The influence of sex, age and BMI on the degeneration of the lumbar spine

Abstract

Previous research on lumbar spine osteophyte formation has focused on patterned development and the relation of age and sex to degeneration within the vertebral bodies. The inclusion of osteophytes originating on the laminae and body mass index (BMI) may result in a more complete evaluation. This study investigates lumbar osteophyte development on the laminae and vertebral bodies to determine whether osteophyte development: (i) is related bilaterally, at different lumbar levels, and superior and inferior margins; (ii) on the laminae and vertebral bodies are reciprocally dependent responses; (iii) is correlated with sex, age and/or BMI. Seventy-six individuals (39 females, 37 males) were randomly selected from a modern skeletal collection (Bass Donated Collection). Osteophyte development was scored in eight regions on each vertebra at all five lumbar levels. A factor analysis considered all 40 scoring regions and Pearson's correlation analyses assessed the relatedness of age and BMI with the consequent factors. The factor analysis separated the variables into two similar factors for males and females defined as: (i) superior and inferior vertebral body scores and (ii) superior laminar scores at higher lumbar levels. The factor analysis also determined a third factor for females defined as: (iii) inferior laminar scores at lower lumbar levels. The severity of vertebral body osteophytes increased with age for both sexes. Additionally for females, as BMI increased, osteophyte severity increased for both the superior laminar margins higher in the column and the vertebral bodies. Dissimilarities between the factors in males and females and the correlation of BMI to osteophyte severity exclusively in females provide evidence for different biomechanical processes influencing osteophyte development.

Introduction

Osteophytes on the vertebral bodies and the laminae pose health problems for many individuals. They can be painful and eventually restrict movement. While osteophytes of the veterbal bodies and laminae are extremely common throughout the spine, these conditions are especially prevalent in the lumbar vertebrae due to the demands of bipedal locomotion (Bridges, 1994). Osteophytes of the vertebral body have been studied extensively and scored in multiple ways to examine patterns of sex, age and ethnicity. Laminar osteophytes, however, are rarely mentioned within the literature. Only a handful of studies have documented the presence of laminar osteophytes and even fewer have considered their significance. Additionally, these previous studies utilized ordinal scoring in conjunction with statistical analyses for interval data, which falsely assumes a linearity of the ordinal scale (Stevens, 1946). An ordinal statistical analysis, such as a factor analysis, is preferred with the use of ordinal data because it results in more conservative findings and avoids exaggerated conclusions.

Osteophytes have been described as bony outgrowths on joint margins, although no clear definition exists (Van der Merwe et al., 2006). Several authors have suggested that differences exist between osteophyte types on vertebral bodies. Specifically, some researchers have distinguished between what they call true osteophytes, or traction spurs, from other kinds of bony spurs. They define these true osteophytes as developing in a horizontal direction out from the vertebral body margin (MacNab, 1971; Revell, 1992; Derevenski, 2000). Alternatively, claw-type osteophytes, or enthesophytes, are defined as an ossification of the enthesis of the anterior longitudinal ligament (MacNab, 1971; Resnick, 1992; Derevenski, 2000). However, Pate et al. (1988) indicate that many intermediate varieties of the two vertebral body osteophyte types exist and conclude that the two types are different stages of the same degenerative process. Hence, this study will treat the different vertebral body osteophyte ‘types’ as a single form of degeneration.

Laminar spurs are designated as bony outgrowths at the point of insertion of ligament into bone (Allbrook, 1957). This description is identical to the enthesophyte definition of body osteophytes, yet laminar bony outgrowths are never referred to as osteophytes, only as spurs or ossifications (Shore, 1931; Allbrook, 1957; Williams et al., 1984; Maigne et al., 1992). Because no significant difference exists between the definitions, vertebral body osteophytes and laminar ossifications will both be referred to as osteophytes within this study.

