Validity of the Trunk Impairment Scale as a Measure of Trunk Performance in People With Parkinson’s Disease

Participants 

PD patients were screened when they visited the consultant neurologist at the University Hospital in Leuven, Belgium. Subjects were also recruited from the neurology ward, receiving inpatient care, as well as from local PD groups, attending outpatient rehabilitation. The consultant neurologist confirmed the diagnosis of PD and determined the Hoehn and Yahr stage. Only patients from Hoehn and Yahr stages I to IV were eligible for inclusion.

Exclusion criteria were evaluated by the observer (LO) after the people with PD agreed to participate in the study. The criteria included the presence of dyskinesia, defined as a grade of more than 2 on the Modified Dyskinesia Scale14; any comorbidity which could impair sitting balance or trunk movements; the presence of a subthalamic deep brain stimulator; a hip endoprosthesis; and dementia, defined by a score of less than 24 on the Mini-Mental State Examination.15

We selected age- and sex-matched healthy controls from an existing database. People with PD and healthy controls were informed about the aim and procedure of the study and were asked to voluntarily sign consent after having agreed to participate in the study. Ethics approval was obtained from the local ethics committee.

Assessments 

We collected age, sex, disease duration, and Hoehn and Yahr stage of the disease from the updated medical records for every PD patient when they visited the consultant neurologist.

All patients included in the study were assessed with part III of the Unified Parkinson’s Disease Rating Scale (UPDRS).16 This part of the UPDRS evaluates motor deficits of persons with PD by looking at the presence of rigidity, bradykinesia, tremor, and gait disorders. Items 12 (turning in bed and handling blanket [hereinafter, turning in bed]) and 13 (falling) of the UPDRS part II were evaluated for every participant as well, because they represented possible important aspects of functional trunk performance.16 The UPDRS is a widely used scale, used across the clinical spectrum of PD, and has established clinimetric properties, including reliability and validity.17 To detect differences in trunk impairment between freezers and nonfreezers, freezers were defined as freezing at least once a week (score >1 on item 3 of the Freezing of Gait Questionnaire (FOGQ), presently the only valid scale to evaluate freezing of gait).18 We also used the FOGQ total score to determine the severity of freezing.

We assessed trunk performance by means of the TIS. The TIS evaluates static sitting balance, dynamic sitting balance, and trunk coordination on a scale from 0 to 23 points, a higher score indicating a better performance.11 Patients’ starting position entailed sitting on the edge of a plinth without back and arm support. The thighs made full contact with the table. The feet were placed hip width apart and flat on the floor. The knee angle was 90°. The arms rested on the legs. Both head and trunk were in a midline position. Static sitting balance (score range, 0–7) evaluated if a subject could maintain a seated position, also with the legs crossed passively by the observer and actively by the subject. Dynamic sitting balance (score range, 0–10) assessed movement pattern and quality of movement when performing lateral flexion of the trunk, initiated from the upper and lower part of the trunk. Finally, coordination (score range, 0–6) scored the possibility to selectively rotate the shoulder and pelvic girdle.

All assessments were carried out by the same observer (LO), carefully instructed in administering the TIS. Applying the TIS was shown by the principal investigator (GV) by means of an instruction video and pilot session with feedback. The items of the UPDRS and FOGQ were administered by experienced PD researchers (A-MW, AN). All assessments were completed within 1 session.

Age and scores on the TIS and its subscales were collected from a healthy controls’ database of 40 subjects. Controls were selected consecutively based on the matching criteria for sex and age. Assessment of healthy controls was carried out by 2 observers who evaluated subjects independently. Both observers trained together in administering the TIS by means of an instruction video and pilot session with feedback by the principal investigator of this study (GV).

Statistical Analysis 

To detect differences in trunk performance between PD patients and healthy controls, we compared scores on the TIS and its subscales statistically using the Wilcoxon ranked-sum test. The unpaired Wilcoxon test was used to evaluate if there were differences in trunk performance between patients in the early stage of the disease (defined as Hoehn and Yahr stages 2 and 2.5) and in the late stage of the disease (defined as Hoehn and Yahr stages 3 and 4) and between freezers and nonfreezers. Spearman rank correlations were calculated and a univariate linear regression analysis was performed between TIS total score and age, sex, disease duration, Hoehn and Yahr stage, UPDRS part III score, score on the UPDRS part II items turning in bed and falling, and FOGQ score to examine the relation between trunk performance and other PD impairments. Significant parameters (P<.05) were retained and were entered as explanatory variables in a forward stepwise multiple regression analysis with the TIS total score as outcome variable to evaluate the combination of variables that determine trunk performance in PD patients. The level of significance for all analyses was P less than .05. Data were statistically analyzed by means of SAS, version 8.2,a and SAS Enterprise Guide, version 2.0.a

Participants 

We recruited 26 PD patients and 26 healthy controls. Our PD subjects included 17 men (65%) and 9 women (35%) with a mean ± standard deviation age of 65±9 years and a mean disease duration of 9±4 years. Descriptive characteristics for the PD patients are shown in table 1. Our sample of healthy controls included 16 men (62%) and 10 women (38%) with a mean age of 65±12 years.

