Taking the shoulders and pelvis into account in the preoperative classification of idiopathic scoliosis in adolescents and young adults (a constructive critique of King's and Lenke's systems of classification)

Shoulder Imbalance in Adolescent Idiopathic Scoliosis: A Systematic Review of the Current State of the Art

Background

Shoulder imbalance (SI) is among the most rated manifestations of adolescent idiopathic scoliosis (AIS) pointed to by patients and spine surgeons. It serves as a criterion to assess the outcome of scoliosis surgery and is also a cause of dissatisfaction for the patients postoperatively. Despite the availability of multiple studies on this issue, a comprehensive survey of the risk factors and preventive measures has yet to be elucidated. The present study aimed to highlight the most recent approach to the evaluation and management of SI, as well as medical counseling about the expectations and limitations of the surgery.

Methods

A systematic literature review using electric databases was conducted, including PubMed, Embase, the Cochrane Library, and Google Scholar, with a well-defined search strategy on SI definition, risk factors, and preventive and surgical recommendations.

Results

A total of 69 articles were identified; SI > 2 cm was the most used cut-off, and its risk factors included the main thoracic Cobb angle > 80˚, preoperative level shoulder, high left shoulder, and higher Risser grade. The most stated strategies to preclude SI were the sufficient correction of the proximal thoracic curve, and moderate correction of the main thoracic and lumbar curve (LC).

Conclusion

Shoulder imbalance should be prevented not only for appearance or satisfaction but also for possible complications such as distal adding-on, new LC progression, or trunk shift postoperatively in AIS patients.

Achieving Shoulder Balance Using Medial and Lateral Radiological Measures in Adolescent Idiopathic Scoliosis

BACKGROUND

Research has shown that postoperative shoulder imbalance is a common problem after spinal fusion in adolescent idiopathic scoliosis (AIS). The best radiographic predictor has not yet been determined and results are inconsistent. This study was to investigate whether using medial and lateral shoulder parameters can effectively achieve postoperative shoulder balance.

METHODS

A prospective database of AIS undergoing posterior spinal fusion were reviewed. Patient demographics and radiological parameters including radiographic shoulder height (RSH), clavicle angle, T1-tilt and first-rib angle at baseline, 6 weeks and last minimal follow up of 2 years were recorded. Correlations between radiological parameters were assessed using Pearson's correlation coefficients. Multivariable linear models identified predictors associated with increased RSH.

RESULTS

219 patients (mean age:13.7 years; 81.7% female) were included. The mean follow-up time was 2.8 years (range:2.0-7.0). The mean RSH at baseline, 6 weeks and last follow up was improved significantly at 95.8%. Preoperative (r=0.8; p<0.001) and post-operative measurements of RSH at 6-week (r=0.9; p<0.001) and last follow up (r=0.9; p<0.001) correlated strongly with clavicle angle measured at respective time-points. In a multivariable linear model, we noted marginal increase in clavicle angle (+4.3°; p<0.001) to be associated with increased RSH. On the contrary, first rib angle and T1-tilt demonstrated moderate to weak correlation with RSH.

CONCLUSION

Clavicle angle is strongly consistent with RSH. First rib angle and T1-tilt as demonstrate medial shoulder balance are moderate to weak correlation. Leveling T1 tilt and first rib angle do not guarantee the postoperative shoulder balance. Level of Evidence: IV.

Sinister! The high pre-op left shoulder is less likely to be radiographically balanced at 2 years post-op

PURPOSE

AIS patients consider shoulder balance an important cosmetic outcome after surgery. We examined the impact of preoperative left shoulder elevation (LSE) and choice of upper instrumented vertebra (UIV) on postoperative shoulder imbalance (PostSI).

METHODS

This was a retrospective cohort study utilizing a prospective AIS database. Patients had Lenke type 1-4 curves and preoperative shoulder height ≥ 1.0 cm. Patients with preoperative LSE and right shoulder elevation (RSE) were compared. Shoulder height difference < 1 cm was considered 'mild', 1-2 cm was 'moderate', and ≥ 2.0 cm was 'severe'.