Because of the close proximity of the vertebral laminae and articular facets, differentiating between laminar and apophyseal osteophytes can be difficult. In most cases, a clear separation exists between osteophytic developments in the two regions. In cases of more extreme degeneration, osteophytes of the articular facets and laminae may resemble a continuous osteophyte formation, but the overall severe degeneration of these vertebrae precludes the necessity of analyzing the laminar osteophytes alone. The continuity of laminar osteophytes with apophyseal osteophytes does not obscure the relative severity of osteophyte growth from the laminar margin. Within this study, osteophytes attributed to the laminae will only include osteophytes originating from laminar margins unless no separation exists between the osteophytes of laminae and articular facets.

All previous studies have demonstrated meaningful trends in the relative positions of the differently sized osteophytes on vertebral bodies. A consensus suggests that body osteophytes increase in severity on the lower lumbar vertebrae. The greater degree of osteophyte development in this region is the result of increased loading at the peak of the lumbar lordotic curve, which experiences increased pressure because of its distance from the line of gravity (Nathan, 1962; Eisenstein, 1977; Bridges, 1994; Knüsel et al., 1997; O'Neill et al., 1999; Van der Merwe et al., 2006). Other researchers have indicated that the formation of osteophytes aids in preventing further slippage of intervertebral discs due to loading (Allbrook, 1957; Resnick, 1985; He & Xinghua, 2006). Still other research has shown an association between osteophyte formation and intervertebral disc degeneration (MacNab, 1971; Margulies et al., 1996; Santiago et al., 1997; Goh et al., 2000), although disc degeneration does not always lead to vertebral body osteophyte formation (Revell, 1992; Ferguson & Steffen, 2003; Oishi et al., 2003). Finally, a few studies have found a consistent increase in severity of osteophyte development as age increases (Stewart, 1958; Nathan, 1962; Snodgrass, 2004). Eisenstein (1977) is the only researcher to report no significant trend between osteophyte development and increasing age, but his subjective and small number of grades might have skewed his results.

Although less examined, definite patterns have been discerned in laminar osteophytes as well. The general consensus on laminar margin osteophytes is that they are ossifications of the ligamentum flavum insertions (Allbrook, 1957; Williams et al., 1982; Maigne et al., 1992; Santiago et al., 1997; Viejo-Fuertes et al., 1998). The ligamentum flavum is the most elastic of the spinal ligaments, maintaining a constant state of tension even during relaxation (Putz, 1992). Because the attachment area of this ligament is not as strong as other ligaments and elastin can only partially repair itself, the ligament often ossifies to prevent separation from the bone (Putz, 1992; Viejo-Fuertes et al., 1998). Maigne et al. (1992) examined thoracic vertebrae and asserted that these osteophytes are caused by rotatory strain. They suggest that the vertebrae capable of more rotation also evinced the largest and most frequent incidence of osteophyte development. Shore's (1931) investigation of thoracic vertebrae led her to conclude that laminar osteophytes are more prevalent where the ligamentum flavum is most stressed, namely at the peak of the thoracic curve. This conclusion is in agreement with Williams et al. (1984) who conjectured that the ossifications might be due to the separation of the vertebral arches during upright weight-bearing activities. In terms of life history factor correlations, Williams et al. (1984) found no definite pattern of laminar osteophyte development in relation to age and sex. To date, no studies have examined the relationship between laminar osteophyte development and weight.

Both vertebral body and laminar osteophytes have been shown to develop in a predictable way. Cited factors leading to osteophyte formation in both lumbar regions include microfractures in the bone (Bick, 1955; Radin et al., 1972; Eulderink, 1992; Felson, 2004), damage to articular cartilage (Stecher, 1958; Radin et al., 1972; Eulderink, 1992; Resnick, 1992; Menkes & Lane, 2004), the ossification of ligament and tendon attachment sites (Bick, 1955; Santiago et al., 1997), instability of ligaments and discs (Culver & Pirson, 1956; Allbrook, 1957; Stecher, 1958; MacNab, 1971; Eulderink, 1992; Margulies et al., 1996; Santiago et al., 1997; Andersson, 1998; Felson, 2004; Menkes & Lane, 2004; He & Xinghua, 2006), general malalignment (Margulies et al., 1996; Felson, 2004), increased loading due to strenuous labor and trauma (Kellgren & Lawrence, 1958; Radin et al., 1972; O'Neill et al., 1999), obesity (Bick, 1955; Liu et al., 1997; O'Neill et al., 1999; Oishi et al., 2003; Felson, 2004), lumbar lordosis (Allbrook, 1957; Bridges, 1994; O'Neill et al., 1999) and small cross-sectional area (Radin et al., 1972). A closer examination of the osteophytic patterns within the vertebral bodies and laminae as well as a search for associations between regions and life history factors will provide new information about where and why lumbar osteophyte development occurs. This information can then be used to identify populations at risk for debilitating back problems. This study investigates osteophyte development on the laminae and the vertebral bodies of the lumbar vertebrae to determine whether: (i) correlated development occurs at the different levels, right and left sides, and superior and inferior margins of the laminae and vertebral bodies individually; (ii) osteophyte development on the laminae and vertebral bodies are mutually exclusive processes or reciprocally dependent responses; and (iii) a relationship exists between intrinsic [body mass index (BMI)] and extrinsic (sex and age) biomechanical influences and one or both regions of osteophyte development.

