Lumbosacral Transitional Disorder as a Missing Link in Symptomatic Scoliosis

Abstract

Back pain in the presence of LSTVs (lumbosacral transitional vertebrae) was originally noted by Mario Bertolotti in 1917. The Lenke classification for scoliosis forms the current international standard. However, the connection between LSTVs and scoliosis is still poorly understood. The aim of this study is to show the connection between scoliosis and LSTVs in terms of their frequency and impact on further development. Our scoliosis outpatient clinic has an examination period that covers the years from 2014 to 2021. If LSTVs are radiologically suspected (AP and lateral X-ray of the spine) according to the Castellvi classification (II–IV), a further MRI investigation is initiated. Scoliosis shape is assigned according to the Lenke classification. Sagittal segmental differences are observed according to the vertical mean vertebral angle (VMVA) and its difference (Diff-VMVA). Assignment to the lumbar pain group is made in cases of a history of chronic or recurrent pain of more than 6 months. Differentiation is made at the initial presentation according to gender (male–female) and age (children and adults). Other associated diagnoses such as family correlations and the question of brace fitting are cited. Finally, the literature is reviewed from a scoliosis perspective and compared with the findings of the authors of this paper. A total of 1332 patients were evaluated, and LSTV Castellvi II–IV was confirmed in 72 of them (58 female and 14 male). The curvature extent of scoliosis in children and adults had a mean Cobb angle of 24.3° with a range from 11° to 55° (n-42) and 32.4° with a range from 12° to 66° (n-30), respectively. This indicates that 75% of n-54 patients were mostly classified as Castellvi II (pseudarthrosis) (IIA, 54.2%; IIB, 20.8%) in the total data. There were few patients classified with Castellvi III (22.2%) and IV (2.8%). A proportion of 87.5% of 72 patients were mostly classified as Lenke 1 (25.0% n-18) and Lenke 5 (62.5% n-45). According to the literature, the Diff-VMVA shows Cobb angles of 9.3° for Castellvi III and 5.3° for Castellvi IV. Half of the patients complain of chronic low back pain for at least 6 months or have recurrent complaints (48.6%). The complaints are mostly classified as Castellvi IIA (27.8%) and B (9.7%). The association of scoliosis and LSTVs yields only 13 studies (PubMed 04/18/2022) that focus on spine surgery issues and not on diagnostics. Attention should be given to the lumbosacral junction in Lenke 1 and Lenke 5. Nearly half of the patients in each category complain of chronic lumbar symptoms. This is notable in individual cases due to its rounding and thus its increased Diff-VMVA. Since only Castellvi II–IV is considered, a comparison of the frequency with other studies is not permissible. In conclusion, for scoliosis Lenke 5 and Lenke 1, the lumbosacral junction should be examined.

1. Introduction

In 1917, Bertolotti stated that LSTVs (lumbosacral transitional vertebrae) can cause low back pain [1]. Castellvi et al. classified LSTVs into four types. Type I includes unilateral (IA) or bilateral (IB) dysplastic transverse processes with a width of at least 19 mm (craniocaudal dimension). Type II shows incomplete unilateral (IIA) or bilateral (IIB) lumbarization/sacralization with an enlarged transverse process that forms a pseudarthrotic joint between itself and the sacrum. Type III describes unilateral (IIIA) or bilateral (IIIB) lumbarization/sacralization with complete bony fusion of the transverse process(es) to the sacrum. Type IV involves a unilateral transition from type II to type III on the contralateral side [2], see Figure 1.

The Ferguson image (AP radiograph with a 30° angle) serves as the standard method for detecting LSTVs. However, this radiograph is not the same as that used to clarify scoliosis. Thus, an additional radiograph becomes necessary. The alternative is an MRI scan, which can also provide additional information for all three spatial planes. Diagnostic errors in determining vertebral body assignment on sagittal radiography and lumbar MRI are also common [3,4,5], see Figure 2.

