Spondylolytic Spondylolisthesis

SPINE Volume 30, Number 6S, pp S12–S21 ©2005, Lippincott Williams & Wilkins, Inc. Spondylolytic Spondylolisthesis A Study of Pelvic and Lumbosacral Parameters of Possible Etiologic Effect in Two Genetically and Geographically Distinct Groups With High Occurrence Thomas E. Whitesides, Jr, MD,* William C. Horton, MD,* William C. Hutton, DSc,*† and Lisa Hodges, BA, BSN, MSA‡ Study Design. An anatomic and radiographic study of archeological skeletal remains from two genetically and geographically distinct groups with high occurrence rates of spondylolytic spondylolisthesis was done. Specimens were Aleut (27% known occurrence rate, n � 48) and Arikara Plains Indians (9% occurrence, n� 250� of 1,062). Objective. To evaluate three radiographic parameters highly correlated with spondylolisthesis (pelvic incidence [PI], sacral table angle [STA], and lumbar index [LI]) in genetically homogeneous populations to determine which may be etiologic or most predictive for lysis. Summary of Background Data. LI has been known to vary with the percentage of slip in lytic spondylolisthesis. Recent clinical studies have shown that PI is also signifi- cantly higher in high-grade slips, and a possible etiologic effect has been ascribed to this association. STA has also been shown to vary between normals, those with only lysis, and those with lysis and slip. The etiologic signifi- cance of STA is unknown. Methods. Radiographic and direct morphologic mea- surement of PI, LI, and STA was done on L5 and reassem- bled sacra and ilia. Statistical analysis of these three pa- rameters among all groups was done. Results. 1) There is a genetically determined difference in the upper sacral tilt (STA) that may be etiologic. 2) Genetically homogeneous groups with a lower STA in normal specimens have an increased occurrence rate of spondylolysis. 3) When there has been pars lysis, changes in the STA occur as well as deformity more caudal in the sacrum. 4) These changes are likely related to remodeling with epiphyseal growth related to changed axial stresses secondary to pars lysis. 5) PI is not a pri- mary etiologic factor in the process. Conclusions. The STA in the normal population for each genetic group varies and relates significantly to the occurrence rate and is thus probably etiologic. STA is more highly associated with the occurrence of pars defect than is PI. Upper sacral deformities appear due to the growth plate response to the changed pressure gradients across the epiphyseal plate rather than interosseous re- modeling of the ilium and acetabular area. Thus, changes in PI would be secondary. Key words: spondylolisthesis, spondylolysis, occur- rence rate, etiology, lumbar index, pelvic incidence, sacral table angle. Spine 2005;30:S12–S21 There have been a variety of studies that have identified radiographic and morphologic parameters in regard to spondylolytic spondylolisthesis. These have primarily been from observations on plain radiographs of sagittal deformity and include Meyerding grade of slip, lumbar index (LI), sacral inclination, sacral angle from the hor- izontal, sacral tilt, slip angle, lumbar lordosis, proximal compensatory curves, pelvic radius, and center of gravi- ty.1–8 Most of these features are descriptive, and little is known about the etiology or prognostic significance of them. Cyron and Hutton9–12 carried out studies of a more biomechanical nature, elucidating some of the mechanisms of pars failure. Hutton et al suggested that spondylolysis is a fatigue failure and that the resulting pars fracture causes a change in the mechanics of the motion segment such that the cascade of spondylolysis and arch separation are set into motion.10 However, no studies have shown statistical relevance for any radio- graphic parameters that may be etiologic, and most stud- ies have not involved strictly defined genetic groups. In the last decade, there has been increasing interest and research in sagittal balance and standing posture, especially as it relates to scoliosis. This has been promul- gated by several authors.2,5,6,13–16 The latter have devel- oped the concept of pelvic incidence (PI) (Figure 1). In a study of spondylolisthesis, they found statistical correla- tion in two areas: sacral slope and PI. The PI increased from 41.8° in the adolescent normal to 64.5° in the adult with isthmic spondylolisthesis. PI has also been found to correlate with the grade of slip by these and other au- thors.13,17–19 Duval-Beaupere’s group stated that PI was etiologic in the occurrence of isthmic spondylolisthesis, suggesting that a larger PI and sacral slope could predis- pose to vertebral slip.16 This would require that the in- creased PI be present prior to lysis or vertebral slip. This observation has not yet been reported. Spondylolisthesis is not present in the newborn ac- cording to a longitudinal study by Fredrickson et al using From the *Department of Orthopaedics, Emory University