RESULTS

407 patients had ≥ 1.0 cm imbalance preoperatively, with 88 (21.6%) LSE. There were no differences in gender (p = 0.855) or age (p = 0.477). Patients with LSE more frequently had Lenke type 2 curves (43.2% vs 16.3%, p < 0.001), while preoperative RSE averaged 1.9 ± 0.9 cm versus 1.6 ± 0.5 cm for LSE (p < 0.001). Those with LSE more often had severe PostSI at 2 years (30.7% vs 5.0%, p < 0.001), and only 26.1% of patients with severe preoperative LSE corrected to mild. In contrast, most patients with RSE had mild PostSI regardless of initial imbalance. When examining only LSE patients, there was no difference in preoperative SH by final UIV (p = 0.101). Further, UIV choice did not impact the proportion of severely unbalanced patients postoperatively (p = 0.446). A PTC > 34.5° was predictive of PostSI ≥ 2.0 cm for patients with preoperative LSE.

CONCLUSION

AIS patients with preoperative LSE are less likely to achieve level shoulders postoperatively. Choice of higher UIV did not affect postoperative shoulder imbalance in this cohort. A PTC > 34.5° was predictive of severe PostSI in patients with preoperative LSE.

LEVEL OF EVIDENCE

II.

Impact of pelvic obliquity on coronal alignment in patients with adolescent idiopathic scoliosis

STUDY DESIGN

Retrospective study.

OBJECTIVE

To investigate pelvic obliquity prevalence and its coronal alignment effects in patients with adolescent idiopathic scoliosis (AIS). Pelvic obliquity observed on standing radiographs could affect coronal alignment; however, its incidence or relationship with coronal alignment remains unclear.

METHODS

Data of 141 patients with AIS (11 men, 130 women; mean age, 14.2 years) were retrospectively analyzed. Pelvic obliquity (pelvic coronal obliquity angle (PCOA) value ≥ 3°) was evaluated on standing radiographs; PCOAs were classified into two groups: right-upward and left-upward. Cobb angle and flexibility of proximal thoracic, main thoracic, thoracolumbar/lumbar curve, radiographic shoulder height (RSH), L4 tilt, and coronal balance were measured. Iliac crest and femoral head height differences were measured to evaluate correlations between pelvic obliquity and leg length discrepancy. We compared patients with and without pelvic obliquity according to Lenke classifications.

RESULTS

Among 141 patients, 33 (23%) showed pelvic obliquity: 12 were type 1, 3 were type 2, 1 was type 3, 13 were type 5, and 4 were type 6. Right-upward obliquity was observed in 25 patients (76%), all B or C curves; left-upward obliquity was observed in 8 (24%) and most were lumbar modifier A curves. PCOA and iliac crest height difference correlated with femoral head height difference. Among patients with Lenke type 1 with and without pelvic obliquity, those with right-upward pelvic obliquity showed significantly greater absolute RSH values, while those with left-upward pelvic obliquity showed significant smaller absolute RSH values. Among patients with Lenke types 5 and 6, those with pelvic obliquity showed significantly greater L4 tilt absolute values.

CONCLUSIONS

Pelvic obliquity was frequently observed in patients with AIS, especially lumbar scoliosis. Right-upward pelvic obliquity influenced by lumbar curves promoted shoulder imbalance and left-upward obliquity compensated for shoulder imbalance in Lenke type 1.

LEVEL OF EVIDENCE

Level 3.

Pelvic obliquity in adolescent idiopathic scoliosis planned for posterior spinal fusion: A preoperative analysis of 311 lower limb axis films

PURPOSE

To analyze the incidence, pattern, and contributing factors of pelvic obliquity among Adolescent Idiopathic Scoliosis (AIS) patients who will undergo surgery.

METHODS

In total, 311 patients underwent erect whole spine anteroposterior, lateral and lower limb axis films. Radiographic measurements included Transilium Pelvic Height Difference (TPHD; mm), Hip Abduction-Adduction angle (H/Abd-Add; °), Lower limb Length Discrepancy (LLD; mm), and Pelvic Hypoplasia (PH angle; °). The incidence and severity of pelvic obliquity were stratified to Lenke curve subtypes in 311 patients. The causes of pelvic obliquity were analyzed in 57 patients with TPHD ≥10 mm.

