Quantitative Evaluation of Upright Posture by X-Ray and 3D Stereophotogrammetry With an Original Marker Placement Protocol in Late Onset Pompe Disease.

Background: Late Onset Pompe Disease (LOPD) is an autosomal recessive muscular disorder characterized by prevailing weakness of trunk and pelvic girdle muscles that causes ventilatory insuciency and postural abnormalities. The most common myopathy phenotype described clinically in LOPD is the Limb Girdle and Diaphragmatic Pattern; spinal deformities, as hyperlordosis, hyperkyphosis and scoliosis diagnosed by x-Ray exam, have been reported in about a third of LOPD patients. The non-specic clinic onset, the similarity of the LOPD phenotypes with other myopathies, inter-individual heterogeneity and the lack of any disease hallmark, make early diagnosis challenging and, if enzyme replacement therapy does not begin timely, about 60% of patients develops severe motor and ventilatory disabilities. Aim of our study was to quantitatively assess the upright posture in a sibship of LOPD adults by x-Ray(xR) and 3D Stereophotogrammetry (St), considered the gold standard to measure spinal angles and whole-body posture respectively, in order to better identify specic alterations more likely to be present in a homogeneous group. Results: Statistical analysis of St parameters showed a larger ankle, knee, elbow, dorsal, S2-C7, heel-S2-C7, heel-S2-nasion angles and a lower sagittal vertical axis (SVA) than healthy controls. Moreover, xR analysis highlighted a lower occipito-cervical, C2-C7 cervical and Cobb dorsal angles, and a trend to lower lumbar lordosis and SVA compared to normal values. Pearson’s coecient analysis was carried on in order to evaluate the correlation between xR and St sagittal spino-pelvic parameters and signicant correlation was found in dorsal and lumbar angles calculated using xR markers placed on spiny apophysis, xR centre of vertebral bodies, Cobb method and St markers. Conclusions: This is the rst study that quantitatively assesses standing whole-body alignment and postural abnormalities in LOPD. These postural alterations are not easily detectable during clinical examination and might be useful to early identify LOPD patients and to facilitate differential diagnosis other proximal myopathies. placement protocol showed high to compared to to and of LOPD with other proximal myopathies.


Background
Pompe Disease is a rare autosomal recessive neuromuscular disorder, due to mutations of the GAA gene codifying the lysosomal acid alpha1-4 -glucosidase enzyme (acid maltase, GAA), and characterized by glycogen accumulation primarily in skeletal, cardiac and smooth muscle [1]. Mutations causing reduction of musculoskeletal GAA activity below 1% the normal, lead to very severe infantile form (IOPD) which is incompatible with life if not treated early with enzyme replacement therapy (ERT) [2]. Levels of GAA activity between 1%-25% the normal, cause late onset form (LOPD) which may occur at any age from 2 to 70 years. This mild form has a better prognosis for life expectance than IOPD form, but about 35% of patients is con ned to wheelchair and/or requires mechanical ventilation before the age of 65, and all LOPD patients show postural and motor disabilities [3]. Recent clinical trials demonstrated that ERT leads to better functional results if it is established earlier and associated with rehabilitation [4,5]. Unfortunately, LOPD diagnosis occurs after an average delay of 10 years, owing to the variability of clinical manifestations, the rarity of the condition and the overlap of signs and symptoms with other neuromuscular diseases [6].
The phenotypic spectrum of LOPD is quite broad considering the different age of onset, (varying from early infancy to late adulthood), the heterogeneity of musculoskeletal involvement and the rate of disease progression, even in patients carrying the same mutations of GAA. Despite these differences, a recent systematic review reported that about 80% of LOPD patients showed a clinical triad that consists of a primary involvement of the pelvic girdle, respiratory and trunk musculature [7]. LOPD patients were also classi ed into three clinical subgroups, the Limb Girdle and Diaphragmatic Weakness (LGDW), the Rigid Spine Syndrome, Scoliosis and Low Body Weight (RSS) and the Cardiocerebrovascular patterns, in relation to their shared prevailing symptoms [8].
MRI studies revealed an involvement of para-spinal lumbar muscles (multi dus, longissimus, iliocostalis muscles) in pre-symptomatic patients with iperCKemia as the only manifestation, whereas abdominal rectus, psoas, iliacus and posterior muscle of the thigh were further involved in symptomatic patients [9]; the severity of muscle changes at MRI-T1w score strongly correlated with muscle weakness, exercise intolerance, decreased endurance, fatigue, di culties during transitions between positions, reaching upright standing (Gowers' maneuver), ambulation, running and climbing stairs [7,9].
The loss of muscular function is associated to skeleton alterations as osteoporosis and spinal deformities. A third of LOPD patients showed scoliosis and only a quarter of them had an adult-onset form; moreover, more than half of LOPD patients with scoliosis had been diagnosed by xR and, in these subgroup, kyphosis and lordosis abnormalities were respectively found in 13.9% and 15.9% of patients, even if the type of abnormalities was never speci ed [10].
To the best of our knowledge, only one study assessed the stabilometric parameters in LOPD using a force platform, but no study has ever analysed quantitatively whole-body standing alignment in LOPD patients [11].
Based on these observations, aim of our study was to quantitatively assess postural abnormalities in a sibship of LOPD adults by gold standard methods: x-Ray for spinopelvic sagittal angles and 3D Stereo-photogrammetry for whole-body upright posture using an original markers placement protocol. This instrumental evaluation could identify speci c postural parameters useful for early and differential diagnosis with other proximal myopathies, in order to start ERT as soon as possible and to quantitatively monitor the treatment outcomes.