School of Medicine, Emory Spine Center, Atlanta, GA; †Veterans Affairs Medi- cal Center, Decatur, GA; and ‡National Institutes of Health, Washing- ton, DC. Acknowledgment date: December 13, 2003. Acceptance date: Decem- ber 21, 2004. The manuscript submitted does not contain information about medical device(s)/drug(s). No funds were received in support of this work. No benefits in any form have been or will be received from a commercial party related directly or indirectly to the subject of this manuscript. Levels of Evidence: Prognostic Studies, Level II. Address correspondence and reprint requests to Thomas E. Whitesides, Jr, MD, 958 Calvert Lane NE, Atlanta, GA 30319-1202; E-mail: twhitesides@comcast.net S12 a population from Pennsylvania.3 No defects were found in 500 consecutive radiographs from newborn infants. In another group of 500 children from the same area fol- lowed from the age of 6 years, the occurrence rate at age 6 years was 4.4%. Racial or genetic characteristics were not reported or analyzed. The occurrence rate that they reported increased slightly until age 18 years when it had become 5.4%. Of the 27 subjects who at age 18 had developed bilateral lytic defects, 20 (74%) had devel- oped listhesis. They observed that the younger in age the defect occurred, the greater was the amount of eventual slippage, and that no slip occurred after age 18 years. The LI correlated statistically with the Meyerding grade/ percentage of slip. No spondyloptosis occurred. They did not measure PI or sacral table angle (STA) as these parameters had not yet been described, nor did they re- port the genetic composition of their population groups. The occurrence of lysis and slip in genetically different genetic groups has been reported to vary from less than 3% in the African20–22 to 45% to 50% in various iso- lated Northern Inuit groups.21 This suggests that studies of genetically homogeneous populations with reported high or low occurrence might be illuminating in identi- fying etiologic or prognostic features. Inoue et al4 recently reported a study on a racially distinct Japanese population in which they identified three significant parameters regarding the presence or absence of lytic defects and listhesis (Figure 1): LI, sacral table index (STI), and STA. STA was initially described by Osterman and Osterman in rabbit lumbosacral exper- iments and termed “sacral endplate angle.”23 The STA is that angle in the midline sagittal plane subtended by a line drawn from the anterior aspect of the superior end- plate of S1 to its posterior edge and a line drawn from that point caudally to the posterior midline point of the caudal endplate of S1 (as noted in Figure 1). Inoue’s study showed a statistically significant and progressive decrease in STA, STI, and LI when lysis and the subse- quent listhesis were present (Table 1). The correlation between STA, STI, and LI with slippage was statistically strongly supported.4 Inoue et al4 made no suggestion that any of these parameters is etiologic for the lysis that allows listhesis to occur. Thus, it is not clear if the changes in LI, STA, and STI are primary (causative) or secondary to the slip (remodeling). Unfortunately, Inoue et al4also did not report on PI. The occurrence rate of spondylolisthesis in the Japanese is reported as 5.6%.24 From the longitudinal study of Frederickson et al,3 spondylolysis antecedes spondylolisthesis. An unsolved question is whether the reported increase in PI or de- crease in STA or LI was present before the onset of spon- dylolysis (and are thus etiologic), or whether these mea- surements are changed by virtue of remodeling of the upper sacrum and/or the L5 vertebra. Also, can one or more of these or other parameters explain the wide oc- currence rate differences between genetic groups that are known to be present? It was thus thought that a study of archeological specimens from genetically isolated groups having high occurrence of spondylolysis might be helpful in further evaluating the significance of these parameters. The purpose of this study was therefore to evaluate three radiographic parameters highly correlated with spon- dylolisthesis (PI, STA, and LI) in these genetically differ- ent populations to determine what correlations may be etiologic or predictive. Materials and Methods Permission was obtained to carry out this study at the National Museum of Natural History of the Smithsonian Institution in Washington, DC using collections archeologically excavated in the mid 20th century primarily by the late Ales Hrdlicka, MD, and T. Dale Stewart, MD, under the auspices of the Smithso- nian Institution.25,26 Three collections with high incidence of pars defects were present from isolated populations and repre- sented material that predated the introduction of progenitive genetic material of heterogenous origin: 1. North Slope Inuit from the Point Hope area (lysis occur- rence of 45%–50%24) 2. Aleuts from the mid-Aleutian Four