The leading question for this investigation was: “How is posturography used in the assessment of static balance in pwPD and atypical parkinsonism?”.

A scientific librarian developed comprehensive and systematic searches. We did not use the PICO framework because neither the Intervention nor the Comparison components could be defined for most studies included in this review, which does not focus on evaluating the effectiveness of a specific intervention on a particular outcome. Therefore, the search strings were developed through an iterative process. The research was conducted in January 2024 within the following databases: Medline, Cinahl, Embase, and Scopus. Furthermore, a manual cross-reference search was performed on the reference lists of included articles. No time limits were set for publications to be included.

The complete search strategies can be consulted in Table 1.

The inclusion criteria were: (a) studies involving adult pwPD or other types of atypical parkinsonism; (b) studies investigating pwPD using static posturography, with any protocol that included barefoot bipedal upright stance acquired with eyes open on a firm surface, which will be from now on considered the “baseline condition” (along with the baseline condition, protocols could then involve other conditions, e.g., open/closed eyes, feet apart/together, firm/foam surface, cognitive tasks); (c) primary peer-reviewed studies (e.g., RCT, clinical study, observational study); (d) full text available in English.

The exclusion criteria were: (a) studies including only healthy individuals or patients with other pathologies; (b) studies not involving humans (i.e., modeling studies); (c) acquisition protocol based on dynamic posturography only (e.g., Sensory Organization Test, Functional Reach Test); (d) sessions acquired while wearing goggles or similar headsets; (e) studies in which patients were provided onscreen feedback of the position of their center of pressure to stabilize during the assessment; (f) patients holding their arms away from the body so that the center of pressure is shifted (e.g., Romberg position).

We were interested in all outcome measures derived from static posturography acquired in the baseline condition.

After all databases were searched, reports were exported to EndNote20 (Clarivate Analytics, PA, USA), and duplicates were removed. The remaining studies were imported to Rayyan (26) online software. Two independent reviewers (LC and MCB) blindly screened their title, abstract, and full text. Discrepancies were discussed in a consensus meeting. If agreement could not be reached, a third researcher was called upon to solve any discrepancy (AM).

Two reviewers (MCB and LC) independently extracted the data. When the computation methods of the parameters were unclear, a third author (AM), a bioengineer with 20 years of experience in instrumental motion and posture analysis, was consulted. When necessary, the authors were contacted by email to obtain missing data from the reports, and information was added if they replied in 1 month.

The following information was summarized in a custom table: first author, year of publication, journal, aims of the study, sample size, patients' demographic and clinical characteristics, any treatment undergone by patients, protocol adopted for the posturographic assessment (including different conditions tested, number and duration of repetitions, and dosage of any drugs or brain stimulation), task conditions (including foot positioning, gaze fixation, and arm positions), technical details of the instrumentations and data processing (e.g., filters applied), any posturographic parameter derived from baseline posture. Parameters (means and standard deviations, median, or ranges) were retrieved from text, tables, or obtained from figures when necessary. Findings were also presented in a narrative synthesis, grouped by the five domains illustrated in the following paragraph.

The three reviewers involved in the previous steps critically appraised the included studies. In case of doubts about the clinical characteristics of the sample, a neurologist with extensive expertise in pwPD was consulted.

To standardize the quality assessment procedure, a tailored set of quality questions was designed, tested on a random sample of 30 studies, and refined progressively until its final version, reported in Table 2.

The questions, listed according to the traditional structure of the paper, explore five domains:

Each question was scored on a three-level basis: 1 for yes, 0.5 for limited details, and 0 for no. For some items, the score Not Applicable (NA) was added. The total score was calculated for each study to evaluate its overall quality. The total score of each question—and, consequently, of each category—was then expressed as a percentage of the maximum achievable score (i.e., considering only the assessable studies), not to penalize studies that received an NA score for some items.

Finally, studies were categorized into three groups: high-quality (total score exceeding 80%), medium-quality (total score ranging between 51 and 79%), and low-quality (total score below 50%) as in Samadi Kohnehshahri et al. (22). We also separately evaluated the quality of each category, determined by the ratio of question scores at each level (low, medium, high) and the total number of questions linked to that category.

We also investigated whether studies published in strictly clinical or mixed (i.e., clinical and bioengineering, bioengineering only, neurophysiology) journals received different scores in the categories of critical appraisal. We labeled the journal as clinical or mixed based on the main “subject area and category” on the Scimago Journal and Country Rank portal (20). When journals were not indexed on Scimago or classified as “multidisciplinary” (e.g., PlosOne), classification was carried out by examining the journal's aims, the study itself, and the authors' affiliations. Critical appraisal analysis was then performed for these two subgroups.

The comprehensive table containing all information extracted from the included studies can be found in Supplementary material 1.

Of the 132 studies, 115 focused on pwPD and 17 on atypical parkinsonism (MSA, PSP). Ninety-five were observational studies, and 37 were interventional studies. The former were generally cross-sectional studies aimed at describing the features of balance management of the included sample. The latter mainly evaluated a drug's or treatment's effectiveness on patients' postural ability.

