Simultaneous progression patterns of scoliosis, pelvic obliquity, and hip subluxation/dislocation in non-ambulatory neuromuscular patients: an approach to deformity documentation

Antero-posterior spine radiographs were measured using the Cobb criteria. Normal spine measurement in assessing for scoliosis using standing, sitting, or supine radiographs for the purpose of this study was considered to be 0° since in this patient population slight deformity beyond 0° generally indicates the beginning of the progression into scoliosis. In the large majority of scoliosis cases, the lumbar component of the curve continued beyond L5 into the sacrum, stressing the need for spinal radiographs to include the lumbosacral junction.

Pelvic obliquity was measured on sitting antero-posterior full spine radiographs. A line was drawn linking the most superior part of each (right and left) iliac crest, the horizontal axis (in the sitting position) was parallel to the lowermost exposure line of the radiograph at the lower end of the pelvis, and pelvic obliquity was the angle formed by these two lines. While the optimal measurement is 0°, we consider any measurement in the range of 0–1.9° to be normal since values of zero are not invariably seen and, in a practical sense, there is often difficulty in measuring in this patient population.

Hip position was quantified by determining the migration percentage assessing the ossified femoral head coverage by the bony acetabulum. Hip subluxation was measured on the antero-posterior bilateral hip radiograph by drawing a perpendicular line (to the horizontal axis) from the outer bony margin of the acetabulum (Perkin’s line) passing distally through the femoral head. The migration percentage (or index) is a measure of the percentage of the ossified femoral head left uncovered by the ossified acetabular roof; that is, the amount (or percentage) of femoral head outside (lateral to) Perkins’ line on the frontal view [9‐11]. This measurement also indicates the percentage of the femoral head covered by the acetabulum and this latter parameter was used to quantify position. Normal coverage of each femoral head by the outer edge of the acetabulum is considered to be 67 % or greater on the antero-posterior radiograph (2/3 of the bone of the femoral head is contained within the lateral bony margin of the acetabulum, with 1/3 outside).

Spinal flexibility which decreases scoliosis deformity can be determined in three ways: (1) antero-posterior spine radiograph in the supine position (out of brace); (2) antero-posterior spine radiograph in the supine position with bending to straighten the spine (i.e., bending towards the convexity and away from the concavity); and (3) antero-posterior spine radiograph in the sitting position in brace. The Cobb angle is then measured and compared with the scoliosis deformity angle in the maximal functional position out of brace.

Pelvic obliquity can often be seen to diminish in sitting antero-posterior spine radiographs in brace, indicating some degree of flexibility in pelvic deformity. It became evident by qualitative observation of large numbers of radiographs that pelvic obliquity also tends to diminish in supine compared to sitting radiographs, at least until rigidity sets in. There is, however, no standardized way to document this. We developed a way to document supine pelvic obliquity (and thus flexibility of the pelvic obliquity deformity) by carefully positioning the patient supine on the radiographic table with the head, trunk, and lower extremities as anatomically straight as possible. The patient was positioned with each segment (trunk, pelvis, lower extremities) centered as closely as possible in relation to either side of the table. The radiograph was centered on the pelvis but included the lumbosacral region of the spine and the proximal half of both femurs. To determine supine pelvic obliquity, the horizontal axis for measurement is a line placed within the pelvis perpendicular to left and right sides of the image and the iliac crest axis is the line linking the most superior point of each iliac crest. The angle formed between these two lines is the supine pelvic obliquity.

The frog lateral hip radiograph determines the extent of hip deformity flexibility. A dislocated or subluxed femoral head on the antero-posterior view can either reduce fully into the acetabulum, partially, or not at all. The migration percentage can be determined on the frog lateral radiograph by the same method of measurement used on the frontal antero-posterior view since the pelvis and acetabulae are in the same position for both projections and it is only the proximal femurs that move with the frog lateral positioning.

The deformity and flexibility assessments are outlined in Table 1 and Fig. 2a–c.

Each deformity is listed using a specific symbol and line linking each study date. Specific symbols along with associated specific shadings (black, gray, white), line thickness, and line pattern (solid, dotted, etc.) were established for sitting AP (antero-posterior) spine out of brace (OOB), sitting AP spine in brace (brace), supine AP spine (supine), and supine bending spine (bending); AP pelvis sitting (sitting), AP pelvis sitting in brace (brace) and pelvis supine (supine); and hip AP (AP) and hip frog lateral (lateral). The age of the patient at the time of each study is shown along the horizontal x-axis while the vertical y-axis numbered from 0 to 120 indicates both deformity and flexibility/rigidity by angular measurements in degrees for scoliosis and pelvic obliquity and the percent coverage of the femoral head by the acetabulum for hip position. Considering the ranges of scoliosis, pelvic obliquity, and femoral head coverage, the values tend to be dispersed at specific regions of the chart in the normal or with early deformity, allowing simultaneous visual appreciation of the pattern of positioning with time. In a normal child or one with very little deformity, the percent coverage of the femoral head is towards the upper part of the 0–120 range while the scoliosis and pelvic obliquity readings fall in the bottom part. With this method of numerical charting, as scoliosis and pelvic obliquity increase (worsen) the values track upwards; as hip position progressively worsens towards full dislocation, however, the percent coverage decreases and a downward slope for the values indicates progressive deformity.

As increasing numbers of patients were assessed, the specific numerical ranges for grading each deformity and its degree of flexibility/rigidity from mild to moderate to severe allowed for an alternative way of tracking deformation. In a qualitative sense these changes are referred to as fully reducible, partially reducible, and non-reducible, although these concepts are not charted. For charting, age was plotted in similar fashion on the x-axis, with grades from 1 (least deformity) to 5 (greatest deformity) placed along the y-axis. The grade correspondences to the numerical ranges of deformity and flexibility/rigidity for each of the three regions are shown in Fig. 2a–c. With this method of charting by grade, worsening in each of the three parameters of deformity always tracks upwards.