Therefore, routinely obtaining sagittal MRI images of the entire spine is recommended. The caudal numbering of C2 on MRI images of the entire spine and the use of Castellvi’s morphologic classification allow for the correct identification of both numeric and morphologic vertebral variations [6,7,8].

The use of MRI has demonstrated a sensitivity of up to 80% in identifying LSTVs. However, sagittal and axial images alone have been shown to increase the likelihood of misdiagnosis [9]. CT scans provide the highest accuracy in assessing bony anatomy and may be useful in considering LSTVs.

Farshadet et al. advocate for measuring segmental differences in the vertical mid-vertebral angle (Diff-VMVA) and vertical anterior vertebral angle (Diff-VAVA) of the three most caudally located segments of the lumbar spine to easily and reliably identify LSTVs. A Diff-VMVA of ≤+10° identified type III and IV LSTVs with a sensitivity of 100% and a specificity of 89% on MRI [10,11].

Chalain et al. [11] state that LSTVs may show increased lordosis of the lumbar spine and a lack of angulation at the lumbosacral junction. Measurement of angles allows for the prediction of the presence of LSTVs.

Other methods of measurement in the sagittal slices of MRI scans are also described for the identification of an LSTV. In this regard, the variability in the vertebral bodies makes it difficult to define L5 and S1 accurately [12]. As indicated, the number of vertebral bodies is not an assured constant.

A high prevalence of variations in the number of thoracic or lumbar vertebrae is also described for scoliosis patients. However, this is mainly crucial in fusion surgery or pseudoarthrectomy [13,14].

According to Quinlan et al., the prevalence in the studied MRI scans of 769 consecutive patients of Bertolotti syndrome is 4.6% in the general population and 11.4% in those with back pain who are younger than 30 [15]. Unilateral LSTVs result in asymmetric biomechanical loading. The side with the extra L5/S1 junction carries a higher proportion of the load, resulting in lateral tilting with the scoliotic curvature. This local loading leads to higher wear and unilateral muscle activity. In addition, asymmetric motion may also influence disc degeneration [16].

The Lenke classification for scoliosis forms the current international standard [17] and focuses on scoliosis surgery. However, there is no generally accepted uniform classification for brace fitting, so various classifications are used.

The Lenke classification is applied in the study by Choon Sung Lee et al. [18]. The prevalence of LSTVs in patients with adolescent idiopathic scoliosis (AIS) is 12.2%. A trend toward early degeneration of the L4-5-level discs is noted in patients with AIS with LSTVs. This article considers the issue from the perspective of a necessary fusion surgery [18].

In the association between scoliosis and LSTVs, the literature focuses on the local loading phenomena and on the degeneration of the adjacent vertebral segment.

Another study by Farshad-Amacker et al. concludes that strengthening the mechanical connection of a lumbosacral transitional vertebra protects the disc at the transitional level. The adjacent cranial segment shows a comparable degree of degeneration to the L5-S1 level in control subjects [19].

A 2015 review by Jancuska et al. outlines the potential pathogenesis and local pathologic burden in LSTVs. The interaction of the HOX genes, responsible for spinal segmentation and also for the effect on the development of the lumbosacral junction, is therefore explained biomechanically [20].

Estimates of the prevalence of LSTVs in the general population vary widely in the literature due to differences in definition and diagnostic modalities, ranging from 4.0% to 35.9%. Apazidis et al. found that type IA is the most common, with a prevalence of 14.7%. However, type I is generally considered to be clinically insignificant [21].

In the work of Min Tang et al., the prevalence of LSTV type 1 diagnoses is the most common. The representational inaccuracy of the processus transversus in its dependence on pelvic tilt should not be neglected [22].

Konin et al. comprehensively address the identification and correct numbering of LSTVs, and they recognize imaging findings related to the development of back pain [23].

The aim of this study is to show the connection between scoliosis and LSTVs in terms of its frequency and impact on further development.

