Surprising Evidence of Pelvic Growth (Widening) after Skeletal Maturity

Alexander A. Berger,1 Ryan May,2 Jordan B. Renner,3 Neal Viradia,1 Laurence E. Dahners1
1 2

Department of Orthopaedics, University of North Carolina, CB #7055, Bioinformatics Building, Chapel Hill, North Carolina 27599-7055, Department of Biostatistics, University of North Carolina, CB #7420, 3101 McGavran-Greenberg Hall, Chapel Hill, North Carolina 27599-7055, 3 Department of Radiology, University of North Carolina, CB #7510, 2006 Old Clinic Chapel Hill, North Carolina 27599-7055
Received 22 December 2010; accepted 5 May 2011 Published online 23 May 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jor.21469

ABSTRACT: Following an increase in length and width during childhood and adolescence, skeletal growth is generally assumed to stop. This study investigates the inuence of aging on the dimensions of the pelvis and the L4 lumbar vertebra during adulthood. The dimensions of the pelvis, L4 vertebra, and femoral heads were calculated for 246 patients who had received pelvic and abdominal Computed Tomography scans from the UNC Health Care System. Linear regression analysis determined the signicance of relationships between age and width of the pelvis. There was a strong correlation between increasing patient age and increasing width of the pelvis at the trochanters, (0.333 mm/year of age p<0.0001), at the iliac wings, (0.371 mm/year of age p < 0.0002), and between the femoral heads, indicating that the bony pelvis widens over 20 mm between the ages of 20 and 80. The pelvic inlet did not enlarge over time while the distance between the hips and the femoral head diameter did signicantly increase. The height of L4 did not increase over time, but the L4 width did increase. These correlations were seen in both genders. Surprisingly, our results suggest that the pelvis and L4 vertebra increase in width after skeletal maturity and cessation of longitudinal growth. 2011 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 29:17191723, 2011. Keywords: bone growth; pelvic widening; bone maturation; longitudinal growth; computed tomography

During childhood and young adulthood our bones grow in both length and width. Bone widens by periosteal apposition, a process which slows during aging.1 Bone widening also occurs in later years and this is thought to be an adaptive response to compensate for the loss of strength produced by endocortical bone loss. In a 7year prospective study of over 600 women, Szulc et al.2 reported that rates of endocortical resorption increased with age, while the rate of periosteal apposition decreased with age, resulting in a decrease in the mass of the bone. Other research also supports an increase in skeletal bone width in postmenopausal women as a result of periosteal apposition.3,4 In a previous radiographic study of trochanteric bursitis, we noted that the width of the entire pelvis increased with age after skeletal maturity (unpublished data). We were surprised by this apparent indication of growth in width of the skeleton post maturity. We searched for previous studies indicating such pelvic enlargement and were unable to nd information in the literature regarding post maturity changes in the actual width of the skeleton (as opposed to simply the dimensions of the bones themselves). In critical discussions of these ndings, there were a number of serious questions advanced. We developed several hypotheses that might explain this unexpected nding. The rst was that, as patients became more obese with age, the greater distance between the pelvis and plain radiographic plates would increase, inducing greater magnication on X-rays and apparent increases in pelvic dimensions. Therefore, we embarked on this present study in which we used
Correspondence to: Laurence E. Dahners (T: 1-919-966-3340; F: 1-919-966-6730; E-mail: led@med.unc.edu)
2011 Orthopaedic Research Society. Published by Wiley Periodicals, Inc.

computed tomography data (not subject to magnication) to more carefully investigate the dimensional changes in the pelvic bones. Another hypothesis was forwarded that, because our data represented a cross section of the population rather than a longitudinal follow-up of individuals, we might have an inadvertent selection bias wherein somehow we had selected larger individuals in our older age groups. Thus in this study we elected to also measure the height of the L4 vertebral body as some indication of whether our older patients were also taller (larger overall). The third hypothesis assumed that the pelvis was truly enlarging and that any such widening of the pelvis would be due to periosteal appositional bone formation rather than true enlargement or growth. Therefore, we also measured the width of the pelvic inlet which we expected would diminish or narrow from the addition of periosteal new bone. We also measured the distance between the femoral heads which we assumed would not be changed by periosteal apposition. Finally we measured the diameters of the femoral heads as they do not have a periosteum with which to produce appositional growth.

METHODS
This study was approved by the ethics committee of the University of North Carolina. We included patients age 2079. Diagnostic exclusion criteria included the presence of either pelvic/vertebral column implants or noticeable abnormalities in the normal anatomy of the pelvis/vertebral column such as scoliosis, osteophytosis, or leg length inequality. Patients The researchers used picture archiving and communication System (PACS) to nd patients receiving abdominal/pelvic Computed Tomography (CT) scans at the University of North Carolina Health Care System (UNC HCS) between January
JOURNAL OF ORTHOPAEDIC RESEARCH NOVEMBER 2011

1719

1720

BERGER ET AL.