Overall, 4,262 patients, aged 43–91, were assessed with static posturography in baseline condition.

The results obtained from the scores of each quality question are reported in Supplementary material 2. Of the 132 studies, 17 (13%) were rated high-quality, 105 (79%) medium-quality, and 10 (8%) low-quality (Figure 2A). The category “clinical aspects” received high-quality scores in half of the studies (see Figure 2B). The “technical aspects” and “study methodology” categories received mostly medium-quality ratings. Finally, the two categories with the lowest scores were “transferability to clinical practice” and “assessment protocol”.

Table 3 shows the results of each item from the five domains, detailing the average rating scores (see Supplementary material 2 for details of individual studies). The total score of each question was calculated as a percentage, excluding studies assessed as “NA” to avoid penalizing those questions. Consequently, some ratios have denominators lower than the total number of included studies (e.g., 74/128).

Of the included studies, 102/132 were strictly clinical, and 30/132 were mixed studies.

Of the 102 clinical studies, 13 (13%), were rated high-quality, 81 (79%) medium-quality, and 8 (8%) low-quality. Of the 30 mixed studies, 4 (13%) were rated high-quality, 24 (80%) medium-quality, and 2 (7%) low-quality.

Regarding clinical studies, the domains with the highest scores were “clinical aspects” and “assessment protocol” (see Figure 3A). The categories with the lowest scores were “transferability to clinical practice” and “technical aspects”. Thirteen clinical studies were cumulatively classified as high-quality. Only one study received high-quality ratings for all five domains (35).

Regarding mixed studies, the domain with the highest score was “technical aspects” (see Figure 3B). The categories with the lowest scores were “clinical aspects” and “assessment protocol”. Four mixed studies received high-quality global evaluation but no study received high-quality ratings for all five domains.

It is also worth mentioning that more than half of the studies received medium-quality scores for the domain “study methodology”.

Despite being rarely mentioned as a key point among the studies, three main clinical uses of the posturographic assessment in pwPD emerge from this review. These are: (1) to evaluate the effectiveness of rehabilitation treatments (37 clinical trials), (2) to monitor the course of the disease, and (3) to distinguish subgroups of patients (95 observational studies).

Only 16 studies have explicitly discussed how the assessment protocol and investigated posturographic parameters can be used in clinical routine (34–36, 38, 39, 42, 43, 50, 63, 72, 110, 115, 138, 147, 148, 150). Surprisingly, the lack of explanation of a clear rationale is particularly notable in studies published in purely clinical journals (see Figure 3A), which, on the contrary, are expected to be more accurate about this. This makes it difficult for readers to translate the literature into their daily clinical practice.

The reasons for this flaw may be manifold. It is possible that the authors took the rationale behind their choices for granted, as they are used to using these tools and dealing with these parameters every day. Also, the authors might have focused on the research-related aspects of their studies, so they did not discuss the clinical rationale and relevance of both their protocol and results. Next, several journals did not require to address this point explicitly. Concerning the choice of the posturographic parameters used in the studies included in this review, their selection may be due to—and limited by—the software of the commercial device used in the study. These, in fact, typically provide only the most used and standard outcomes, such as COP velocity and area, which may not be the most informative ones when assessing pwPD.

It is known that it takes about two decades to implement evidence into clinical practice (156). To bridge this time gap, the literature must focus on clinical relevance and transferability of findings. Many scientific journals require a short paragraph on clinical implications when submitting a manuscript. In line with this, future studies should discuss which additional information the posturographic test provides other than the clinical assessment, why it is essential to perform it in pwPD, and the possible contribution of the assessing protocol (task and parameters) to differential diagnosis or early recognition of symptoms or treatment effect assessment or patient status monitoring.

This systematic review included both pwPD and patients with atypical parkinsonism. To our knowledge, this is new in the literature, as this is the only review that included both populations. Despite being rarer, diseases such as PSP and MSA manifest clinical symptoms at onset that can be confused with PD and, therefore, deserve to be investigated as well and differentiated.

Two domains with similar percentages of high-quality studies are “assessment protocol” (32%) and “technical aspects” (31%). Many studies obtained fair scores at the critical appraisal as they correctly provided the necessary information. However, several technical and assessment-related aspects still deserve consideration.

This review has some limitations that need to be considered.

First, we included studies that reported balance assessment under the “baseline” condition. So, other protocols (e.g., dynamic posturography), which could provide additional relevant information and valuable hints for clinical practice, were not considered.

The search was conducted on four databases and by hand searching the references, but some relevant papers not indexed through the keywords entered in the search strings may have been missed.

Finally, the appraisal we used to assess the included studies on posturography in pwPD is new to the literature and is not a validated tool, as it was constructed ad hoc by the authors. Although we relied on similar examples to conceive the items (16, 23, 24) and develop the scoring system (22), followed an iterative process, and consulted with experts for its design, it is possible that some important aspects were not considered.