Islands area (Kagamil) (lysis occurrence of 27%) 3. Arikara Plains Indians from South Dakota (lysis occur- rence of 9%) Inuit Specimens. At the time of this study, consultation for the repatriation of these remains was either in progress or im- minent. The federal legislation allows for Native American tribes to request the return of biologically or culturally associ- ated remains. The Point Hope Inuit group, with an occurrence rate of 45% to 50% and been finalized for repatriation, and was thus unavailable. The other two groups (Aleut and Figure 1. Methods of determining PI, STA, STI, and LI.13,18 STA is that angle in the midline sagittal plane subtended by a line drawn from the anterior aspect of the superior endplate of S1 to its posterior edge and a line drawn from that point caudally to the posterior midline point of the caudal endplate of S1. Table 1. Aleut Results Normal (n � 19) Bilateral Pars Defects (n � 9) Pelvic incidence 51.5� 10.77 50.2� 9.00 Sacral table angle 94.9� 4.47 90.2� 3.56 (P � 0.01) Lumbar index 0.85� 0.04 0.82� 0.09 While there is no significant difference between normal and those with bilat- eral pars defects in PI, there is a significant difference in STA (P � 0.01). A trend is apparent in LI. Note: All data are from radiographs. S13Possible Etiologic Effects • Whitesides et al Arikara) were available, but with limited access, due to the evaluation process in progress for their repatriation. Aleut Specimens. We initially studied the semimummified remains from a mid-Aleutian Island area that had been isolated for millennia. The burial sites on Kagamil Island existed before the time of arrival of Russians and had not been added to since. We were able to reassemble both immature (Figure 2) and skeletally mature (Figure 3) specimens. Some were not useful for study due to animal predation or taphonomic loss of essen- tial bones. Specimens were reassembled using masking tape to recon- nect the sacroiliac (SI) articulations and, if both pubic areas were present, the pubic articulation. Any specimen without firm and distinct SI interlock was not measured. After an ap- propriate titanium femoral head sizing shell had been inserted and taped into each acetabulum to exactly locate it for imaging, they were radiographed using high-resolution mammography technique maintaining a true lateral projection (Figure 3). When both iliac wings were present, the radiograph beam was centered at the center of the cephalad S1 endplate. When only one iliac wing was present, the radiograph beam was centered at the anatomic center of the acetabulum. This conforms to the bifemoral axis. Comparison between the two projection meth- ods showed a mean standard error of 1.1°. The entire recon- structable group of Aleuts was radiographed, and all measure- ments were taken from direct tracings of the radiographs (Figure 2). In the Aleut group, of the 48 adult specimens present, the percentage of L5 bilateral lytic pars defects was 27%. Only one unilateral lysis was present. Three child specimens without de- fects were present. Arikara Specimens. A time period to study the much larger collection of Arikara Indians from South Dakota was granted at a later date. This collection came from settlements dating from the 1650s situated along the Missouri River region in northern South Dakota, these sites being abandoned just before or just after European contact. Thus, European genetic admix- ture is improbable. The collection size of 1,062 specimens was too large to study in the allotted time; thus, a random sample was reviewed. Examination continued until large enough num- bers in each group had been examined to permit appropriate statistical analysis. Of the 250� reviewed, 140 specimens were Figure 2. An intact mummified lumbar spine and sacrum to which the iliac wings have been reattached with firm SI joint in- terlock. The radiographic image of this specimen demonstrates the use of an overlay of tracing paper on which to draw and measure parameters such as PI, STA, and LI. Figure 3. Radiographs were made with mammography film technique in a true lateral pro- jection after the iliac wings have been reattached at the sacrum and pubis with titanium acetabu- lar sizing shells inserted. The protractor adaptation and its use for direct measurement of STA are depicted. Comparison be- tween this direct method and the radiographic method shown in Figures 2 and 3 produced a mean standard error of � 1.1°. S14 Spine • Volume 30 • Number 6S • 2005 found usable for study. For expedience due to the time con- straints involved, a direct method for measuring STA was de- veloped using a protractor adaptation (Figure 3) and was used on all Arikara specimens. Forty-six of the 140 specimens were also radiographed to obtain PI data and to corroborate the direct protractor method of STA measurement with the radio- graphic measurement (mean standard error, �1.1°). The collection was cataloged in the order that they were excavated