RESULTS

The mean Cobb angle was 64.0 ± 17.2°. Sixty-nine patients had a TPHD of 0 mm (22.2%). The TPHD was <5 mm in 134 (43.0%) patients, 5-9 mm in 104 (33.4%) patients, 10-14 mm in 52 (16.7%) patients, 15-19 mm in 19 (6.1%) patients, and ≥20 mm in only 2 (0.6%) patients. There was a significant difference between the Lenke curve types in terms of TPHD (p = 0.002). L6 curve types had the highest TPHD of 9.0 ± 6.3 mm followed by L5 curves, which had a TPHD of 7.1 ± 4.8 mm. In all, 44.2% of L1 curves and 50.0% of L2 curves had positive TPHD compared to 66.7% of L5 curves and 74.1% of L6 curves which had negative TPHD. 33.3% and 24.6% of pelvic obliquity were attributed to PH and LLD, respectively, whereas 10.5% of cases were attributed to H/Abd-Add positioning.

CONCLUSIONS

76.4% of AIS cases had pelvic obliquity <10 mm; 44.2% of L1 curves and 50.0% of L2 curves had a lower right hemipelvis compared to 66.7% of L5 curves and 74.1% of L6 curves, which had a higher right hemipelvis. Among patients with pelvic obliquity ≥10 mm, 33.3% were attributed to PH, whereas 24.6% were attributed to LLD.

Adolescent spinal pain: The pediatric orthopedist's point of view

INTRODUCTION

Ten to twenty percent of persons experience spinal pain during growth. Causes are diverse in adolescents, and it is essential to determine etiology rapidly so as to guide optimal management.

HYPOTHESIS

It is important for the pediatric orthopedist to understand the natural history of conditions inducing spinal pain.

MATERIAL AND METHODS

A retrospective study included 116 adolescents presenting with spinal pain at the Hôpital Nord (Marseille, France) between January 1, 2009 and January 1, 2014. Malignant tumoral etiologies were excluded. Mean patient age was 13.6 years. Risser ranged between >0 and <5. Interview and clinical examination (skin, spine, neurologic examination, general clinical examination) were systematic; depending on results, complementary examinations (imaging, biology, biopsy) were prescribed.

RESULTS

There were 32 cases of non-specific adolescent low back pain, 31 of lumbar or thoracolumbar scoliosis, 23 of spinal growth dystrophy (Scheuermann's disease), 13 of isthmic lysis, 5 of spondylolisthesis, 8 of transitional lumbosacral hinge abnormality, 2 of discal hernia, 1 of osteoid osteoma and 1 of eosinophil granuloma. Treatment was often non-operative when diagnosis was sufficiently early. In case of failure, surgery could generally be considered.

DISCUSSION

Correctly indicated non-operative management or surgery changes the natural history of these pathologies. The aim of treatment is to resolve pain in adolescence, as it risks becoming chronic and disabling by adulthood.

Should Shoulder Balance Determine Proximal Fusion Levels in Patients With Lenke 5 Curves?

STUDY DESIGN

Multicenter review of prospectively collected data.

OBJECTIVES

To identify the frequency of an opposite high shoulder in Lenke 5 patients and evaluate factors that influence preoperative and postoperative shoulder balance.

SUMMARY OF BACKGROUND DATA

A high left shoulder is an indication to extend the fusion proximally in a right thoracic curve. Some apply a similar rule to high right shoulders in patients with left thoracolumbar/lumbar curves.

METHODS

A prospective multicenter adolescent idiopathic scoliosis database was queried for patients with Lenke 5 curves and minimum 2-year follow-up. Preoperative and postoperative shoulder height differences were recorded and categorized by the opposite shoulder (right shoulder in a left thoracolumbar curve) as high (greater than 1 cm), level (0-1 cm), and low (less than 1 cm). Preoperative and postoperative radiographic variables and Scoliosis Research Society questionnaire scores were evaluated.

RESULTS

Of the 104 patients identified, 37% had level shoulders and 53% had a high opposite shoulder. A high shoulder was associated with a greater mean thoracic Cobb (31°) than a level (24°) or low shoulder (26°) (p = .008). Postoperatively, 64% of patients had level shoulders (less than 1 cm); 93% had a shoulder difference less than 2 cm. Preoperative lumbar Cobb was a significant predictor of postoperative shoulder height (p = .051). A slightly greater proportion of preoperative high shoulders (36%) had a nonselective fusion than those with level (27%) or low (9%) shoulders. Among the 29 patients with a preoperative moderate or significant high shoulder (greater than 2 cm), 3 continued to have a high shoulder greater than 2 cm that was not influenced by fusing the thoracic spine. There were no significant differences in preoperative or postoperative Scoliosis Research Society scores based on shoulder height (p > .05).

