberger's model are not valid when applied to the rib cage.     

Analysis of human chest wall motion using a two-compartment rib cage model.

We present a model of chest wall mechanics that extends the model described previously by Macklem et al. (J. Appl. Physiol. 55: 547-557, 1983) and incorporates a two-compartment rib cage. We divide the rib cage into that apposed to the lung (RCpul) and that apposed to the diaphragm (RCab). We apply this model to determine rib cage distortability, the mechanical coupling between RCpul and RCab, the contribution of the rib cage muscles to the pressure change during spontaneous inspiration (Prcm), and the insertional component of transdiaphragmatic pressure in humans. We define distortability as the relationship between distortion and transdiaphragmatic pressure (Pdi) and mechanical coupling as the relationship between rib cage distortion and the pressure acting to restore the rib cage to its relaxed configuration (Plink), as assessed during bilateral transcutaneous phrenic nerve stimulation. Prcm was calculated at end inspiration as the component of the pressure displacing RCpul not accounted for by Plink or pleural pressure. Prcm and Plink were approximately equal during quiet breathing, contributing 3.7 and 3.3 cmH2O on average during breaths associated with a change in Pdi of 3.9 cmH2O. The insertional component of Pdi was measured as the pressure acting on RCab not accounted for by the change in abdominal pressure during an inspiration without rib cage distortion and was 40 +/- 12% (SD) of total Pdi. We conclude that there is substantial resistance of the human rib cage to distortion, that, along with rib cage muscles, contributes importantly to the fall in pleural pressure over the costal surface of the lung.     

Mechanical coupling of the rib cage, abdomen, and diaphragm through their area of apposition

Although volumetric displacements of the chest wall are often analyzed in terms of two independent parallel pathways (rib cage and abdomen), Loring and Mead have argued that these pathways are not mechanically independent (J. Appl. Physiol. 53: 756-760, 1982). Because of its apposition with the diaphragm, the rib cage is exposed to two distinct pressure differences, one of which depends on abdominal pressure. Using the analysis of Loring and Mead as a point of departure, we developed a complementary analysis in which mechanical coupling of the rib cage, abdomen, and diaphragm is modeled by a linear translational transformer. This model has the advantage that it possesses a precise electrical analogue. Pressure differences and compartmental displacements are related by the transformation ratio (n), which is the mechanical advantage of abdominal over pleural pressure changes in displacing the rib cage. In the limiting case of very high lung volume, n----0 and the pathways uncouple. In the limit of very small lung volume, n----infinity and the pathways remain coupled; both rib cage and abdomen are driven by abdominal pressure alone, in accord with the Goldman-Mead hypothesis. A good fit was obtained between the model and the previously reported data for the human chest wall from 0.5 to 4 Hz (J. Appl. Physiol. 66:350-359, 1989). The model was then used to estimate rib cage, diaphragm, and abdominal elastance, resistance, and inertance. The abdomen was a high-elastance high-inertance highly damped compartment, and the rib cage a low-elastance low-inertance more lightly damped compartment. Our estimate that n = 1.9 is consistent with the findings of Loring and Mead and suggests substantial pathway coupling.     

Effects of the rib cage on thoracic spine flexibility.

Besides protecting the internal organs of the thorax, the rib cage is the site of numerous muscle attachments. It also decreases the overall flexibility of the thoracic spine. This study developed finite element (FE) models of the thoracic spine with and without the rib cage, and the effects of the rib cage on thoracic spine flexibility were determined. The numerical models were validated by comparing the maximum rotation of the models for several loading cases with experimental data in the literature. After adapting the material properties for the discs and ligaments, the calculated maximum rotations differed from the measured median values by less than 1 degrees without the rib cage and by less than 2.5 degrees with it. The rib cage decreased the mean flexibility of the thoracic spine by 23% to 47%, depending on the loading plane. Assuming the ribs to be rigid beams required a corresponding reduction of ligament stiffnesses in order to achieve the same agreement of the maximum rotations with the measured median values. Interconnecting the FE thoracic spine model plus rib cage with the existing detailed FE lumbar spine model improves the simulation of force directions of muscles attached to the rib cage or thoracolumbar spine. In addition, such a model is suitable for determining the effects of lumbar spine implants on spinal balance.     

Action of human respiratory muscles inferred from finite element analysis of rib cage.

The actions of several human respiratory muscles have been inferred from finite element analysis of the rib cage. The human model is based on anatomic and mechanical measurements in dogs and human cadavers. As in an earlier canine model, the external and internal (interosseous) intercostal muscles were found to cause, respectively, inspiratory and expiratory displacements of the rib cage, in agreement with the two-dimensional geometric analysis of Hamberger. When extended to three dimensions, Hamberger's analysis helps explain why muscles at the side of the rib cage produce changes in the anteroposterior diameter, whereas muscles at the front and back of the rib cage cause changes in the transverse diameter.     

Numerical investigations of rib fracture failure models in different dynamic loading conditions

Rib fracture is one of the most common thoracic injuries in vehicle traffic accidents that can result in fatalities associated with seriously injured internal organs. A failure model is critical when modelling rib fracture to predict such injuries. Different rib failure models have been proposed in prediction of thorax injuries. However, the biofidelity of the fracture failure models when varying the loading conditions and the effects of a rib fracture failure model on prediction of thoracic injuries have been studied only to a limited extent. Therefore, this study aimed to investigate the effects of three rib failure models on prediction of thoracic injuries using a previously validated finite element model of the human thorax. The performance and biofidelity of each rib failure model were first evaluated by modelling rib responses to different loading conditions in two experimental configurations: (1) the three-point bending on the specimen taken from rib and (2) the anterior–posterior dynamic loading to an entire bony part of the rib. Furthermore, the simulation of the rib failure behaviour in the frontal impact to an entire thorax was conducted at varying velocities and the effects of the failure models were analysed with respect to the severity of rib cage damages. Simulation results demonstrated that the responses of the thorax model are similar to the general trends of the rib fracture responses reported in the experimental literature. However, they also indicated that the accuracy of the rib fracture prediction using a given failure model varies for different loading conditions.     