The underlying causes of the three inter-related deformities differ significantly depending on the neuromuscular disorder involved even though there are similarities in the clinical and radiographic appearances. The triadic complex is initiated by asymmetric muscle function due to either weakness or spasticity or both. Secondary deformity then worsens as irregular growth of soft tissues and the skeleton is superimposed on the abnormal positions. In general, the earlier the triad develops the worse the deformity, since growth abnormality is superimposed on positional deformity and rigidity increases with time.

Deformity in each component tends to develop early in the first decade (ages 2–5) in severe quadriparetic CP, SMA type II, the structural congenital myopathies and congenital muscular dystrophy (especially if merosin-negative). Deformity develops in DMD most rapidly in the second decade after loss of walking, especially in the absence of steroid management [13]. Steroids have been shown to prolong ambulation and decrease scoliosis in DMD [14]. The almost invariable early onset and progression of thoracolumbar scoliosis in severely involved neuromuscular patients has been well documented [2, 15‐17]. The scoliosis deformity worsens with the severity of the underlying disorder, as demonstrated in Duchenne muscular dystrophy [13, 18], cerebral palsy [2, 19], and spinal muscular atrophy [17]. When patients are non-ambulatory, the severe neuromuscular scoliosis, unlike the patterns seen in ambulatory idiopathic scoliosis patients, almost always extends to involve the lumbosacral region where it contributes to pelvic obliquity [1‐3, 5‐8, 15‐18]. Further evidence of the inter-related nature of the three sites of deformity is provided by specific studies on hip subluxation and dislocation in relation to more proximal scoliosis and pelvic obliquity in cerebral palsy [19‐22], spinal muscular atrophy [21, 23, 24], and Duchenne muscular dystrophy [3, 18, 21].

In the severe neuromuscular disorders of early childhood, the lumbar and thoracolumbar scoliosis that commonly develops almost invariably includes and extends beyond the L5 vertebra into the sacrum. A hypothesis of the sequence of events is that the scoliosis leads to an associated pelvic tilt or obliquity, since there is no space for a compensatory scoliosis correction at the lowermost spinal levels. As the pelvis tilts, the hip on the high side begins to sublux, although initially in a relative sense since it is actually the acetabulum that tilts away from its normal coverage of the femoral head. The entire pelvis tilts with the low side adjacent to the convexity and the high side to the concavity of the scoliosis curve. The hip (femoral head) shows improved coverage on the low side but progressive subluxation deformity on the high side. The femoral head and neck are already positioned into coxa valga and anteversion due to non-ambulation and they further adduct relative to the malpositioned acetabulum. The normal proximal femoral development into varus is slowed in non-ambulatory patients where a coxa valga develops with the neck-shaft angle remaining high and often approaching 180° on the antero-posterior hip radiograph (Fig. 1b). Persisting femoral anteversion contributes to this radiographic finding of coxa valga. The proximal femoral deformities of coxa valga and increased anteversion combined with acetabular subluxation due to the pelvic obliquity progressively predispose to femoral head subluxation and dislocation; in addition, there is lateral acetabular under-development due to asymmetric pressure of the malpositioned head against the lateral acetabular cartilage mass, further minimizing its depth and limiting head coverage. The triad of deformities regardless of cause then follows the classic pattern of orthopedic deformity, being flexible early in the developmental sequence (passively correctable), then partially rigid and eventually markedly rigid (minimal to no change in supine position or with manual manipulations) due to soft tissue contractures and bony deformity. This documentation approach allows for awareness of the simultaneous development of these three inter-related regional structural deformities. The pathoanatomic changes of the femoral–acetabular complex are demonstrated in Fig. 1b.

We had noted previously the essentially invariable scoliosis in non-ambulatory neuromuscular patients to be lumbar or thoracolumbar, to extend and be continuous into the sacrum, and to be associated with a pelvic obliquity. In a subset of 116 patients with scoliosis, the incidence of pelvic obliquity was 98 %.

In then assessing the patterns of inter-related deformity, variability can be seen in the relative progression of deformity in the entire group of patients. Spinal deformities (scoliosis) developed more rapidly than hip deformities in 44 %, spine and hip deformities developed at the same time in 40 %, and hip deformities developed more rapidly than spine deformities in only 16 %. This variability was assessed statistically using Pearson’s chi-squared test where p = 0.0501, an almost significant valuation. In the latter group (hip > spine), 7 of the 10 patients were in the CP group (Table 2). We feel that an adducted hip alone in a CP patient has little tendency, in a mechanical sense, to tilt the pelvis or spine above. It is in the severely quadriparetic CP patients with global involvement that the triad of deformities invariably occurs. Some non-ambulatory cerebral palsy patients with more regional involvement (as distinct from those with global spastic quadriparesis) show hip deformity developing a considerable period of time before pelvis and spine deformity and consequently hip displacement is not always at the high pelvic side. Hodgkinson et al., in a study of 234 non-ambulatory patients with cerebral palsy, clearly recognized the triad of deformities but could not find a relationship between the side of the pelvic obliquity and the side of the scoliosis convexity or hip displacement. They found, however, that pelvic obliquity was a risk factor for hip displacement and that eventually those deformities were a direct risk factor for scoliosis [2]. In the weakest SMA patients (type I and the most involved type II) both hips may sublux and dislocate very early, even before an eventual marked pelvic obliquity establishes itself. This variability in the entire series is seen in the finding of three temporal groups of progression and the variability in the p-value (p = 0.0501, almost significant) (Table 2).