Classification in Brief: SRS-Schwab Classification of Adult Spinal Deformity

History

The term adult spinal deformity encompasses a large and complex group of pathologies (scoliosis, kyphosis, etc) with diverse clinical and radiographic presentations. The exact prevalence of adult spinal deformity is not known, but it has been reported to be as high as 60% among patients older than 65 years of age [17, 24]. In spinal deformity, most adults present with pain as their primary concern, which differs from adolescent patients with deformity who are more concerned with cosmesis [4]. When treating adults with spinal deformity, special attention should be placed on sagittal alignment, because this radiographic parameter has been the most associated with pain [5-7, 16].

A classification system designed by Aebi et al. [1] in 2005 grouped spine deformity based on etiology. This scheme proved simple to use, but surgical planning and treatment decisions were difficult using this system [20]. In 2006, Schwab et al. [18] sought to use radiographic parameters that correlated with patient-reported outcomes to design an adult idiopathic scoliosis classification system. The initial study involved 947 adult patients and they were classified into five deformity apex curve types and two modifiers (lumbar lordosis and subluxation). Also in 2006, the Scoliosis Research Society (SRS) published a largely descriptive classification system with seven curve types and three modifiers accounting for regional sagittal, lumbar degenerative, and global alignment [13]. Although this system was reliable, it was mainly descriptive without taking into account clinical parameters, making it less valuable in practice [20]. Then in 2012, Schwab and the SRS published a hybrid classification system (the SRS-Schwab classification) [20]. Similar to the initial classification of Schwab et al. [18], this one correlated radiographic deformity with patient-reported outcomes, but the new classification [14] added radiographic pelvic parameters such as pelvic incidence, lumbar lordosis, pelvic tilt, and sagittal vertical axis. The rationale behind the SRS-Schwab classification was that pelvic alignment is related to sagittal spinal alignment, and the postoperative pelvic parameters are tightly correlated both to pain and spine-related disability [10].

Purpose

Schwab’s previous classification system, which focused on clinically relevant parameters, proved to be reliable and clinically relevant, but it lacked sagittal spinopelvic parameters and some descriptive detail [23]. The newer SRS-Schwab classification system added these spinopelvic parameters linked with pain and disability [10] to continue Schwab’s original goal of maintaining clinical relevance, which the older SRS classification system was lacking. The scheme’s developers also sought consistent characterization of a clinical condition, a way to establish treatment recommendations that was lacking in the previous Aebi classification system [23], and a basis to compare different treatment options [20].

Description of the SRS-Schwab Classification System

The SRS-Schwab classification system (Table 1) uses frontal and sagittal full-length radiographs that are divided into coronal curve types and sagittal curve modifiers. The coronal curve types are classified as follows:

Table 1.

SRS-Schwab classification system

Curve type T (Fig. 1A): thoracic major curve > 30° (apical level of T9 or higher);

Fig. 1 A-B.

(A) This patient’s radiographs show a lumbar curve type (single curve: lumbar 41° with apex at L2), PI-LL: ++ (21°), PT: + (27°), SVA: + (6.2 cm). (B) This patient’s radiographs show a double curve type (double curve: thoracic 72°, lumbar 59° with apex at L2), PI-LL + (15°), PT: + (24°), SVA: 0 (1.6 cm). PI-LL = pelvic incidence minus lumbar lordosis; PT = pelvic tilt; SVA = sagittal vertical axis.

Curve type L: lumbar or thoracolumbar major curve > 30° (apical level of T10 or lower);

Curve type D: double major curve with each curve > 30° (Fig. 1B);

Curve type N: normal or no coronal curve > 30° (ie, no major coronal deformity).

The first sagittal modifier takes into account two radiographic parameters, pelvic incidence and lumbar lordosis to calculate the difference between the two. This measurement is important for surgical planning for patients with a small lumbar lordosis relative to their pelvic incidence to achieve proper postoperative lumbar alignment [20]. Lumbar lordosis is the measurement of the sagittal Cobb angle between the superior endplate of S1 and the superior endplate of L1. Pelvic incidence is the angle between the midpoint perpendicular axis of the sagittal endplate and a line from the midpoint of the sagittal endpoint to the bicoxofemoral axis. Patients are classified as having a pelvic incidence minus lumbar lordosis modifier “0” if the mismatch is < 10°, modifier “+” if it is between 10° and 20°, and modifier “++” if > 20°.

Pelvic tilt is another sagittal modifier important to measure because a high pelvic tilt reflects a compensatory mechanism that reduces the apparent global sagittal alignment [10]. Patients with greater pelvic tilt often need larger corrections (osteotomies) to reduce the risk of postoperative failures [20]. Pelvic tilt assesses the degree of retroversion by measuring the angle between the line from midpoint on the sacral endplate to the bicoxofemoral axis and the vertical axis through the bicoxofemoral axis. Patients are classified as pelvic tilt modifier “0” if the measurement is < 20°, modifier “+” if it is between 20° and 30°, and modifier “++” if the pelvic tilt is > 30°.