CONCLUSIONS

Half of all Lenke 5 curves have a high opposite shoulder that is influenced by the size of the compensatory thoracic curve. Postoperatively, most patients had level shoulders. Inclusion of the thoracic spine did not influence postoperative shoulder balance.

Chest cage angle difference and rotation of main thoracic curve are independent risk factors of postoperative shoulder imbalance in surgically treated patients with adolescent idiopathic scoliosis

STUDY DESIGN

Retrospective case series of surgically treated adolescent patients with scoliosis.

OBJECTIVE

To assess the prevalence and independent risk factors for postoperative shoulder imbalance in surgically treated adolescent patients with idiopathic scoliosis.

SUMMARY OF BACKGROUND DATA

Despite recent reports that have identified risk factors for postoperative shoulder imbalance, the relative risks remain unclear.

METHODS

A retrospective review of 85 consecutive patients treated with thoracic fusion with a minimum 2-year follow-up (mean, 3.1 yr) was conducted to investigate the patient radiographical measurements and demographics. Shoulder height difference (SHD) was measured as the graded height difference of the soft tissue shadows. A SHD more than 2 cm indicated an unbalanced shoulder. Patient demographics and radiographical data were studied to determine risk factors for postoperative SHD. The potential risk factors included age, sex, Risser sign, Cobb angle, flexibility, and apical vertebral rotation (AVR) of the main curve, upper-instrumented vertebra level, SHD, and clavicle chest cage angle difference (CCAD). Univariate and multivariate logistic regression analyses were performed to determine the independent risk factors for postoperative shoulder imbalance.

RESULTS

Of the 85 patients, 21 patients presented postoperative shoulder imbalance. The univariate analysis indicated age, Risser sign, Cobb angle of the main curve, AVR of the main curve, and CCAD as risk factors, but the multivariate logistic regression analysis showed that only AVR of the main curve and CCAD were independent risk factors for postoperative shoulder imbalance (AVR, P = 0.04, odds ratio (OR): 3.54; CCAD, P = 0.01, OR: 5.10).

CONCLUSION

Postoperative shoulder imbalance was observed in 25% of the surgically treated adolescent patients. The CCAD and AVR of the main thoracic curve were independent risk factors for postoperative shoulder imbalance in surgically treated patients with adolescent idiopathic scoliosis. The significant correlation between CCAD and postoperative shoulder imbalance seen in this study strongly suggests that the relationship of the shoulder girdle and chest cage has a role in maintaining shoulder balance.

Classification system of the sagittal standing alignment in young adolescent girls

PURPOSE

The purpose of this cohort study was to classify sagittal standing alignment of pre-peak height velocity (pre-PHV) girls, and to evaluate whether identified subgroups were associated with measures of spinal pain. This study further aimed at drawing attention to similarities and differences between the current postural classification and a previous system determined among pre-PHV boys.

METHODS

557 pre-PHV girls [mean age, 10.6 years (SD, 0.47 years)] participated in the study. Three gross body segment orientation parameters and five specific lumbopelvic characteristics were quantified during habitual standing. Postural subgroups were determined by cluster analysis. Logistic regression was applied to assess the relationship between postural subgroups and spinal pain measures (pain and seeking care, assessed by self-administered questionnaire). Chi-square statistics, independent samples T test, and distribution-based methods were used for comparison with postural categorization in pre-PHV boys.

RESULTS AND CONCLUSION

Among pre-PHV girls, clinically meaningful posture clusters emerged both on the gross body segment and specific lumbopelvic level. The postural subtypes identified among pre-PHV girls closely corresponded to those previously described in pre-PHV boys, thereby allowing the use of the same, working nomenclature. In contrast to previous findings among pre-PHV boys, no associations between posture clusters and spinal pain measures were significant in girls at pre-PHV age. When comparing discrete 'global' alignment scores across corresponding posture types, some intriguing differences were found between genders which might involve different biomechanical loading patterns. Whether habitual posture forms a risk factor for developing spinal pain up to adulthood needs evaluation in prospective multifactorial follow-up research.

