Correlation between qualitative balance indices, dynamic posturography and structural brain imaging in patients with progressive supranuclear palsy and its subtypes
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.185417
Source of Support: None, Conflict of Interest: None
Objectives: To compare the clinical, balance, and radiological profile of progressive supranuclear palsy (PSP) of Richardson type (PSP-R) and Parkinsonian type (PSP-P).
Keywords: Balance; Berg balance scale; dynamic posturography; imaging; magnetic resonance imaging; progressive supranuclear palsy; Richardson's type; Tinetti performance-oriented mobility assessment scale
Progressive supranuclear palsy (PSP) is the second most common neurodegenerative movement disorder after Parkinson's disease (PD)., Recent studies have subclassified PSP into different clinical phenotypes.,, The two most common PSP subtypes are Richardson's syndrome (PSP-R) and PSP- Parkinsonism More Details (PSP-P).,, Of the subtypes, PSP-R is the most common subtype, having the typical clinical features of PSP.,, Dynamic posturography (DP) is a method to quantitatively assess the balance abnormalities in different directions. It detects dynamic balance as well as directional balance of postural reflexes and may help in differentiation of early PSP from early PD. In addition, there are various studies on advanced neuroimaging methods in different types of Parkinsonism, which have attempted to differentiate the clinical phenotypes of PSP., However, there is paucity of studies which have attempted to differentiate the various subgroups of PSP based on balance abnormalities.
This study was done to investigate the balance and radiological profile of patients with PSP and its subtypes. Patients with PSP were prospectively assessed using balance scales,,, DP  and high-resolution magnetic resonance imaging (MRI) (3T). To the best of our knowledge, this is the first study to correlate DP with high-resolution MRI imaging in patients with PSP.
Consecutive patients satisfying the inclusion and exclusion criteria of probable and possible PSP as per the National Institute of Neurological Disorders and Stroke-SPSP criteria  were recruited from the Neurology Department of the National Institute of Mental Health and Neurosciences (NIMHANS), Bangalore, a tertiary referral center in Southern India. Only those PSP patients who were ambulant and were able to undergo DP were recruited. Patients with PSP were further subclassified into PSP-R and PSP-P based on the asymmetric Parkinsonism, tremors at onset, limb rigidity, and levodopa responsiveness.
The controls were age - and gender-matched healthy subjects who were caregivers or family members of patients and gave consent to participate in the study. They were neurologically examined with emphasis on detection of abnormalities of eye movements, impaired balance and gait, and extrapyramidal signs.
The study was approved by the Institute Ethics Committee and written informed consent was taken from each subject.
All patients underwent the following clinical assessments: (i) Unified PD rating scale (UPDRS) part III,, (ii) PSP rating scale (PSPRS), (iii) Berg balance scale (BBS),, and (iv) Tinetti performance-oriented mobility assessment (POMA) scale with its subcomponent scores: Tinetti POMA balance (TPB), Tinetti POMA gait (TPG), Tinetti POMA total (TPT). The details of these scales are given in [Supplementary Table 1].[Additional file 1]
The patients and controls underwent DP (Biodex Balance System, Biodex, USA), and MRI of brain in a 3-Tesla scanner (Philips, Achieva Scanner).
Dynamic posturography (Biodex Balance System)
This system consists of a suspended computerized circular platform that can tilt 20° in all directions from the horizontal. The system's microprocessor-based actuator controls the extent of the surface instability of the platform. The surface instability can be adjusted from Level 8 (most stable) to Level 1 (least stable). Level 8 was used for the examination of balance in all the subjects.
The subject's hand and leg dominance was assessed through Edinburg inventory. All the subjects were explained and oriented to the test in training mode before the actual test. The subjects were tested for (a) dynamic stability, which included the ability to control balance in all directions (overall balance index [OBI]), front to back (anterior–posterior index [API]), and side-to-side (mediolateral index [MLI]); and, (b) the limits of stability (LOS) in all the directions. The degree of impairment of balance indices was obtained from the difference between the actual and predictive values.
LOS is defined as the area over which a subject can safely move without changing the base of support towards the target, which appears randomly in eight directions only once. The DP instrument (Biodex, USA) is designed to test the LOS by displaying a blinking target on a screen placed in front of the subject. The subjects are instructed to move their centers of mass toward the target, without changing the foot position. In the present study, LOS in eight directions are represented by forward (FW), backward (BW), right (RT), left (LT), FW-RT, FW-LT, BW-RT, and BW-LT vectors.
Higher scores in the UPDRS III, PSPRS, and balance indices are indicative of a poor balance whereas the higher scores are representative of better stability with the BBS, Tinetti POMA and LOS scales.