Subject populations
Seven LOPD siblings, sharing the same GAA mutation compound, were recruited in Neuromuscular Disease Centre of the University of Campania "L.Vanvitelli" and assessed using clinical scales (Medical Research Council and Range Of Movement), X-Ray (xR) and 3D Stereophotogrammetry (St) in upright posture. Fourteen healthy individuals, age and sex-matched, were used as controls for St-parameters.
Moreover, radiographic spinopelvic sagittal parameters [16][17][18][19][20][21][22][23] were assessed by x-Ray (OperaG80HF -GMM -Seriate (BG) -Italy), positioning patient in perfect orthostatic position and acquiring plain lm radiograms in a lateral view; radiopaque markers were placed on the spiny apophysis of C7-T7-T12-L3-L5-S2 ( Figure 2, A) in according to DB-Total protocol, to study the correlation between St and xR parameters. Using Matlab, radiological angles were evaluated and plotted on the x-Ray images, as shown and explained in gures 2 B, C. To this aim, an expert radiologist manually identi ed landmark points (vertebral corners, the pro le of the sacral plate, and the pro les of femoral heads) on radiographic images using Osirix; landmarks coordinates were input into ad hoc developed Matlab script.

Statistical analysis
As the aim of the study was to individuate the upright posture parameters more suitable to a characterisation of the LOPD patients compared to the healthy population, for each St and xR parameter the distribution between the two populations was analysed. For clinical and St parameter healthy population was recruited as control group, while for xR normative values were found in literature [16,17,18]. In particular, baseline features and ROM values have been tested for clinical differences between patients and controls using Wilcoxon-Mann-Whitney test and t-test, respectively.
Moreover, Wilcoxon-Mann-Whitney test was used to evaluate separation between patients and healthy subjects for St-parameters, while t-test was used for xR parameters. P-values less than 0.05 were considered statistically signi cant different.
Pearson's correlations coe cients (including con dence intervals) have been computed between St and xR parameters in order to assess the reliability of DB Total markers placement protocol (high correlation coe cient >0.7 and < -0.7). All statistical analysis has been performed in R [13].

Results
Baseline features of LOPD patients (L) and controls (C) are: gender= 4F, 3M in L; 8M, 6F in C; mean age= 57.1 ± 6.5 in L; 46.7 ± 14.9 in C; mean BMI= 26.2 ± 4.2 in L; 23.4 ± 3.0 in C; no signi cant difference was found between two groups. LOPD group showed a lower strength in ex-extensor trunk and tibialis anterior muscles (average MRC=2), while the other dorsi exors ankle (extensor digitalis and hallucis longus) were not compromised; moreover, a signi cant lower active e passive dorsi exion (a-and p-ROM) of ankle was found (Table 1, Figure 3).
As regard the evaluation of upright posture, stereophotogrammetric analysis showed signi cant differences in sagittal parameters, with a larger dorsi exion ankle, exion knee, dorsal, S2-C7, Heel-S2-C7 and Heel-S2-Nasion angles and a lower Sagittal Vertical Axis (SVA) in the Pompe group compared to the healthy controls (Table 2).
About radiographic spinopelvic measurements, signi cant differences were found in sagittal angles with a reduction of occipito-cervical (O-C2), C2-C7 cervical (C2-C7-CA) and Cobb Dorsal angles (C-DA) and a trend of reduction of SVA in LOPD patients (Table 3) respect to the normal values.
Moreover, a strong correlation was found between xR and St parameters: 1) dorsal and lumbar angles calculated using xR markers placed on spiny apophysis (xR-mks) and St markers (St-mks); dorsal angles calculated using 2) St markers (St-mks) and xR centre of vertebral bodies (xR-cvb), 3) xR-mks and xR-cvb and 4) xR-mks and Cobb Angle (Table 4).