Pelvic morphology, body posture and standing balance characteristics of adolescent able-bodied and idiopathic scoliosis girls

The purpose of this study was to determine how pelvic morphology, body posture, and standing balance variables of scoliotic girls differ from those of able-bodied girls, and to classify neuro-biomechanical variables in terms of a lower number of unobserved variables. Twenty-eight scoliotic and twenty-five non-scoliotic able-bodied girls participated in this study. 3D coordinates of ten anatomic body landmarks were used to describe pelvic morphology and trunk posture using a Flock of Birds system. Standing balance was measured using a force plate to identify the center of pressure (COP), and its anteroposterior (AP) and mediolateral (ML) displacements. A multivariate analysis of variance (MANOVA) was performed to determine differences between the two groups. A factor analysis was used to identify factors that best describe both groups. Statistical differences were identified between the groups for each of the parameter types. While spatial orientation of the pelvis was similar in both groups, five of the eight trunk postural variables of the scoliotic group were significantly different that the able-bodied group. Also, five out of the seven standing balance variables were higher in the scoliotic girls. Approximately 60% of the variation is supported by 4 factors that can be associated with a set of variables; standing balance variables (factor 1), body posture variables (factor 2), and pelvic morphology variables (factors 3 and 4). Pelvic distortion, body posture asymmetry, and standing imbalance are more pronounced in scoliotic girls, when compared to able-bodied girls. These findings may be beneficial when addressing balance and ankle proprioception exercises for the scoliotic population.

Clavicle chest cage angle difference (CCAD): a novel predictor of postoperative shoulder imbalance in patients with adolescent idiopathic scoliosis

STUDY DESIGN

Retrospective case series of surgically treated patients with adolescent scoliosis.

OBJECTIVE

To establish a new radiographical measurement method to determine the best preoperative predictor of postoperative shoulder balance.

SUMMARY OF BACKGROUND DATA

Shoulder balance is an important aspect of the overall cosmetic balance in patients with adolescent idiopathic scoliosis. Despite recent reports, it is still difficult to estimate the postoperative shoulder balance accurately.

METHODS

A retrospective review of 89 consecutive patients who had thoracic fusion with a minimum 2-year follow-up (mean, 3.1 yr) was conducted to investigate the radiographical measurements and patient demographics. The shoulder height difference (SHD) was measured as the graded height difference of the soft tissue shadows. SHD more than 2 cm indicated an unbalanced shoulder. The clavicle chest cage angle difference (CCAD) was established and evaluated. The CCAD was graded as grade A: no imbalance (<0°), grade B: mild imbalance (0°-10°), and grade C: significant imbalance (>10°).

RESULTS

Of the 89 patients, 22 patients had a moderate or significant SHD at 2 years postoperatively and were categorized as the unbalanced shoulder group (unbalanced SD). A significant difference was observed in preoperative CCAD between the balanced and unbalanced SD groups (P = 0.01). The intraclass correlation coefficient for CCAD was 0.94 among the observers. CCAD was consistent from the preoperative to the final postoperative follow-ups in both groups. The classification of the CCAD preoperatively indicated that 12 of 22 (54.4%) patients who were classified into the postoperative unbalanced SD group showed grade 3 CCAD preoperatively, whereas only 9 of 67 (13.4%) patients who were classified in the postoperative balanced SD group had grade 3 CCAD preoperatively.

CONCLUSION

The developed method to predict postoperative shoulder balance was demonstrated to be easy to perform, reliable, and practical. Additionally, we classified the estimation of postoperative shoulder imbalance by preoperation CCAD. The results of this comprehensive review will guide spinal surgeons in their preoperative planning and in the surgical management of adolescent idiopathic scoliosis to reduce postoperative shoulder imbalance.

LEVEL OF EVIDENCE

4.

Anterior instrumentation (dual screws single rod system) for the surgical treatment of idiopathic scoliosis in the lumbar area: a prospective study on 33 adolescents and young adults, based on a new system of classification

OBJECTIVES/PURPOSE

The choice of anterior instrumentation in the treatment of lumbar scoliosis in adolescents and young adults is not a new topic for the authors. The first results achieved using the Dwyer surgical modality were reported by one of the authors followed by the results achieved using Zielke (VDS) instrumentation. Today, new techniques and new instrumentations have been developed that challenge the instrumentation choices. Here we describe how the new system of classification of scoliotic curves we developed has been used as a basis for treating idiopathic scoliosis in lumbar area in adolescents and young adults using an anterior approach.