Materials and methods

Seventy-six individuals (39 females, 37 males) were examined from the Bass Donated Skeletal Collection, which is housed at the University of Tennessee in Knoxville, Tennessee. These individuals were randomly selected to better quantify the variation of osteophyte development. Only adult individuals (ages 20–110 years) of European ancestry with associated age and sex information as reported at time of death were included. Standing height and weight at time of death, when available, were used to calculate BMI {mass (kg)/[height (cm)/100]²}. Individuals with obvious pathology, such as ankylosing spondylitis or diffuse idiopathic skeletal hyperostosis, or trauma were excluded from the study to reduce the influence of extraneous factors. Additionally, to facilitate comparison among individuals and across vertebral levels, only individuals with five lumbar vertebrae were included.

Osteophyte development was scored by visual impression in eight different regions using a scale from 0 to 4 on each lumbar vertebra for each lumbar level according to the criteria outlined in Table 1 and in Figs 1 and 2. For the vertebral body, the superior left, superior right, inferior left and inferior right margins were each scored separately. Similarly, for the lamina, the superior left, superior right, inferior left and inferior right margins were each assigned individual scores. These eight regions (Fig. 3) were scored on each lumbar vertebra for all five levels for a total of 40 scores per individual. A measure of intraobserver error was used to ensure the reliability of this new scoring method, although similar tests are not reported by other researchers. This test was carried out by scoring 10 individuals twice, with a 3-week interval between the two trials. For the vertebral body, 85.5% precision was achieved, and for the lamina, 80% precision.

Table 1.

Standards for osteophyte scoring

Score	Description
Vertebral body
Stage 0	No evidence of osteophytosis; smooth rim with no scalloping or osteophytic points (no reactive bone activity)
Stage 1	Minor evidence of osteophytosis; one or two small osteophytic points (< 2 mm in length and width) but no larger osteophytes that protrude above the rim and/or the beginnings of arthritic lipping but no horizontally-projecting lipping
Stage 2	More developed osteophytosis; three or more small osteophytic points or larger osteophytes and/or horizontally-projecting arthritic lipping at least 3 mm in length or the fusion of multiple osteophytic points that protrude
Stage 3	Arthritic lipping/fused osteophytes that extend out either superiorly or inferiorly at least 3 mm in height (towards the center of the vertebral body or the adjacent vertebra)
Stage 4	Either partial or complete fusion of arthritic lipping/fused osteophytes between adjacent vertebrae
Lamina
Stage 0	No evidence of osteophytosis; rolling/smooth contour allowable along the laminar margin but no pronounced bumps, osteophytic points, or raised ridge
Stage 1	Minor evidence of osteophytosis; one or two small osteophytic points (< 2 mm in length and width) but no larger osteophytes
Stage 2	More developed osteophytosis; three or more small osteophytic points or one or two larger osteophytes
Stage 3	Three or more larger osteophytes and/or fused osteophytes (multiple points coming off 1 or more osteophytes) that cover < 50% of the area created by the articular facet and the midline
Stage 4	Three or more larger osteophytes and/or fused osteophytes that cover 50% or more of the area created by the articular facet and the midline

Fig. 1.

Scoring criteria for the vertebral bodies. The arrows indicate vertebral body margins that exemplify each of the five degeneration scores.

