Gait disturbance(s) after suffering a stroke may significantly affect one's daily activities. Thus, assessment of a patient's gait ability immediately after a stroke and establishment of a treatment plan are most important before beginning rehabilitation. Gait ability is closely associated not only with muscle strength of upper and lower extremities of the clinically affected side, but also with balance control. Balance impairment may cause not only an extension of one's hospitalization after stroke, but also a deterioration of the motility and functional level [1]. Many studies have reported that balance control was a significant clinical marker for predicting overall prognosis of patients at initial stages of stroke [2,3,4]. Indeed, although standing posture is impossible in most patients at the initial stage of stroke, most previous studies on balance of hemiplegic patients after stroke mainly focused on standing balance. The Berg Balance Scale, which is currently widely used to assess balance and gait ability of patients with stroke, consists of items closely related to daily activities; it has also been proven to be closely associated with prognosis of patients with stroke. However, despite its usefulness, limitations, such as low sensitivity, inadequate items for assessing sitting balance, and the ceiling effect, have also been reported [5,6,7].

Many researchers suggested development of an assessment method to objectify sitting balance via clinical assessment. However, many methods showed a methodological limitation by simply reflecting balance maintenance time rather than functional assessment [8]. Verheyden et al. [9] presented the concept of trunk performance, which is closely associated with actual activities, including sitting balance; they developed the Trunk Impairment Scale (TIS) to objectively assess trunk performance objectively. They validated reliability and internal consistency of TIS and confirmed the correlation of TIS with sitting balance, gait, and functional ability in patients with stroke [10,11]. The TIS provides a convenient assessment and is more advantageous than previous assessment tools as it can provide a more classified assessment due to a less pronounced ceiling effect [9]. Reports have indicated that the assessment of patient's trunk performance via TIS reflected the functional level of patients with stroke in the chronic stage better than other typically-used variables, such as early Korean version of Modified Barthel Index (K-MBI), the patient's age, and the time required for hospitalization, which are known to be correlated with the functional prognosis of patients with stroke [9,12,13]. However, these studies had numerous limitations. The study was conducted on some patients with stroke in the chronic stages, the patients did receive continuous follow-up. Furthermore, although the clinical usefulness of trunk performance may vary depending on the condition of patients with stroke, and in particular, the clinical importance of trunk performance may vary depending on the gait ability at the initial stage of stroke, TIS was assessed without consideration to patients' condition. In addition, only a few studies have been conducted for TIS as a predictor of functional prognosis after stroke.

This study aimed to investigate the relationship between initial trunk performance and functional outcomes according to gait ability, and whether initial trunk performance is of predictive value in terms of functional prognosis in patients with stroke.

This study aimed to investigate the clinical importance of TIS according to gait ability in patients with stroke, as well as to confirm the relationship of initial TIS and functional prognosis of rehabilitation. Prognosis prediction via clinical assessment has been investigated by myriad researchers, thanks to easy conduct without a particular device [17,18,19]. Many studies have reported that a patient's sitting balance in initial stage after stroke was closely related to the functional level of patients. These studies emphasized that sitting balance should be the focus of clinical examination and statistical analysis. Recently, new methods have been offered to measure and assess sitting balance: the trunk control test (TCT), Postural Assessment Scale for Stroke-trunk control (PASS-TC), and TIS have been the most commonly used instruments as tools to assess trunk balance performance. TCT assesses a patient's performance in both lying and sitting posture; both have been validated to be related to the recovery and functional level of the muscle strength of the upper and lower extremities. However, TCT poorly reflects the quality of movement, and is weakly related to the actual movement of the trunk muscle. Furthermore, the 'ceiling effect' cannot be ruled out as TCT consists of items easily conducted by patients with stroke [1,2]. PASS-TC was developed by selecting five items of trunk balance maintenance. The instrument was validated to be used as a tool to assess patient's performance in daily activities. However, PASS-TC poorly reflects the functional level of chronic patients with stroke, and the ceiling effect cannot be ruled out.