from the archeological site without separation by sex, age, pathologic condition, or in any other manner. The collection was not curated with pathologic conditions segre- gated into different storage locations, so overselection for uni- lateral or bilateral lytic pars defects would be impossible. Skel- etal analysis during cataloging identified an overall combined occurrence of unilateral and bilateral defects of 10%; however, the rate of bilateral defects was not recorded. We were given the catalog numbers of the specimens identified by the curators with either unilateral or bilateral lysis. These specimens were then, in addition to the original random group, consecutively examined from the collection of 1,062 until a total of 70 had been examined. Seven of this 70 were found to have a unilateral defect, giving an estimated 9% occurrence rate of bilateral de- fects for the Arikara. As in the Aleut, no dome-shaped sacral table surfaces were found. All had a flat surface in the midsagittal plane extending from the posterior aspect of the sacral table to within 5 mm of the anterior lip. This was expected as both previous clinical studies of the Inuit, Aleut, and Athabasca populations of Alaska had found no spondylolisthesis above a Meyerding Grade II.27,28 The Arikara are of genetically similar origin to the Athabasca. Not all normal L5 vertebrae were radiographed in the Aleut and Arikara groups. It was impossible to temporarily accu- rately position the L5 vertebrae with separate arches to the sacrum. This caused radiographic magnification differences that made calculation of the STI inherently difficult. Accurate recreation of the erect posture was also not possible. Thus, the Meyerding Grade and/or percent of slip and other parameters dependent on the erect posture could not be determined. These factors limited the parameters for which one could obtain ac- curacy to PI, STA, and LI, and the number of observations of LI was small. Statistics. Means and standard deviations were computed to compare the average measurements of STA, PI, and LI across specimens. To test whether there was a statistically significant difference between abnormal and normal specimens indepen- dent two-sample pooled t tests were used for the Aleut. A one- way analysis of variance followed by Fisher’s LDS pair-wise comparisons was performed for the Arikara as there were three test groups (normal, unilateral lysis, and bilateral lysis). Results are reported as statistically significant if the P value was less than or equal to 0.05. In addition, the Pearson correlation coefficient was calculated to summarize the association be- tween STA and PI for Aleuts and Arikara separately. Results The data for the Aleut and Arikara are shown in Tables 2 and 3. Statistical analysis of the Aleut data failed to show any significant difference between the normal and the lytic abnormals in PI. However, the STA difference was significant (P � 0.01) with the STA being lower in the cases with lysis. The LI was also diminished with lysis, but the trend observed did not reach statistical sig- nificance, most likely because of low numbers (Tables 2 and 3). As only one unilateral lysis was present in this small group, no statistical evaluation of this was possi- ble. Recent clinical studies corroborate our documented percentage of occurrence in the Aleut and document no progression beyond Meyerding Grade II.25,26 The Arikara statistical analysis showed progressively significant differences in PI and STA between those that were normal, had a unilateral defect, and had bilateral defects (Table 4). The STA is lower in the bilateral lysis group (88.0° � 4.2°) compared with the normal Table 3. Inoue Japanese Data4 Normal (n � 310) Bilateral Defect No Slip (n � 213) Bilateral Defect With Slip (n � 154) Comparison Normal versus Bilateral With Slip (P ) Sacral table angle (STA) 97.4� 4.2 97.8� 4.3 91.6� 5.7 �0.0001 Sacral table index (STI) 94.4� 3.8% 96.6� 3.4% 102.5� 5.8% �0.0001 Lumbar index (LI) 0.890� 0.052 0.826� 0.05 0.803� 0.062 �0.001 These data show a significant decrease in STA, STI, and LI from normal to bilateral pars lysis without slip and then an especially significant difference between normal and those with spondylolisthesis.4 Table 2. Adult Arikara Data Normal (n � 83) Unilateral Defect (n � 7) Bilateral Defects (n � 23) Pelvic incidence 47.1� 4.4 57.3� 5.0 61.7� 4.2 (P � 0.04) (P � 0.002) Sacral table angle 98.1� 4.6 94.9� 5.0 88.0� 4.2 (P � 0.07) (P � 0.001) Lumbar index 0.84 0.77 0.79 STA is significantly lower in those with lysis (P � 0.001) and PI is significantly lower also in those with lysis (P � 0.002). A trend is present for LI that did not reach significance due to low numbers. Note: Data are from radiographs and direct measurement. Table 4. Group Occurrence Rate Versus STA Genetic Group Occurrence Rate of Bilateral Defects (%) STA (normal adult mean) (°) Aleut 27 94.9� 4.5 Arikara 9 98.1� 4.4* Japanese 5.6 97.2� 4.2† The rate o