Fig. 2.

Scoring criteria for the laminae. The arrows indicate superior or inferior laminar margins that exemplify each of the five degeneration scores.

Fig. 3.

Superior and inferior scoring regions on each lumbar vertebra include: (1) superior left vertebral body margin, (2) superior right vertebral body margin, (3) superior left laminar margin, (4) superior right laminar margin, (5) inferior right vertebral body margin, (6) inferior left vertebral body margin, (7) inferior right laminar margin and (8) inferior left laminar margin.

All statistical analyses were performed with spss version 17.0. A Bonferroni correction was employed (α = 0.05/2 = 0.025) throughout all statistical analyses in order to accommodate repeating the same tests on the male and female subsamples (Perneger, 1998). Averages and standard deviations of age and BMI were calculated for males and females individually (Table 2). Kolmogorov–Smirnov tests were used to check the normality of the age and BMI distributions for both males and females. Two sample t-tests not assuming equal variances were then used for the normally distributed data and Mann–Whitney U-tests for the non-normally distributed data to determine suitability for comparison between the subsamples based on age and BMI.

Table 2.

Descriptive statistics for the sample according to age and BMI

Sex	n	Mean (SD)	Kolmogorov–Smirnov test	Two-sample t-test	Mann–Whitney U-test
Age
Male	37	54.4 (21.2)	0.095 (P = 0.200)	−1.908
Female	39	62.5 (15.1)	0.078 (P = 0.200)	P = 0.061
BMI
Male	26	24.8 (5.7)	0.131 (P = 0.200)		359.00
Female	34	30.0 (11.1)	0.193 (P = 0.020*)		P = 0.216

^*Denotes a significant result at the 0.025 level (two-tailed).

A factor analysis was performed in which all 40 osteophyte scoring regions were simultaneously considered. Because this analysis focused on establishing whether or not correlated osteophyte development occurs at the different sides, levels, and regions, a correlation matrix was employed as the basis of comparison. Similarly, an unweighted least-squares extraction method was utilized to reduce the large number of categorical variables to a smaller number of components while considering both the unique and common variance in the variables. An oblimin rotation of the factor loadings was used for the final results because it provides greater flexibility in searching out patterns in the variables regardless of their correlations as well as allowing factors to be correlated. This rotation produces a pattern matrix, which distinguishes clusters of interrelated variables, as well as a structure matrix, which measures the correlation of the variables with the clusters (Hatcher, 1994). Only the pattern matrix needs to be examined to interpret the factors (Rummel, 1970), hence this analysis will not consider the structure matrix. Males and females were considered individually and then compared. Regardless of the size of the sample, components can be considered significant and reported as long as they either have four or more loadings above 0.600 in absolute value or three or more loadings of 0.800 in absolute value (Guadagnoli & Velicer, 1988). To ensure the significance of factor loadings, the critical value required for significance for an ordinary correlation was doubled (Stevens, 2002). For females, with a sample size of 39 and 40 variables, the doubled critical value is r = 0.632. For males, with a sample size of 37 and 40 variables, the doubled critical value is r = 0.650.

Finally, to investigate any effects that age or BMI may have on osteophyte formation, Pearson's correlation analyses were conducted with age, BMI, and the consequent factors. BMI was used instead of height and weight as separate variables for a better understanding of how these life history factors interact to affect osteophyte development. A small number of the individuals (five females, 11 males) had no associated weight information and therefore no BMI information. Factors were saved as variables by the Bartlett method. Once again, males and females were considered individually and then compared. Figure 4 provides a visual representation of the distribution of age and BMI for the sample.

Fig. 4.

Scatterplot of age and BMI for the male and female subsamples. The lines of best fit depict the different distributions of the two subsamples.

Results

The Kolmogorov–Smirnov tests (Table 2) showed that BMI is not normally distributed in the female sample. Age for both males and females and BMI for males were normally distributed. Accordingly, a two-sample t-test was used and evinced that the mean ages in the male and female samples were not significantly different from each other and were therefore suitable for comparison (Table 2). Likewise, the Mann–Whitney U-test provided evidence that the male and female samples were suitable for comparison based on BMI (Table 2).