2. Methods

In 2014, a patient presented with chronic low back pain 37 years after undergoing scoliosis surgery. During the CT scan, LSTVs were diagnosed as the cause of the symptoms. This case became the starting point for a targeted evaluation of the lumbosacral junction in all scoliosis patients (Cobb angle > 10°). Since then, if LSTVs are radiologically suspected (with an AP and lateral X-ray of the spine showing a bony bridge or a rounded promontory), further investigation is initiated, and this became the inclusion criteria for this study. Exclusion criteria are a Cobb angle of < 10° and Spina bifida. The examination period covers the years from 2014 to 2021.

Initial Data at Baseline

The input data are characterized by the fact that patients were generated from the scoliosis outpatient clinic. Thus, scoliosis forms the first criterion for data collection, and a comparison of prevalence with other studies is not justified. The decision to only include patients with LSTV Castellvi II–IV in this study leads to a further reduction in patient numbers. At the same time, however, diagnostic certainty is achieved, which is further enhanced by the free availability of MRI and CT.

The responsible MRI institute is instructed to provide the sagittal, axial, and coronary slices when LSTVs are questioned. Only in cases of a contraindication for MRI examination (patients with a long-distance spinal fusion or cochlear implant) was CT examination with 3D reconstruction ordered. All patient examinations were performed by a single orthopedic specialist.

LSTVs were assigned according to the Castellvi classification.

Patients with the Castellvi I classification (height of transverse process > 19 mm) were not included in this study because imaging inaccuracy has been shown to only allow for an exact measurement of the processus transversus in the Ferguson image. In addition, this additional radiograph is prohibited for radiation hygiene reasons.

Therefore, this study contained the data of patients with the Castellvi II–IV classification. In Castellvi II for pseudarthrosis and Castellvi III for ossification, a distinction is made between A (one side affected) and B (both sides affected). In Castellvi IV, one side is pseudarthrosis and one side is ossification. This distinction is guaranteed with high precision and was therefore chosen for this study.

Scoliosis shape is assigned according to the Lenke classification with curve type 1–6 and lumbar spine modifier A-B-C.

Sagittal segmental differences are made according to the vertical mean vertebral angle (VMVA) and its difference (Diff-VMVA) for the three most caudally located segments of the lumbar spine.

Assignment to the lumbar pain group is made in cases of a history of chronic or recurrent pain of more than 6 months.

Differentiation is made at the initial presentation according to gender (male–female) and age into the group of children (Risser 0–4 and <16 years of age) and adults (Risser 5 and >16 years of age).

Other associated diagnoses such as family correlations and the question of brace fitting are cited.

Finally, the literature is reviewed from a scoliosis perspective and compared with the findings of the authors of this paper.

3. Results

During 2014–2021, 1332 patients were evaluated for scoliosis in the scoliosis outpatient clinic, and further clarification of the lumbosacral junction was initiated in 116 of them (MRI n-111 and CT n-5).

LSTV Castellvi II–IV was confirmed in 72 patients, comprising 58 females and 14 males. Of these, 42 were identified prior to growth completion (Risser 0–4, age 8–16 years) and 30 were identified once growth completion had already been reached at baseline (Risser 5, age 16–76 years).

The curvature extent of scoliosis in children and adults had a mean Cobb angle of 24.3° with a range from 11° to 55° (n-42) and a mean of 32.4° with a range from 12° to 66° (n-30), respectively.

Two female patients in the adult group with unknown LSTVs up to that point had undergone posterior spinal fusion 30 and 37 years prior and presented with chronic lumbar pain.

Brace treatment was recorded in 10 of 42 children, whereas 7 of 30 adult patients had a history of such treatment.

Concerning the accompanying bony pathologies, vertebral body and rib malformations were found in children (n-6) as well as in adults (n-4). Intraspinal changes included syringomyelia (n-1), sacral cyst (n-1), and spina bifida (n-2). Other malformations such as congenital clubfoot, longitudinal radiation defect, and cartilaginous exostosis disease were present in the group of children.

When family history was obtained, a pair of twins with LSTVs and a mother–daughter relationship with LSTVs were found (

Table 1. Data at initial examination.