Magnetic resonance imaging of the brain
The MRI sequences included T1-weighted, T2-weighted (T2W), fluid-attenuated inversion recovery, and T1 MPRAGE sequences. Qualitative analysis of imaging included the identification of the specific signs, namely, the “Humming bird' sign (HBS) or the “Penguin silhouette” sign, the “Morning glory” sign (MGS), and generalized cortical atrophy (GCA).,,,, The quantitative analysis of MRI included measurements of the AP diameter of midbrain in the axial plane, the area of midbrain, the area of pons, and the ratio of the area of midbrain to the area of pons, as per Oba et al.,,,,,
SPSS 16 (SPSS for Windows, Version 16.0. IBM, Chicago) was used to analyze the data. Student's t-test [two-tailed, independent] was used to find the significance of the study parameters on a continuous scale between the two groups. Chi-square/Fisher exact test was used to determine the significance of the study parameters on categorical scale between two or more groups. Pearson correlation was determined between the clinical severity scores with radiological, balance and clinical variables. A P value of < 0.05 was considered statistically significant.
At the end of the study period, 29 patients with PSP were evaluated. There were 17 cases of the PSP-R subtype with a mean age of 62.5 ± 6.6 years and 12 cases of PSP-P subtype with a mean age of 59.0 ± 8.9 years (P = 0.06). The distribution of men and women between the two types of PSP was gender matched. The age at onset of symptoms (P = 0.05) and the age at falls (P = 0.09) were considerably lower in the case of PSP-P patients relative to the PSP-R patients. The demographic characteristics of the patients as well as controls are represented in [Table 1]. The levodopa responsiveness was better in the PSP-P subtype as compared to the PSP-R subtype; however, it did not reach statistical significance [Table 2]. Thirty healthy controls were enrolled into the study.
Qualitative balance scores
Among the clinical balance scales used, the PSPRS, TPG, and TPT scores were found to be statistically significant among the subtypes of PSP (P = 0.045; P = 0.031; and, P = 0.037) while UPDRS-III was not significant (P = 0.7). The PSP-R subtype performed poorly in comparison to the PSP-P subtype on these scales [Table 2].
Dynamic posturography (quantitative balance evaluation)
The mean values of balance indices were almost similar between the subtypes of PSP as compared to controls. The most significant of all the parameters in DP was LOS (P < 0.001) and the PSP-R subtype had lower scores [Table 3]. Significant correlations between the progressive supranuclear palsy rating scale and the qualitative and quantitative balance parameters were noted [Supplementary Table 2].[Additional file 2]
Overall progressive supranuclear palsy
There was a significant correlation of PSPRS with BBS (r = −0.642, P < 0.001), TPT (r = −0.516, P = 0.004), TPG (r = −0.449, P = 0.013), and TPB (r = −0.505, P = 0.004). LOS-BW-LT had significant positive correlation with BBS (r = 0.381, P = 0.038), TPB (r = 0.417, P = 0.022), and TPT scores (r = 0.362, P = 0.049).
Progressive supranuclear palsy of Richardson type
API had significant negative correlation with BBS (r = −0.589; P = 0.044) and TPT (r = −0.0562; P = 0.057) scores.
Progressive supranuclear palsy-Parkinsonian type
All the quantitative indices such as OBI (r = −0.79; P = 0.002), API (r = −0.589; P = 0.044), MLI (r = −0.604; P = 0.037), and LOS-BW-LT (r = 0.610; P = 0.035) had significant correlation with BBS. LOS-BW-LT also had significant correlation with TPB (r = 0.599; P = 0.039) and TPT (r = 0.57; P = 0.053) scores. There was also significant negative correlations of OBI with TPG (r = −0.772, P = 0.003) and TPT (r = −0.683, P = 0.013) and MLI with TPG (r = −0.688, P = 0.013).
The PSP-R subtype, compared to the PSP-P subtype, had more often radiological signs such as HBS (P < 0.001), MGS (P < 0.008), and GCA (P < 0.001). These signs were absent in controls, except for MGS, which was present in 23.3% of controls [Table 4]. The most statistically significant quantitative radiological parameter was the area of the midbrain (P < 0.002) and the ratio of midbrain to pons (P < 0.013). These parameters were worse in the PSP-R patients, better in the PSP-P patients, and normal in controls [Table 4].