Verheyden et al. [9] suggested that the conventional assessment methods poorly reflected trunk performance, which is actually used in daily activities, although trunk performance after stroke significantly affects daily activities, suggesting the TIS as a new assessment method. And it is suggested that each TIS subscale as well as total TIS had sufficient reliability and validity for use in clinical practice.

In this study, initial TIS and functional outcomes at 4 weeks and 6 months after stroke were significantly higher in ambulatory patients than in non-ambulatory patients. However, initial TIS assessed in ambulatory patients was no significant correlation with K-MBI at 6 months after stroke. It was suggested that trunk performance was not integral factor for functional outcomes if patients' gait ability was preserved after stroke. In addition, of the TIS subscales, TIS-S showed no significant correlation with K-BMI at 4 weeks after stroke. Although Verheyden et al. [9] suggested TIS as a tool with no ceiling effect, the results of this study showed that ceiling effect was not ruled out in the assessment of TIS-S in ambulatory patients, with 34 patients (62.97%) scoring 7 (maximum score) and 16 patients (29.63%) scoring 6. This is likely to be attributable to the fact that total score of initial TIS was not correlated with K-MBI at 6 month after stroke. However, the additional studies is needed to support this result and clinical implication of trunk performance in ambulatory patients after stroke should be confirmed by means of various objective measurements, such as isokinetic muscle testing, EMG analysis or motion analysis.

To investigate the relationship of TIS with performance in daily activities, K-MBI subitems were analyzed for their correlation with initial TIS. In previous studies, it had been suggested that sitting balance is a crucial component to perform ADL and it might be because K-MBI assess ADL, which requires sitting balance for feeding and dressing. However, as a result of this study, all of K-MBI subitems had a significant correlation with initial TIS, such as bowel and bladder control, which is not related with trunk movement, and this relationship was remained over the time course. That is trunk performance have relevance to general performance of ADL as well as a specific motion. Of the TIS subscales, TIS-C showed a significant correlation with only 5 K-MBI subitems for all the time points measured and these subitems which had a significant correlation for K-MBI had little in common feature. It suggested that floor effect could not exclude in TIS-C of non-ambulatory patients, with 36 patients (44.45%) scoring 0 or scoring 1. This should be confirmed in future studies.

Of the TIS subscales, TIS-S evaluates if a patient can maintain an upright sitting position with and without the legs crossed. Because of ADL are performed from a seated position, static sitting balance could be seen as a prerequisite condition. In previous study, it was considered as a significant predictor for functional outcome [12]. However, in our study, even if participants were limited to non-ambulatory patients after stroke, TIS-D was the most important factor when predicting K-MBI and FAC at 4 weeks and 6 months after stroke. TIS-D examines the ability to perform lateral flexion of trunk and reflected dynamic trunk movement. Eventually it might be more essential for performing ADL than static trunk control and supports the importance of active trunk rehabilitation.

This study had the following limitations. First, our study was conducted retrospectively, using the medical records of subject patients. For this reason, bias in patient selection cannot be ruled out. Second, limited assessment tools were used to assess the patients' functional independence and gait level. Although K-MBI and FAC have been widely used in patients with stroke, and their reliability and eligibility have already been validated [8], other assessment tools, such as the trunk control test, are required for comparison to show the usefulness of TIS. Nonetheless, the results of our study again confirmed the clinical usefulness of TIS, particularly according to gait ability after stroke. In addition, this study is meaningful as it was conducted to investigate which is considered as predictor of ADL and gait function for the establishment of a better rehabilitation program.

In conclusion, the strong relationship between trunk performance and functional outcomes in patients with stroke emphasizes the importance of trunk rehabilitation. In particular, clinicians should evaluate initial TIS after stroke, especially TIS-D; this can be used as a useful predictor of performance in functional prognosis in patients with stroke. However, the application of TIS in clinical practice and stroke research is needed consideration on patient's gait ability and severity of trunk impairment.