	Age Average	Cobb Angle Average	Cobb Angle Range	Scoliosis Treatment
Children Risser 0–4 n-42	13.3 years	24.3°	11°–55°	n-10 brace
Adult Risser 5 n-30	31.4 years	32.4°	12°–66°	n-7 (bracing at childhood)
				n-2 (dorsal spine fusion)
	Malformations
Risser 0–4 n-42	n-9 patient	vertebrae (n-4), rib malformation (n-2), syringomyelia (n-1), longitudinal radiation defect (n-1), congenital clubfoot (n-1), cartilaginous exostosis disease (n-1)
Risser 5 n-30	n-5 patient	vertebrae (n-2), rib malformation (n-2), sacral cyst (n-1), spina bifida (n-1)
Genetic	n-1 mother and daughter
	n-1 twins

).

3.1. LSTVs in Castellvi Classification

The scoliosis patients were mostly classified as Castellvi II (pseudarthrosis), with 75% of the n-54 patients (IIA, 54.2%; IIB, 20.8%) in the total data. In the present group of patients, the sex distribution showed a bias toward the female sex comparable to idiopathic scoliosis (19.4% male and 80.6% female). This distribution pattern was also maintained in the individual Castellvi classification (

Table 2. Castellvi classification and gender.

Castellvi Classification	Total Data		Male		Female
	n-72		19.4%	n-14	80.6%	n-58
IIA	54.2%	n-39	11.1%	n-8	43.1%	n-31
IIB	20.8%	n-15	5.6%	n-4	15.3%	n-11
IIIA	12.5%	n-9	1.4%	n-1	11.1%	n-8
IIIB	9.7%	n-7	1.4%	n-1	8.3%	n-6
IV	2.8%	n-2			2.8%	n-2

).

3.2. Scoliosis Severity

The Castellvi classification showed no correlation with curvature extent. This was widespread in all groups with similar mean Cobb angles between 25.4° and 39° (

Table 3. Castellvi classification and scoliosis.

Castellvi Classification		Extent of Curvature Cobb Angle	Average of Curvature Cobb Angle
IIA	n-39	12–46°	26.4°
IIB	n-15	21–42°	27.1°
IIIA	n-9	16–66°	35.0°
IIIB	n-7	12–60°	25.4°
IV	n-2	38–40°	39.0°

).

3.3. Curvature Shape in the Lenke Classification

Assignment of scoliosis was mostly classified as Lenke 1 (25.0% n-18) and Lenke 5 (62.5% n-45) in 87.5% of 72 patients (

Table 4. Lenke classification and low back pain.

Lenke			lumbar Spine Modifier			Chronic Low Back Pain
			A	B	C
1	25.0%	n-18	n-7	n-7	n-4	9.7%	n-7
2	4.2%	n-3
3	4.2%	n-3
4	1.4%	n-1
5	62.5%	n-45	n-10	n-13	n-22	32.9%	n-23
6	2.8%	n-2

).

Thus, attention should be given to the lumbosacral junction in Lenke 1 and Lenke 5. The lumbar spine modifier does not provide further evidence of LSTV differentiation. Nearly half of the patients in each category complain of chronic lumbar symptoms (Lenke 1 and 5 n-63, with chronic low back pain n-30) (Table 4).

3.4. VMVM Vertical Mid-Vertebral Angle and Diff-VMVA in the Lowest Vertebral Segments

According to the literature, the vertical mid-vertebral angle in the sagittal profile is 30.2° at S1-L5, and VMVA L5-L4 shows a value of 18.8°, giving a Diff-VMVA of 11.5°. The Castellvi II group shows a Diff-VMVA of 12.4° (n-54), and the Castellvi III group shows 9.3° (n-16). The Castellvi IV group shows a Diff-VMVA of 5.3°, which is significantly decreased, but only in a small number of patients (n-2). This is notable in individual cases due to its rounding and thus its increased Diff-VMVA. However, in the calculation of the mean values, no clear differentiation from the sagittal profile to the Castellvi classification can be concluded in the present data (Figure 3,

Table 5. Castellvi classification and vertical mid-vertebral angle.