Blunt Liver Injury with Intact Ribs under Impacts on the Abdomen: A Biomechanical Investigation

Abdominal trauma accounts for nearly 20% of all severe traffic injuries and can often result from intentional physical violence, from which blunt liver injury is regarded as the most common result and is associated with a high mortality rate. Liver injury may be caused by a direct impact with a certain velocity and energy on the abdomen, which may result in a lacerated liver by penetration of fractured ribs. However, liver ruptures without rib cage fractures were found in autopsies in a series of cases. All the victims sustained punches on the abdomen by fist. Many studies have been dedicated to determining the mechanism underlying hepatic injury following abdominal trauma, but most have been empirical. The actual process and biomechanism of liver injury induced by blunt impact on the abdomen, especially with intact ribs remained, are still inexhaustive. In order to investigate this, finite element methods and numerical simulation technology were used. A finite element human torso model was developed from high resolution CT data. The model consists of geometrically-detailed liver and rib cage models and simplified models of soft tissues, thoracic and abdominal organs. Then, the torso model was used in simulations in which the right hypochondrium was punched by a fist from the frontal, lateral, and rear directions, and in each direction with several impact velocities. Overall, the results showed that liver rupture was primarily caused by a direct strike of the ribs induced by blunt impact to the abdomen. Among three impact directions, a lateral impact was most likely to cause liver injury with a minimum punch speed of 5 m/s (the momentum was about 2.447 kg.m/s). Liver injuries could occur in isolation and were not accompanied by rib fractures due to different material characteristics and injury tolerance.     

Relationship between parasternal intercostal length and rib cage displacement in dogs

The relationship between parasternal intercostal length and rib cage cross-sectional area was examined in nine supine dogs during passive inflation and during quiet breathing before and after phrenicotomy. Parasternal intercostal length (PSL) was measured with a sonomicrometry technique, and rib cage cross-sectional area (Arc) was measured with a Respitrace coil placed around the middle rib cage. During active inspiration as well as during passive inflation, PSL decreased as Arc increased. However, the relationship between PSL and Arc during active inspiration, whether in the intact or phrenicotomized animal, was almost invariably different from that during passive inflation, so that the same increase in Arc was associated with a greater decrease in PSL in the former than in the latter instance. This difference between passive inflation and active inspiration is probably due to the active contraction of the parasternals during inspiration and the consequent caudal displacement of the sternum. In upright humans, the sternum moves cephalad and not caudad during inspiration, so the relationship between PSL and Arc during active breathing might be similar to that during passive inflation.     

Shape and size of the human diaphragm in vivo

Serial computerized tomograph (CT) sections at 5-mm intervals of a human diaphragm in relaxed and contracted states were obtained in one subject while he held his breath and lay supine in a CT scanner. All sections for one state were scanned at the same chest wall configuration as monitored by rib cage and abdominal dimensions, using magnetometers. Sections were scanned at relaxed functional residual capacity and after inspiring approximately 1 liter in such a way that rib cage dimensions increased only slightly. Models of the diaphragm dome in the two states were constructed from the sets of serial sections. Diaphragm length and volume displaced were measured, the zone of apposition of diaphragm to rib cage was mapped, and the line of the diaphragm silhouette in anteroposterior and lateral X-rays identified. Coronal and sagittal sections were constructed. In the inspiration studied, the diaphragm movement displaced 680 ml. Meridian lines in sagittal, coronal, and transverse directions over the right hemidiaphragm dome shortened by 6.7-7.2 cm, but over the left dome by only 4.0-4.3 cm. Lines of X-ray silhouettes were close to meridian lines, and estimates of shortening were similar to those made previously from X-rays. The peculiar saddle shape of the muscle may help the hemidiaphragms to operate independently, the fibers of the saddle acting as an anchor for midline directed fibers of the hemidiaphragm domes. The shape of the diaphragm also has implications for the distribution of transdiaphragmatic pressure and for the kind of distortion of the lower rib cage margin that is seen during inspirations at high lung volume.     

Pubertal timing, hormones, and body composition among adolescent Turkana males

The Turkana, like other East African pastoral groups, are known for their tall adult stature, achieved despite a blunted growth spurt during adolescence and continued growth into the early 20s. To investigate the hormonal mechanisms associated with the pattern of slow and continued adolescent growth, we collected data on hormonal status, height, weight, and trunk skinfolds and ethnographic self-reports of testicular maturation in a cross-sectional sample of 35 nomadic and 37 settled Turkana males aged 14-24. Hormonal determinations included testosterone (T), sex hormone-binding globulin (SHBG), and dehydroepiandrosterone (DHEA) in blood, in addition to urinary DHEA. Self-reports of testicular maturation showed no difference between settled and nomadic subpopulations. However, nomadic boys exhibited significantly higher levels of T, DHEA, and SHBG. Of all the hormones, only SHBG showed a significant relationship with age. Multiple regression models show blood T and SHBG to be significant independent predictors of achieved height as well as weight, controlling for age. Our results suggest that onset of puberty is substantially delayed among Turkana males, and that bioavailable T is related to growth in stature during adolescence. We suggest that SHBG acts to mediate the effects of energy availability on adolescent growth in this energetically limited population. Our findings may also have implications for understanding adolescent growth among Homo erectus.