In the factor analysis, the first two components in the male and female subsample contain at least four loadings above 0.600, qualifying them as statistically significant (Table 3). The third component in the female subsample, however, only has two factor loadings of 0.762 and 0.862. Similarly, the third component in the male subsample only has two factor loadings above 0.600. Hence, this analysis will only report on the first three factors, considering the first two factors significant and only noting the non-significant trends in the third factor.

Table 3.

Pattern matrix for the first three components for males and females

	Male components			Female components
	1	2	3	1	2	3
L1BodySL	0.808*	−0.128	0.015	0.722*	0.203	−0.059
L1BodySR	0.774*	−0.028	0.119	0.759*	0.016	−0.039
L1BodyIL	0.812*	−0.052	0.198	0.710*	−0.109	−0.164
L1BodyIR	0.773*	0.060	0.018	0.790*	0.011	−0.005
L2BodySL	0.857*	0.212	−0.125	0.734*	−0.232	0.055
L2BodySR	0.735*	0.118	0.177	0.697*	0.181	0.111
L2BodyIL	0.774*	−0.206	0.099	0.618	0.034	−0.174
L2BodyIR	0.830*	−0.010	0.057	0.823*	0.134	−0.160
L3BodySL	0.926*	0.167	−0.108	0.551	−0.124	0.210
L3BodySR	0.727*	−0.036	−0.159	0.686*	0.075	−0.026
L3BodyIL	0.775*	−0.087	0.136	0.764*	0.020	−0.114
L3BodyIR	0.833*	−0.188	0.170	0.796*	−0.114	−0.192
L4BodySL	0.872*	0.022	0.140	0.737*	−0.105	−0.018
L4BodySR	0.778*	0.028	0.153	0.644*	−0.087	−0.184
L4BodyIL	0.783*	−0.111	−0.020	0.645*	−0.362	0.128
L4BodyIR	0.824*	−0.005	−0.033	0.482	−0.262	0.107
L5BodySL	0.779*	0.139	−0.203	0.415	0.006	0.049
L5BodySR	0.664*	0.205	−0.259	0.496	0.235	0.106
L5BodyIL	0.779*	0.074	0.075	0.337	0.064	0.128
L5BodyIR	0.692*	−0.054	0.264	0.531	0.136	0.017
L1LamSL	−0.114	0.538	0.084	−0.107	0.711*	0.072
L1LamSR	−0.221	0.596	0.140	0.066	0.638*	0.194
L1LamIL	−0.130	0.389	0.393	−0.105	0.462	0.460
L1LamIR	−0.112	0.439	0.441	−0.173	0.504	0.277
L2LamSL	−0.058	0.588	−0.003	−0.014	0.719*	−0.151
L2LamSR	0.037	0.726*	−0.158	−0.043	0.662*	−0.397
L2LamIL	0.237	0.226	0.537	0.115	0.486	−0.190
L2LamIR	0.100	0.111	0.413	0.012	0.303	0.081
L3LamSL	−0.007	0.858*	0.059	−0.073	0.583	−0.016
L3LamSR	0.038	0.837*	−0.063	−0.117	0.551	−0.384
L3LamIL	−0.025	0.098	0.347	0.086	0.519	0.029
L3LamIR	0.134	0.036	0.619	−0.068	0.227	0.237
L4LamSL	0.056	0.436	0.031	0.058	0.593	0.107
L4LamSR	0.187	0.569	0.039	−0.191	0.511	−0.079
L4LamIL	−0.135	0.301	0.339	0.052	0.324	−0.193
L4LamIR	0.051	0.115	0.107	0.082	0.175	−0.249
L5LamSL	0.085	0.414	0.028	0.072	0.454	0.272
L5LamSR	0.195	0.623	0.000	0.200	0.415	0.057
L5LamIL	0.231	−0.027	0.553	−0.156	0.081	0.862*
L5LamIR	0.121	−0.452	0.638	0.285	0.084	0.762*
% of total variance	34.926	14.701	6.479	23.939	14.564	7.476

^*(bolded text) Denotes a significant factor loading at the 0.05 level (two-tailed).