Significant correlations among radiological parameters and quantitative, qualitative, and clinical rating scales in progressive supranuclear palsy and its subtypes
Overall progressive supranuclear palsy
There was a significant correlation of the midbrain axial AP diameter (r = 0.4; P = 0.03) and the ratio of midbrain to pons with BBS (r = 0.4; P = 0.02), indicating that these are worse in patients with midbrain atrophy. In DP, API correlated negatively with the midbrain axial AP diameter (r = −0.5; P = 0.01) and midbrain area (−0.39; P = 0.03). The LOS-BW correlated positively with the area of midbrain (r = 0.49; P = 0.001) and the midbrain to pons ratio (r = 0.57; r = 0.001). In addition, LOS-BW-LT correlated with the midbrain to pons ratio (r = 0.37; P = 0.43) [Supplementary Table 3].[Additional file 3]
Progressive supranuclear palsy-Parkinsonian subtype
There was a significant negative correlation between the midbrain axial AP diameter and the DP measures (OBI: r = −0.6; P = 0.024 and MLI: r = −0.6, P = 0.04), suggesting that lower the diameter, higher the balance abnormality. There was also a positive correlation with qualitative balance scores (TPG: r = 0.7, P = 0.006 and TPT: r = 0.7, P = 0.018) again indicating worsening of balance scores with lower midbrain axial AP diameter [Supplementary Table 2].
Progressive supranuclear palsy-Richardson subtype
The API values negatively correlated with the midbrain axial AP diameter showing that the midbrain atrophy was associated with abnormally high API scores (r = −0.5, P = 0.039). The LOS-BW score positively correlated with the area of midbrain and midbrain to pons ratio (both having r = 0.8 and P < 0.001) and the LOS-BW-LT score also positively correlated with the area of midbrain (r = 0.6, P = 0.012) and the midbrain to pons ratio (r = 0.7, P = 0.001). These indicate that worse balance scores (higher scores) were associated with disproportionately increased midbrain atrophy.
The cut-off value of the AP diameter of the midbrain in T2W MRI was 13.63 mm, which had a sensitivity of 70% and a specificity of 96.7%. The cutoff value of midbrain area was 129 mm 2 with a sensitivity and specificity of 90% and 93%, respectively. The cut-off ratio of the area of midbrain to pons in our study was 0.23 with a sensitivity of 90% and a specificity of 86.7%.
The progressive balance abnormality in PSP is due to the early involvement of the brainstem structures. There are only a few studies that have investigated the relation of balance parameters (BBS, Tinetti POMA test, and DP indices) in PSP and its subtypes, with the other known clinical and radiological markers. In this study, we assessed the clinical, balance, and radiological parameters in patients with PSP and its subtypes and looked for any correlations among these parameters.
In our study, BBS was applied to all patients and as expected, all the patients scored lower values on BBS implying the greatest risk of falls. BBS was lower in the PSP-R subgroup compared to the PSP-P subgroup; however, there was no statistically significant difference in the response to levodopa.
BBS assesses only the static balance and does not include the gait measurement. Therefore, we used the Tinetti POMA test for assessing balance and gait. All the patients scored 18 indicating that they faced an extremely high risk of falls. Once again, there was no improvement following levodopa administration. The PSP-R subgroup had lower scores compared to the PSP-P subgroup, with no significant response to levodopa in either group. The lower response to levodopa was possibly due to a smaller sample size.
Quantitative balance parameters in the major groups by dynamic posturography
The mean values of the balance indices (OBI, API, and MLI) were similar in the PSP and control groups implying that in the early stages of PSP, these balance indices may not be sensitive measurements. In PSP, the mean overall as well as the component LOS scores in all the seven directions was significantly lower than in the controls. The scores, especially the overall status, as well as in the FW, FW-LT, BW, BW-RT, and BW-LT directions were significantly lower compared to controls. These findings suggest that the patients had significant difficulty in shifting their weight toward an unanticipated directional change even with visual cues in the respective directions. The overall LOS and LOS in the BW direction are very significant and are practically in close concordance with the clinical observation (historical as well as using the pull test) wherein it is observed that most of the patients of PSP lose balance in the BW direction. Hence, LOS is the most sensitive measure of change of balance in the early stages of PSP.
The only LOS score which was found to be statistically significant between the two subtypes was the LOS-RT, which implies a significant difficulty in shifting weight toward the right-sided direction in the R subtype. However, this isolated observation needs to be validated in a larger number of patients.
In our study, HBS was seen in 80% of the patients due to the disproportionate atrophy of the rostral midbrain. It was mainly seen in the mid-sagittal plane. In their study, Kato et al., noted the HBS sign in all the eight patients of PSP while it was not seen in any of the PD patients or controls. Adachi et al., defined the HBS sign as the concavity of lateral tegmental margin of the midbrain that could be demonstrable in 4 out of the 5 patients (80%) of PSP. They correlated the presence of HBS sign with vertical supranuclear palsy.