Table 1. Descriptive Characteristics of the 26 PD Patients

	Parameter (Possible Range)	Median	Interquartile Range	Range
	Hoehn and Yahr stage (1–5)	2.5	2–3	2–4
	UPDRS part III score (0–108)	25	20–30	8–36
	UPDRS part II, turning in bed item score (0–4)	1	1–2	0–4
	UPDRS part II, falling item score (0–4)	0	0–1	0–4
	UPDRS part II, freezing item score (0–4)	1	0–2	0–4
	FOGQ score (0–24)	6	2–12	0–18
	Total TIS score (0–23)	18	16–20	11–23
	Static sitting balance score (0–7)	7	6–7	5–7
	Dynamic sitting balance score (0–10)	9	7–10	3–10
	Coordination score (0–6)	3	2–4	1–6

Seven PD patients were in Hoehn and Yahr stage 2, 7 in stage 2.5, 11 in stage 3, and 1 in stage 4. Scores on the turning in bed item of the UPDRS part II showed that 6 patients had no difficulty during turning in bed. Twelve subjects were slow and clumsy but independent. Five persons were able to turn in bed only with great effort and 3 subjects were able to initiate the movement but could not conclude it without assistance. Nineteen patients had no history of falls. Three subjects fell rarely, 2 subjects indicated to fall less than once a day, and 2 patients fell more than once a day. According to our definition, the sample consisted of 10 freezers and 16 nonfreezers. All patients were evaluated in the on-phase of the medication cycle.

Trunk Performance in PD 

The results of the Wilcoxon rank-sum test, evaluating differences in trunk performance between PD patients and healthy controls, are shown in table 2. Patients with PD had significantly lower scores on the total TIS (P<.000), static sitting balance subscale (P=.005), and coordination subscale of the TIS (P<.000). Although the PD patients obtained relatively lower scores on the dynamic sitting balance subscale of the TIS in comparison with the healthy controls, the difference did not reach the level of significance (P=.053).

Table 2. Comparison of Trunk Performance Between PD Patients and Healthy Controls

	Parameter	PD Patients Median (IQR)	Healthy Controls Median (IQR)	Wilcoxon U Statistic	P
	Total TIS (0–23)	18(16–20)	22(21–23)	23.60	<.000
	Static sitting balance (0–7)	7(6–7)	7(7–7)	7.91	.005
	Dynamic sitting balance (0–10)	9(7–10)	10(9–10)	3.74	.053
	Coordination (0–6)	3(2–4)	6(5–6)	30.88	<.000

Abbreviation: IQR, interquartile range.

Further analysis revealed that healthy controls scored significantly better on the total TIS (Mann-Whitney U=12.91, P<.000) and coordination subscale (U=22.51, P<.000) compared with the PD patients in the early stage of the disease (Hoehn and Yahr stages 2 and 2.5). Patients with PD in the early stage scored significantly higher for the total TIS (U=4.19, P=.047), static sitting balance subscale (U=5.89, P=.029), and dynamic sitting balance (U=4.79, P=.033) in comparison with the PD patients in Hoehn and Yahr stages 3 and 4. We found no significant differences for the scores on the total TIS and its subscales between freezers and nonfreezers (U=−0.7, P=.484).

No PD patient obtained the lowest score on any of the subscales or on the total TIS. From our sample of patients, 19 (73%) reached the maximum score on the static sitting balance subscale, 10 (38%) on the dynamic sitting balance subscale, 2 (8%) on the coordination subscale, and 1 patient (4%) reached the maximum score on the total TIS.

Determinants of Trunk Performance in PD 

Table 3 shows the Spearman correlation coefficients and the results of the univariate linear regression analysis between TIS total score and age, sex, disease duration, Hoehn and Yahr stage, UPDRS part III score, UPDRS part II scores on items turning in bed and falling, and FOGQ score for the PD group.

Table 3. Correlation and Linear Regression Analysis Between TIS Total Score and 8 Demographic and Clinical Parameters of 26 PD Patients

	Variable	r	R2	P
	Age	−.66⁎	.32	.003
	Sex	−.32	.10	.117
	Disease duration	−.23	.03	.408
	Hoehn and Yahr stage	−.39	.09	.130
	UPDRS part III	−.68⁎	.54	<.001
	UPDRS part II, turning in bed	−.48⁎	.15	.049
	UPDRS part II, falling	−.41⁎	.07	.182
	FOGQ	−.27	.09	.152

NOTE. Values are Spearman rank correlation coefficients (r), explained variance as a result of the univariate regression analysis (R2), and accompanying P value.

⁎ P value for the Spearman rank correlation coefficients was less than .05.

Our analysis showed that lower scores on the total TIS correlated significantly with older age and a higher score on part III of the UPDRS and the turning in bed and falling items of part II of the UPDRS, although the latter item did not show a significant result in the univariate analysis. Therefore the falling item was not retained for the forward stepwise multiple linear regression analysis. This analysis showed that trunk performance in PD patients correlated significantly with a combination of part III score of the UPDRS (partial R2=.54, P<.000) and age (partial R2=.09, P=.030).

Study Limitations 

A limitation of this study was that the PD patients and healthy controls were not exactly matched for sex and age. Control subjects were selected retrospectively from an existing database. The first matching criterion that we used was sex; the second criteria was age. Future studies could ask the participation of the partner of the PD patient as a control subject. This could also control for characteristics such as level of activity between both groups. Another limitation of this study was that a reliability study of the TIS in PD was not carried out. However, extensive psychometric analysis has been performed for the TIS in stroke patients, persons with multiple sclerosis and subjects after traumatic brain injury, in which reliability, validity, internal consistency, and measurement error were established.11, 12, 13 Because the TIS seems a useful tool for PD patients on the basis of the present findings, an evaluation of its reliability, measurement error, predictive validity, and responsiveness needs to be undertaken. Our results point to various trunk impairments in different stages of PD that might be amenable to rehabilitation intervention. The effectiveness of certain aspects of trunk rehabilitation for PD patients has already been investigated.22, 23, 24 A specific standardized and responsive clinical tool for measuring trunk performance would contribute to setting up further experimental studies in this domain.