Abbreviations follow the format vertebral level, vertebral body or lamina designation, superior or inferior surface, and left or right side. For example L1BodySL is the superior left vertebral body margin on the first lumbar vertebra.

The first three components explain 56.1% of the total variance in males and 46.0% of the total variance in females (Table 3). A few factor trends were detected for both males and females. First and foremost, the first components only contain vertebral body margins, whereas the second and third components contain laminar margins. Next, the second components only consist of superior margins, whereas the third components are only made up of inferior margins.

Among males, the factor analysis separates the variables according to three different components (Table 3). The first component contains all 20 vertebral body margins. The second component consists of the second and third superior right laminar margins as well as the third superior left laminar margin. Finally, the third component does not contain any statistically significant factor loadings.

Among females, the factor analysis again separates the variables according to three different components (Table 3). The first component contains all of the first vertebral body margins, most of the second, third and fourth vertebral body margins, and none of the fifth vertebral body margins. The second component consists of the first and second superior left laminar margins as well as the first and second superior right laminar margins. Finally, the third component is composed of the fifth inferior left and right laminar margins.

Regarding the Pearson's correlations, age was positively correlated with the first component (vertebral body margins), whereas BMI was not significantly correlated to any of the components for the male subsample (Table 4). For the female subsample, age was positively correlated with the first component (vertebral body margins) and BMI was positively correlated with both the first and second components (vertebral body margins and superior laminar margins).

Table 4.

Pearson's correlations

	Males		Females
	Age	BMI	Age	BMI
Vertebral bodies	0.855 P < 0.001*	−0.066 P = 0.748	0.589 P < 0.001*	0.442 P = 0.015*
Superior laminar margins higher in the column	−0.226 P = 0.192	0.246 P = 0.227	−0.199 P = 0.267	0.440 P = 0.015*
Inferior laminar margins lower in the column	0.100 P = 0.568	0.025 P = 0.902	−0.083 P = 0.645	0.080 P = 0.673

^*Denotes a significant result at the 0.025 level (two-tailed).

Discussion

The results show that osteophyte development on each vertebral body margin progresses with a high degree of coordination with other vertebral body margins at the different lumbar levels in males and females (vertebral body factor). In the laminae, fewer interdependent relationships exist amongst the laminar regions, but a differentiation between superior margins and inferior margins is noteworthy in both males and females (first and second laminar factors). Finally, age is related to osteophyte development throughout the vertebral body margins in males and females, whereas BMI is associated with vertebral body and superior laminar margin osteophyte development exclusively in females.

Both vertebral body and laminar osteophytes have previously been reported to exhibit patterned incidence. The majority of studies examining these two degenerative processes have focused on where in the column the greatest osteophytic development occurs (Culver & Pirson, 1956; Allbrook, 1957; Stewart, 1958; Nathan, 1962; Eisenstein, 1977; Bridges, 1994; Knüsel et al., 1997; O'Neill et al., 1999; Van der Merwe et al., 2006; Weiss & Jurmain, 2007). The current project aimed to look beyond patterns of more severe osteophyte development and determine: (i) whether patterns of association exist between the osteophyte development at the different levels, left and right sides, and superior and inferior surfaces of the laminae and vertebral bodies individually; (ii) if these patterns exist, whether a relationship exists between the two regions of osteophytic growth; and (iii) whether and how sex, age and BMI are related to osteophyte development.

The results indicate that the vertebral bodies and laminae demonstrate patterns of correlated osteophyte development within the lumbar levels and the superior and inferior margins. The factor analysis provides evidence for extensive correlation of vertebral body margin osteophyte development in L1–L5 in males and L1–L4 in females. The exclusion of L5 in females is interesting and not readily explainable. Additionally, the factor analysis evinces that males and females exhibit different grouping of correlated laminar margins. In females, the first and second vertebral superior laminar margins make up one correlated group. Another grouping, which can only be considered a trend, contains the inferior laminar margins of the fifth vertebra. In males, the second and third superior right margins as well as the third superior left margin make up a correlated group of laminar margins. The groups of correlated laminar margins in males and females do not readily point towards any specific cited factors, but the differences seen between males and females are evidence that different biomechanical factors are working in the two sexes. The clustering of superior margins with other superior margins and inferior margins with other inferior margins is also noteworthy in both males and females. Previous studies found a higher prevalence of osteophytes on the cranial surfaces than on the caudal surfaces of vertebral bodies as well as in areas of greatest stress to the ligamentum flavum on the laminae (Shore, 1931; Nathan, 1962; Gloobe & Nathan, 1973). This suggests that the ligamentum flavum attachment site may be weaker on the superior surface.