The last sagittal modifier group is the global alignment modifier, which is based off of the sagittal vertical axis. An increase in the sagittal vertical axis is associated with increased pain and disability [20] and is defined as the distance between the posterosuperior corner of the sacrum and the sagittal C7 plum line. Patients are classified as having a sagittal vertical axis modifier “0” if the measurement is < 40 mm, modifier “+” if it is 40 to 95 mm, and modifier “++” if > 95 mm.

Validation

Three validation studies have been performed, and all found generally high intra- and interobserver reliability of the SRS-Schwab classification (Table 2). The first was performed by the designers of the classification [20], whereas the others were performed by groups of spine subspecialists and trainees [12, 15]. The three studies had intraobserver repeatability κ coefficient values ranging from 0.67 to 0.87 (which represents substantial to almost perfect reliability according to one scale [11]) and interobserver values ranging from 0.55 to 0.75 (which represents moderate to substantial reliability).

Table 2.

Studies evaluating reliability of the SRS-Schwab classification system for adult spinal deformity

Individual components of the classification system showed better reliability than the overall scores. The sagittal vertical axis modifier had the highest reliability (0.90-0.97), whereas the pelvic incidence minus the lumbar lordosis had the lowest reliability, although still more than adequate (0.70-0.88) [12, 15, 20]. Although the studies had differing levels of experience in the reviewing surgeons (Table 2), there was no difference noted in association with reliability.

The SRS-Schwab system focuses more on spinopelvic parameters compared with earlier systems because they have been shown to correlate with better health-related quality-of-life (HRQoL) scores after surgery [2, 10]. Schwab et al. [19] conducted a study in 2010 evaluating the relationship between spinal alignment parameters and planning corrective surgery. The data in the study resulted in recommendations on how much surgical correction was needed to achieve the best HRQoL scores. Another study showed the classification system can predict patient disability and preference for nonoperative versus operative treatment decisions [21, 22]. Patients with thoracolumbar and primary sagittal deformities had poorer health status and greater disability than other curve types. This study also found operative patients to have worse spinopelvic modifier grades than nonoperative patients with increasingly worse grades correlating to worsening HRQoL scores. A recent study also found a worse prognosis with patients undergoing surgery with a higher sagittal alignment modifier [14]. The importance that Schwab et al. [20] has placed on sagittal realignment objectives in surgical treatment of adult spinal deformity is potentially the most important contribution to the Schwab-SRS classification given the strong correlation with postoperative disability and pain. These studies emphasize the utility in the SRS-Schwab system in regard to treatment decisions and prognostic information.

Limitations

In the 6 years since the SRS-Schwab classification system was first published, there have been only three studies measuring its reliability. This can be a limitation associated with this system, although they all are reasonably well designed and have general agreement on acceptable reliability rates. The system also has 108 possible combinations, which may make it difficult to achieve a high level of reliability with providers not familiar with using the scheme, although the evidence in these three studies argues against it because orthopaedic trainees have shown at least moderate agreement.

Aside from reliability, a good classification system should be able to guide treatment decisions and give prognostic information. Studies that focus specifically on the SRS-Schwab classification system and clinical correlation are sparse, but have shown good prognostic and decision-making results [2, 21]. However, a study by Hallager et al. [8] found that the age and etiology of spinal deformity affected HRQoL and should not be neglected because it is the SRS-Schwab scheme. Lafage et al. [9] also showed that sagittal spinopelvic parameters vary with age and correction should be adjusted accordingly to not overcorrect, which would result in worse HRQoL scores. Although this is hardly a limitation to the SRS-Schwab system, it is important to be aware of these variations when making surgical decisions based on this classification system.

Since the SRS-Schwab scheme was developed, one new system created by Berjano and Lamartina [3] was developed to assist surgeons with operative planning. The goal for their system was to identify selected fusions to avoid full curve fusion. This would reduce surgical risk while maintaining a larger ROM and reduce the risk of junctional disease and decompensation. Compared with Berjano and Lamartina’s scheme, the SRS-Schwab classification system is limited when it comes to operative planning.

Another limitation with the SRS-Schwab classification system is it is only helpful for adult idiopathic scoliosis. The newer Berjano and Lamartina scheme [3] in comparison focuses on degenerative deformities in adult life. Clinicians need to understand which system to use for the appropriate condition.

Another difference between Liu et al.’s and Nielsen et al.’s studies was the amount of pretraining using the system. The former had 30 sets of radiographs in their pretraining, whereas the latter had only six sets without any formal training. Nielsen et al. had lower published rates of inter- and intraobserver reliability, perhaps related to the small number of surgeons (n = 3) participating and the lack of a formalized training process.

Conclusion and Uses

The inter- and intraobserver agreement for the SRS-Schwab classification system for adult spinal deformity ranges from moderate to almost perfect. Although the small number of studies [12, 15, 20] evaluating the reliability of this classification system is a limitation, it remains a widely used system for grading adult spinal deformity. Studies have shown this system has utility in predicting prognosis [2, 10, 14] while helping to guide treatment decisions for surgeons [19, 21]. The authors believe that any further research or improvement to this classification system should focus on guiding treatment. There are few studies using this scheme to make treatment decisions and those that do only used certain modifiers. Overall, the SRS-Schwab classification system allows communication between surgeons, is a reliable system, and is correlated with HRQoL measures giving it prognostic value.