In our study, we found MGS to be present in 60% of PSP patients and 23.3% of healthy controls. The presence of MGS in our healthy controls could be the result of generalized atrophy as described by Mori et al. Mori et al., evaluated eight consecutive cases of PSP and concluded that MGS had a low sensitivity and also added that HBS was superior to MGS. In our study, although HBS was absent in all the controls, 23.3% of the controls had MGS. This observation suggests that HBS may be more specific than MGS in aiding in the diagnosis of PSP. In the past, the sensitivity of the HBS has been reported to range from 55% to 100%.,,
HBS and MGS were relatively more frequent in the R subtype (94%) compared to the PSP-P subtype (66%). This finding is in agreement with the previous radiological study by Agosta et al., in which the authors compared the PSP-R patients with the PSP-P subgroup and found a higher level of white matter atrophy at the level of midbrain in the former group.
In the current study, the mean axial diameter of the midbrain in the PSP patients was 12.7 mm in contrast to 15.1 mm in the controls (P < 0.001). The mean diameters were even lower in the PSP-R subtype compared to the PSP-P subtype (11.9 mm vs. 16.4 mm) [P < 0.001]. The cut-off value of the AP diameter of midbrain in T2W MRI was ≤13.63, which had a sensitivity of 70% and a specificity of 96.7%. Schrag et al., Warmuth-Metz et al., and Asato et al., have used different cut-off points and found that the sensitivity increases using a lower cut-off value.,,,
The mean midbrain area in our study was 93 mm 2 in PSP patients compared to 147 mm 2 in the controls (P < 0.001). Slowinski et al., reported the mean midbrain area of 99 mm 2 in the PSP patients as compared with 141 mm 2 in the control group. Oba et al., reported the mean midbrain area of 56 mm 2 in PSP patients and 117 mm 2 in controls.
In our study, the highest sensitivity was noted for the area of midbrain and the ratio of midbrain to pons, followed by the axial diameter of the midbrain. We found the range of midbrain to pons ratio in PSP was 0.18 and 0.29 in controls (P < 0.001). Slowinski et al., reported similar findings with the ratio of 0.2 and 0.25 in the PSP and controls, respectively. The ratio of midbrain to pons of 0.18 in our study showed a diagnostic sensitivity of 90% and a specificity of 86.7% in helping to differentiate PSP patients from controls. Cosottini et al., reported an 86% sensitivity and a 100% specificity of the midbrain to pons ratio in differentiating PSP from PD and multiple system atrophy-Parkinsonism (MSA-P). However, Quattrone et al., did not find this ratio to be useful for the differentiation of PSP from PD and MSA-P. The cut-off values of the ratio of area of midbrain to pons in our study was ≤ 0.23 with a sensitivity of 90% and a specificity of 86.7%. Oba et al., measured the sensitivity and specificity of the ratio of area of midbrain to pons (cut-off value of 0.12) to the extent of 100%. On subtype analysis, the PSP-R patients were noted to have even lesser values of ratio of midbrain to pons relative to the PSP-P patients (0.15 mm vs. 0.21 mm).
Correlation of radiological parameters with balance scales
In the overall PSP group, the midbrain to pons ratio and the area of midbrain positively correlated with the BBS score. In the PSP-P subgroup, the midbrain axial AP diameter showed a significant positive correlation with the TPG and TPT scores. However, in the PSP-R group, no such correlations were observed. These correlations suggest that either the BBS or some of the subscores of Tinetti POMA may reflect the true disease severity associated with the midbrain atrophy, at least in the PSP-P group.
It was observed that the most important correlations of the midbrain atrophy were found in the overall PSP group, and to a lesser extent, in the PSP-P group, and least in the PSP-R group. This suggests that the qualitative and quantitative measurements of severity of balance in PSP-P are better markers of the predominant pathology (evident as atrophy on MRI) in the midbrain. On the other hand, apart from the midbrain, other areas of brain may also be affected in the PSP-R subtype.
To conclude, the clinical severity scale (PSPRS) correlated well with BBS and TPT only in the overall PSP group; however, PSPRS correlated with LOS in all the three groups. PSPRS did not correlate with the imaging in any of the three groups. The DP score correlated well with midbrain atrophy in the overall PSP group and the subtypes of PSP. The qualitative balance scale abnormality was an indicator of midbrain atrophy only in the overall PSP group (with BBS) and the PSP-P group (with TPT) but not in the PSP-R group. We found that the qualitative and quantitative scales for balance correlated well in the PSP overall group and in PSP-P group but not in PSP-R group.
Thus, our study shows that the measurements of balance severity in the PSP-P group correlate with the predominant pathology of the midbrain (midbrain atrophy); while in PSP-R subtype, the balance abnormalities could be a result of pathology in different or overlapping areas. Further studies should be directed to correlate the qualitative and quantitative balance abnormalities with serial imaging of the brain in larger cohorts of patients of each subtype of PSP.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2], [Table 3], [Table 4]