	Vertical Mid-Vertebral Angle
	VMVA S1-L5	30.2°	n-72
	VMVA L5-L4	18.8°	n-72
	Diff-VMVA	11.5°	n-72
Castellvi Classification
IIA	Diff-VMVA	12.4°	n-54
IIB
IIIA	Diff-VMVA	9.3°	n-16
IIIB
IV	Diff-VMVA	5.5°	n-2

).

The finding by Farshad et al. that a Diff-VMVA of <10° indicates LSTV III and IV agrees with our data [10].

3.5. Low Back Pain

Half of the patients complain of chronic low back pain for at least 6 months or have recurrent complaints (48.6%). The complaints are mostly classified as Castellvi IIA (27.8%) and B (9.7%). The gender distribution of chronic low back pain (8.3% male and 40.3% female) is comparable to that in the symptom-free group (9.7% male and 41.7% female). Thus, a gender-specific difference cannot be ascertained (

Table 6. Castellvi classification and low back pain.

Castellvi Classification	Low Back Pain Total		Low Back Pain Male		Low Back Pain Female
	48.6%	n-35	8.3%	n-6	40.3%	n-29
IIA	27.8%	n-21
IIB	9.7%	n-7
IIIA	4.2%	n-3
IIIB	5.5%	n-4
IV	1.4%	n-1
	Symptom-free Total		Symptom-free Male		Symptom-free Female
	51.4%	n-37	9.7%	n-7	41.7%	n-30

).

3.6. The Literature

The association of scoliosis and LSTVs yields only 13 studies (PubMed 18 April 2022) that focus on spine surgery issues.

However, care should be taken to exclude Bertolotti syndrome in patients younger than 30 years with persistent low back pain due to its congenital origin, as suggested in a review. If LSTVs are identified, usually with an MRI or CT scan of the lumbosacral spine, and there is no additional compelling or obvious source for the patient’s symptoms, a conservative, stepwise treatment plan is recommended. There are limited postoperative outcome data for patients undergoing surgical treatment, with promising but variable results.

However, diagnostic imaging such as X-ray, CT, MRI, and scintigraphy, as well as diagnostic injections when needed, still result in suboptimal diagnostic accuracy, and superior diagnostic methods are yet to be found [24].

4. Discussion

Lumbal skolioses in combination with LSTVs are malformations and therefore are not susceptible to brace treatment.

An urgent question for future research is as follows: what are the limitations for conservative treatment? This concerns the pain treatment of the affected segment as well as the adjacent segment. Brace treatment for scoliosis must be differentiated from idiopathic scoliosis in terms of its expected effectiveness.

The question of the further clarification of the pathological significance of Castellvi type I should be discussed. The assumption that the height of the transverse process of 19 mm correlates with the pathology must be clarified. The presented work should be an impetus for further studies.

Attention should also be given to the concomitant diagnoses of the malformations, which indicate a developmental history of LSTVs.

The pair of twins and the mother–daughter relationship should be noted. Familial relationships in scoliosis are common knowledge and should be considered for LSTVs as well.

LSTVs in Castellvi classification:

A correlation must be assumed since the sex-specific distribution pattern of LSTVs corresponds to that in idiopathic scoliosis. Focusing on the Castellvi II group (pseudarthrosis) always requires a history and investigation of complaints.

Scoliosis severity as a function of Castellvi classification:

No distinction can be made between unilateral and bilateral pathological changes in LSTVs on the basis of the extent of scoliosis. Also, no prediction can be derived for progression in the present data. The available data from adolescents before the end of growth are too small for this so far.

The curvature shape in the Lenke classification:

The classification of LSTVs as mostly Lenke 5 and Lenke 1 requires a specific examination of the lumbosacral junction in the diagnosis of this form of scoliosis (Figure 4). In contrast, the distribution of chronic low back pain does not provide any further differentiation and shows a comparable distribution as that in the symptom-free patient group.

Diff-VMVA in the lowest vertebral segments:

A decreased Diff-VMVA value corresponds to a rounded promontory and may be taken as an indication of LSTVs. In these cases, a generous indication for MRI examination should be made.