Figure 2. Width between trochanters versus age.

Figure 1. Measurements taken on pelvis and vertebra.

head distance (center to center) was calculated by adding 2 the calculated radius of the femoral head to the minimum distance between the surfaces of the femoral. This method eliminated possible errors in determining the location of the centers of the femoral heads. The greatest height and width of the L4 vertebra were also measured at its midpoint. Statistical Methods Data analysis for this project was performed using linear regression. Each measurement outcome was separated by gender and plotted against patient age. Linear regression lines were t to each gender group, allowing quantitative estimation of the change in measurement over time. p-Values were then calculated for each line, giving an estimate of the statistical signicance of the relationship seen. Source of Funding There was no external funding source, and funding did not play a role in the investigation.

1, 2000 to April 31, 2009. The researchers collected 246 participants for the study by randomly selecting 21 2 male and 21 2 female patients in each 10-year age group from 20 to 79 (i.e., 20 males and 20 females who were 2029, 20 males and 20 females who were 3039, etc.). Procedure (Methods) The study was a retrospective review of CT radiographs of the abdomen/pelvis. The IMPAX Picture Archiving Communication System (PACS) (AGFA HealthCare, Greenville, SC) was used to digitally retrieve the radiographs and make measurements. All of the distances were measured by using the calipers provided by the PACS software. The software enables straight-line measurements with an accuracy of 0.01 mm. Figure 1 diagrams the measurements made for each subject, which included the distances between the outermost edges of the greater trochanters and of the iliac wings, and the greatest width of the pelvic inlet. The diameter of the femoral head was measured on the axial cut CT for which the AP diameter was the greatest. The inter-femoral

RESULTS
Our results are displayed in Table 1 and graphically in Figures 28 (with triangles representing 10-year means). The mean value for trochanteric width for all patients was 295.3 mm, with females having a mean value of 303.5 mm and males 286.8 mm. The mean

Table 1. Average Pelvic Measurements and Effect Sizes Average (cm) Trochanter width Males Females Iliac wing width Average pelvic inlet width versus age Average femoral head diameter versus age Inter-femoral head distance versus age L4 vertebral height versus age L4 vertebral width versus age 29.53 30.35 28.68 25.44 9.89 4.42 16.73 3.00 4.22 Effect (Slope) mm/Year of Age 0.333 0.366 0.275 0.371 0.026 0.055 0.114 0.014 0.062 Total Increase Age 2080 (mm) 20 22 17 22 1.6 3.3 6.8 0.9 3.7

p-Value <0.0001 0.0001 0.0012 <0.0001 0.536 0.0005 0.0015 0.187 0.0003

JOURNAL OF ORTHOPAEDIC RESEARCH NOVEMBER 2011

SURPRISING EVIDENCE OF PELVIC GROWTH

1721

Figure 3. Width between iliac crests versus age.

Figure 5. Interfemoral head distance versus age.

value for iliac wing width for all patients was 272.6 mm, with females having a mean value of 276.8 mm and males 268.1 mm. The association between age and pelvic widening was signicant. The trochanteric width increased 0.333 mm per year (p < 0.0001) or a total of 20 mm from age 20 to 80. The effect was (0.366 mm/year, p 0.0001) in males and in females (0.275 mm/year, p 0.0012). The iliac wing width increased 0.371 mm/year of age (p < 0.0002) or 22 mm from age 20 to 80. In men the effect was 0.398 mm/year of age (p < 0.0001) while in females the effect was 0.330 mm/year of age (p 0.0003). While the width of the pelvic inlet did not signicantly increase, the distance between the femoral heads and the femoral head diameters did. The distance between the centers of the femoral heads increased 0.114 mm/year of age (p 0.0015) or 6.84 mm from age 20 to 80. Men had a greater increase in this distance of 0.126 mm/year of age (p 0.0136) compared to women (0.050 mm/year of age, p 0.0486). The diameter of the femoral heads also increased 0.055 mm/year of age (p 0.0005) or a total of 3.3 mm

between the ages of 20 and 80. The effect was higher in males (0.058 mm/year of age, p 0.0001) than females (0.043 mm/year of age, p 0.0045). The L4 vertebral height did not increase in a significant manner, but the L4 width did increase signicantly. The L4 width increased 0.052 mm/year of age (p 0.0003). The women had a greater increase in L4 width of 0.071 mm/year of age (p < 0.0001) compared to men who had an increase of 0.044 mm/year of age (p 0.0313).

DISCUSSION
These results continue to surprise us. They do support our unexpected earlier nding that pelvic and vertebral dimensions enlarge in width after skeletal maturity and the cessation of longitudinal growth. The use of CT data has allowed us to discard the hypothesis that our earlier results were due to magnication errors brought on by increased obesity in older subjects. However, though the lack of an increase L4 vertebral height would appear to dismiss the hypothesis that we do not have a selection bias with larger subjects in our older groups, the femoral head data, which we expected would also show no increase (because there is no known mechanism for

Figure 4. Width at pelvic inlet versus age.