MATERIALS

A prospective study was carried out between 1998 and 2010 at two hospital centers on 33 adolescents and young adult with idiopathic lumbar scoliosis involving curves of three kinds, on whom surgical treatment was performed using a single solid rod. Topography of curves: our system of classification includes curves corresponding to the following three type of scoliosis: Type K I: double thoracic and lumbar curves (lumbar predominant) scoliosis (17 cases) mean age 16 years all female patients. Mean Cobb angle of lumbar curve 41°. Mean Cobb angle of thoracic curve 28°. The lumbar curve was left hand convex in 15 cases and right hand convex in 2 cases. Horizontal tilting of L4 mean value 22°. C7 offset mean value 3 cm. Type K IV A: unbalanced thoracolumbar scoliosis (13 cases) mean age 17 years, ten female patients and three male patients. Mean Cobb angle of thoracolumbar curve 39°. The thoracolumbar curve was left hand convex 4 times and right hand convex 9 times. Horizontal tilting of L4 mean value 18°. C 7 offset mean value 2.5 cm. Type K VI A: real lumbar (three cases). Age: 17, 15 and 13 years; all female patients. Cobb angle of the lumbar curve 66°, 29° and 70° (all LH convex). Horizontal tilting of L4: 40°, 20° and 46°. C 7 offset: 7 cm, 1 cm and 4 cm.

METHODS

Surgical instrumentation: We used the EUROS AZUR anterior instrumentation for all the procedures. Cages have been used on five patients at the lower stages. Number of vertebrae instrumented: mean five vertebrae. The patients did not wear postoperative orthosis. Mean duration of procedure: 3 h 50 min. Mean blood loss: 350 cm(3).

RESULTS

Type K I scoliosis (17 cases): Mean follow-up: 6 years. Correction of the lumbar curve Cobb angle: the mean angle has been corrected from 41° to 21°. Number of vertebrae instrumented: 4:6 times and 5:11 times. Correction of the upper thoracic curve Cobb angle: mean angle corrected from 28° to 19°. Correction of L4 horizontal tilting: mean residual was 7°. Correction of C 7 offset: mean 0.7 cm. Type K IV A scoliosis (13 cases): mean follow-up: 4 years. Correction of the lumbar curve Cobb angle: the mean angle has been corrected from 39° to 16°. Mean number of instrumented vertebrae: 5 (4:4 times, 5:6 times and 6:3 times.) Correction of L4 horizontal tilting: mean residual 5°. Correction of C 7 offset: mean 0.7 cm. Type K VI A scoliosis (three cases): mean follow-up: 7, 2 and 4 years; Correction of the lumbar curve Cobb angle: the angles have been corrected from 66° to 15°, from 29° to 11° and from 70° to 28°. Number of instrumented vertebrae: 5, 4 and 6. Correction of L4 horizontal tilting: residual tilting of 8°, 7° and 17°. Correction of C 7 offset: 1 cm, 0 cm and 1 cm.

COMPLICATIONS

There has been no report early or late septic or vascular or neurological complications. Instrumentation failure: there were three cases of screw breakage, all occurred on the lowest implant. Revision surgery was undertaken in both cases, only the last plate needed to be replaced and the rod could be kept without any other modification of the construct. In both cases, fusion has been achieved without any loss of correction. The mean loss of correction of the main curve was 2.5° for the three series.

CONCLUSIONS

Anterior instrumentation of lumbar idiopathic scoliosis gives highly satisfactory morphological and functional results, since the lumbar musculature is spared and the instrumentation placed at the apex of the curvature has selective effects. Despite our preference and that of other surgeons throughout the world for anterior instrumentation, we are still a minority in comparison with the users of posterior instrumentation. There are several reasons for this reticence, including surgeons' training and ideas about pedicular screw fixation, but the main reason has been the lack of a sufficiently exact system of classification. Previous comparative studies between the anterior and posterior approaches have been biased by the use of an excessively restrictive mode of classification (lumbar/thoracolumbar) of the curves. Real lumbar scoliosis, unbalanced thoracolumbar scoliosis and thoracic and lumbar double curve (lumbar predominant) scoliosis should be properly defined before being compared.