In both male and female groups, the vertebral body factor significantly correlates with age. These findings are in agreement with previous age-related osteophyte studies, which found more severe vertebral body osteophyte development in older individuals (Stewart, 1958; Nathan, 1962; Snodgrass, 2004). Additionally in females, the vertebral body factor as well as the first laminar component are significantly correlated with BMI. The correlation between osteophytes of the vertebral body and BMI provides evidence for the hypothesis that vertebral body osteophytes are an adaptive response to increased loading. Because females generally have vertebral bodies that are narrower than those of males of the same weight, the smaller cross-sectional area of their vertebral bodies is stressed more than with a comparable load in males (Davis, 1961; Taylor & Twomey, 1984; Gilsanz et al., 1994). Additionally, other researchers have found a significant association between body weight and the prevalence and severity of osteophyte development throughout the body (Kellgren & Lawrence, 1958; Liu et al., 1997). Furthermore females of European ancestry have been noted to exhibit significantly less lumbar curvature than males of similar ethnic origins (Trotter, 1929). When a relatively straight lumbar column is loaded, it has been shown to straighten even more (Trotter, 1929). This different loading alignment may force the ligamentum flavum in females to compensate for some of the stresses normally resisted by the anterior longitudinal ligament. Greater stress upon the ligamentum flavum would very likely impact osteophyte development on the laminae. For both of these cases, BMI necessarily impacts loading and therefore osteophyte development in females more than in males. This case may be especially true with this sample due to a non-significant difference in BMI between males and females. BMI is not significantly correlated with any of the components in males. The lack of correlation is most likely due to the greater width and therefore cross-sectional area of the vertebral bodies and the more curved lumbar column, which better distributes the load on the vertebral bodies and reduces the stresses upon the ligamentum flavum as discussed above. However, the association between osteophyte development and lumbar lordosis needs to be explored further. While Trotter (1929) observed males to exhibit more lumbar curvature than females, other studies have found the reverse to be true (Gelb et al., 1995; Vialle et al., 2005; Whitcome et al., 2007) and still others have found no significant differences between the sexes (Jackson et al., 2000; Giglio & Volpon, 2007).

Conclusions

This study demonstrates clear patterns within lumbar vertebral body and laminar osteophytes in the male and female samples. In general, the vertebral body margins are extensively correlated to each other (vertebral body factor) and age in both males and females. The laminar margins, evincing less widespread correlations, are related according to superior and inferior surfaces (first and second laminar factors). This study does not provide evidence for correlated osteophyte development occurring between the vertebral bodies and laminae.

The different trends found in males and females support previous findings that degenerative responses in males and females are different (Allbrook, 1957; Kellgren & Lawrence, 1958; Nathan, 1962; Liu et al., 1997). Although widespread vertebral body margin correlations exist for both sexes, only osteophyte formation at the first four vertebral levels exhibits interdependent development in females, while coordinated development occurs at all five levels in males. Additionally, different groups of laminar margins are correlated with each other in males and females. The independent osteophyte development of the vertebral body margins of the fifth lumbar level in females and the different patterns of interdependent laminar osteophyte development seen in males and females represent important biological differences between the sexes in vertebral morphology, weight transmission and degenerative responses. Finally, the relationship between BMI and certain laminar regions in females is unique when compared with males. This finding provides important groundwork for the future exploration of the effects of weight and stature on osteophyte development. As obesity is becoming a more widespread problem throughout the world, a more in-depth study of the effect of BMI on lumbar spine degeneration is warranted. Overall, further exploration of the differences between male and female osteophyte development is needed to facilitate better sex-specific treatment for the various pathologies of the spine and therefore enable individuals of both sexes to stay more mobile.