Figure 6. Diameter of femoral heads versus age.

JOURNAL OF ORTHOPAEDIC RESEARCH NOVEMBER 2011

1722

BERGER ET AL.

Figure 7. L4 vertebral height versus age.

enlargement of the femoral heads) did show enlargement of the femoral heads with age. We also were unable to conrm our third hypothesis, that widening of the pelvis was simply due to periosteal appositional bone formation, because we observed no signicant change in the pelvic inlet width whereas appositional formation would presumably result in a reduction in the width of the inlet. If periosteal apposition is due to the previously described bone enlargement in response to osteoporosis,2 one would expect that there would be more marked differences between the enlargement of male and female subjects as well as a signicant change in the slope of the line (as represented by the triangular decade means) after age 45 in women, which we did not observe in this data. The signicant increases in inter-femoral head distance also cannot be explained by simple periosteal apposition and would appear to require some conformational change in the pelvis. Finally, it is difcult to understand the enlargement of the femoral head that was detected. A large knowledge gap exists in terms of effects of ageing on the femoral head diameter. One study done the examining the relationship of the radiographic joint space width in the hip with age found no changes with age.5 Our

study is the rst to specically look at the femoral head. Our rst thought was that the enlargement was due to calcication of the deeper layers of the cartilage (apparent enlargement only). It may be, but 3.3 mm of enlargement of the diameter would represent 1.6 mm of calcication of the cartilage on each side of the head over the years from age 20 to 80 which seems to be a great deal of calcication as the total thickness of the cartilage is only 23 mm (joint space 45 mm). The most important limitation of this study, like the rst study, is its cross sectional nature. Following the same individuals longitudinally from age 20 to 80 would be ideal but we are unaware of any archives that would allow such an analysis. Certainly, the desirable accuracy of CT (unaffected by magnication) has only been available since the 1980s. We do hope, however, that other investigators may have access to an archive of plain radiographic data following a set of the same patients and extending over many decades. Such an archive could be used to conrm or refute our ndings (even though such an effort would be compromised to some degree by magnication issues). Additonally, due to the cross sectional nature of the study, we had to assume that the populations studied for each age group were similar, and reected the community as a whole. However, it is possible that patients with a higher body mass index are more likely to have CTs performed, and that the average BMI for each age group is higher with each decade of age. The ethical considerations of performing serial CT scans in healty subjects might be addressed by future research in other vertebrates which may help to address some of these concerns. Other limitations to this present study include the possibility that our randomly selected patient population may have had unrecognized bone deformities, but we believe that the number of patients with such issues would have been small enough to have limited effects. The sample size of 246 subjects was also relatively small for a population study, nonetheless the statistical signicance of the detected increases in pelvic dimensions were strong.

CONCLUSION
To our surprise, this study appears to conrm that the widths of the dimensions of the pelvis and L4 vertebral body as well as the distance between, and the diameter of, the femoral heads continue to enlarge after skeletal maturity and the cessation of longitudinal growth.

ACKNOWLEDGMENTS
The authors of this article have no competing nancial interests. The authors would like to acknowledge the support staff of the University of North Carolina department of Orthopaedics and Radiology for their contributions to the project.

REFERENCES
Figure 8. L4 vertebral width versus age.
JOURNAL OF ORTHOPAEDIC RESEARCH NOVEMBER 2011

1. Balena R, Shih M-S, Partt A. 1992. Bone resorption and formation on the periosteal envelope of the ilium: A

SURPRISING EVIDENCE OF PELVIC GROWTH histomorphometric study in healthy women. J Bone Miner Res 7:14751482. 2. Szulc P, Seeman E, Duboeuf F, Sornay-Rendu E, Delmas PD. 2006. Bone fragility: Failure of periosteal apposition to compensate for increased endocortical resorption in postmenopausal women. J Bone Miner Res 21:18561863. 3. Ahlborg HG, Johnell O, Turner CH, Rannevik G, Karlsson MK. 2003. Bone loss and bone size after menopause. N Engl J Med 349(4):327334.

1723

4. Bouxsein ML, Myburgh KH, van der Meulen MC, Lindenberger E, Marcus R. 1994. Age-related differences in cross-sectional geometry of the forearm bones in healthy women. Calcif Tissue Int 54(2) 113118. 5. Goker B, Sancak A, Arac M, Shott S, Block JA. 2003. The radiographic joint space width in clinically normal hips: Effects of age, gender, and physical parameters. Osteoarthritis Cartilage 11(5) 328.

JOURNAL OF ORTHOPAEDIC RESEARCH NOVEMBER 2011