The finding by Farshad et al. that a Diff-VMVA of < 10° is indicative of LSTV III and IV is confirmed in their data [10].

Lumbalgia:

Half of the patients in this study complain of chronic low back pain and are mostly classified as Castellvi II (pseudarthrosis). A gender-specific difference cannot be ascertained and may thus be attributed to anatomical change. Any scoliosis with complaints in the lumbosacral transition region must therefore be referred to a differentiated examination.

Prevalence of LSTVs:

Patients classified as Castellvi I were not included in this study because the delineation with the 19 mm width of the processus transversus was found to be too blurred. The radiological appearance of the width of the processus transversus in the radiograph depends on the pelvic tilt and the radiographic focus. In MRI examination, the imaging quality of 1.5-Tesla devices is insufficient for measurements. Even with 3.5-Tesla devices, the sectioning does not always provide clear results. Scoliosis inevitably leads to unclear results due to its curvature and rotation. A reliable classification would only be possible with a differentiated examination focused on this issue. This does not mean that patients with LSTV Castellvi I may not also have complaints due to this change.

Under these conditions, the data presented cannot be compared with the prevalence of other studies. The central message of this study is that if scoliosis is present, the possibility of LSTVs should also be considered. This is especially true when scoliosis is combined with lumbar symptoms.

Jancuska et al. (2015) focus on the clinical significance of LSTVs in their review. The problem of diagnosis is also addressed as follows: “LSTV are often inaccurately detected and classified on standard AP radiographs and MRI” [25].

McGrath et al. (2021) also address the frequency distribution, noting that the complaints particularly affect Castellvi II, III, and IV [26].

Apazides et al. (2011) also describes the frequency distribution problem and the related high differences in individual studies [21].

The study by McGrath et al. describes the lumbosacral ligamentous apparatus and thus goes one step further in differentiation. However, the result requires a very high quality of MRI examination. This quality difference problem between individual radiological institutes became apparent from the beginning of their first examinations and was solved by the fact that all examinations were performed by one institute according to the agreed focus [27].

Under these conditions, the frequency distribution of LSTVs may not be directly compared with individual studies from the literature. This is certainly the main reason for the high differences in frequency distribution between individual studies.

The literature: Scoliosis and LSTVs:

The topic of scoliosis and LSTVs is the focus of the PubMed (18 April 2022) search with 13 entries.

Among them, nine papers deal with the relationship of the surgical approach to scoliosis and LSTVs. In particular, they focus on the question of surgical fusion distance and adjacent disc degeneration [28,29,30,31,32,33,34,35].

There are only four contributions for clinical issues related to scoliosis and LSTVs.

In one paper, the incidence of LSTVs in young men in military school in Turkey is surveyed. In another paper, the number of thoracic and lumbar vertebral bodies in adolescent idiopathic scoliosis is shown. There is a paper that is also devoted to the degeneration of the adjacent vertebral segment. A final paper deals with the clinical significance of LSTVs in patients with idiopathic scoliosis [18,19,36,37].

Limitation:

In MRI examinations, there are imaging inaccuracies that depend on the angulation of the device setting to the vertebral bodies. Thus, in any lumbar scoliosis, orthograde imaging of each vertebral body is not guaranteed. Thus, blurring cannot be avoided. A CT scan would image the bony structures more accurately. For reasons of radiation hygiene, this is prohibited as a routine examination.

A comparison of the frequency with other studies is not permissible since only Castellvi II–IV is considered.

Another problem becomes apparent in the present work on scoliosis:

How can LSTV Castellvi IIA be differentiated from a decompensated lumbar curve with a contact zone of the processus transversus with the os sacrum? The following question also emerges related to the increasing age of the patients and degenerative lumbar curvatures: “What is the cause and what is the result?” Keep in mind that it is a malformation scoliosis with all its therapeutic limitations. In conclusion, for scoliosis Lenke 5 and Lenke 1, the lumbosacral junction should be examined.