Discussion
The aim of the study was to quantitatively assess the muscular weakness of speci c anatomic districts and related postural alterations in upright standing of LOPD patients. Our results showed a de cit of trunk muscles, in accordance to previously reported [1,7,9], but also of tibialis anterior although extensor digitalis and hallucis longus were not compromised (Table 1). This condition masks the de ciency of tibialis anterior, never mentioned in literature, and explains the characteristic morphology of the patients' feet during active and passive ankle dorsiexion that result decreased ( Figure 3, Table 1).
Moreover, this is the rst study that evaluate whole-body standing alignment by 3D-stereophotogrammetry and, at the same time, spinopelvic sagittal angles using x-Ray, in LOPD patients.
An original marker placement protocol (DB Total protocol, Figure 1, A), created ad hoc extending standard Helen Hayes M.M. protocol (Figure 1, A-blue markers) in order to study new speci c stereophotogrammetric parameters (Figure 1 B, C), was introduced to better analyse the postural abnormalities of LOPD patients. In particular, LOPD group showed a signi cant larger dorsal angle, indicating a attening of the dorsal curve, and a trend of lumbar hypolordosis, in disagreement to the literature that previously reported an excessive kyphosis and/or lordosis [7,8,10]. Moreover, a larger S2-C7 angle and a lower Sagittal Vertical Axis revealed a C7 position posterior to the sacrum; these ndings, associated to a larger dorsi exion ankle, exion knee, Heel-S2-C7 and Heel-S2-Nasion angles (Table 2), showed a typical sagittal upright posture of LOPD patient, as a result of a biomechanical compensation strategy of the spinal extensor muscles de cit [9,14,15].
Radiographic spino-pelvic results showed a signi cant lower Cobb Dorsal Angle (Table3) of Pompe group respect to normative values [16,17,18] con rming the reduction of dorsal kyphosis emerged in stereophotogrammetric analysis; moreover, a signi cant lower O-C2 and larger C2-C7 cervical angles were found in LOPD patients, underlining a straightening of the cervical lordosis.
No signi cative difference was found for Cobb Lumbar angle, but only a trend of hypo-lordosis in accordance with St-results. Finally, only a reduction trend of SVA was found compared to normal values.
Other studies [24] analysed sagittal spine parameters introducing a larger number of markers placed on the spine than standard protocols, but no correlation study was performed between stereophotogrammetric and radiological sagittal angles. Instead, in our study a strong correlation was found (Table 4), highlighting a high reliability of DBTotal protocol in matching markers to spinous processes of the reference vertebrae and in detection of dorsal angles; a lower correlation was found in lumbar angles probably due to the higher thickness of the adipose tissue present in this area.
The results of the present study should be evaluated and considered with respect to some limitations. The relatively small sample size of the populations recruited, due to the low incidence of the rare disease, prevents to any de nitive conclusion because of small statistical power. Investigation over larger population and/or meta-analysis are required. Nevertheless, the genetic homogeneity of the LOPD patients, the accuracy of our markers placement protocol demonstrated by x-Ray exam and the exclusive choice of sagittal stereophotogrammetric parameters [25,26], make our results reliable. Further studies, considering these quantitative outcome measures, would be needed to better  1  3  4  4  4  4  4  5  4  2  3  4  5  4  5  -10  0   2  2  2  4  4  4  4  5  4  3  4  5  5  3  5  5  10   3  2  2  5  3  3  3  4  3  1  4  5  5  4  4  -          Typical morphology of LOPD patients' feet at rest (a, c) and during active dorsi exion (b, d) on frontal (a, b) and sagittal (c, d) plane.