Introduction

An increasing number of people suffer from neck pain, which is associated with substantial disability as well as economic and psychological distress [1]. Neck pain ranks fourth among the most prevalent causes of disability, with a prevalence rate exceeding 30% every year [2].

Neck pain is a common manifestation of musculoskeletal disorders of the upper quadrant, including the scapular region, and cervical and upper thoracic spines. Considering the thoracic spine, various postural and inlet parameters have been previously researched in both asymptomatic and symptomatic subjects for neck pain but with conflicting results. A recent review highlighted a positive moderate correlation between age and kyphosis in healthy adults. The authors reported that kyphosis increases with aging, with significant variability between 40 and 60 year-olds [3].

A growing body of research suggests the importance of maintaining spinal sagittal balance to keep spinal pain-related issues at bay, and to maintain a balanced upright posture with a horizontal gaze, thus minimizing the energy expenditure for kee** the line of gravity aligned [4,5,6]. However, evidence is scarce regarding the relationship between thoracic postural alterations and neck pain, with even lesser emphasis on the functional status.

Moreover, an expanding body of evidence links poor posture to neck pain [7,8,9,10]. However, the specific thoracic posture-related factors that contribute to neck pain are still not fully understood. The craniocervical region of the body is positioned over the thoracic inlet (or outlet), which is anatomically bound by the manubrium of the sternum, first thoracic vertebra (T1), and first rib on either side [11]. So, it is likely that the orientation of T1 (T1 slope) can influence the sagittal balance of the craniocervical region, which in turn can reflect onto the symptoms related to cervical spine degeneration, or the possible mechanical impact on the surrounding muscles (e.g., sternocleidomastoid, semispinalis) [6]. An altered thoracic inlet angle (TIA) has been reported to be a risk factor for spine degeneration [6, 12]. In addition, T1S has been established as an important parameter in the planning of spinal surgeries, owing to its direct relationship with the cervical sagittal axis [13]. Likewise, postural alterations such as increased thoracic kyphosis have also been reported to result in a risk of fall [14] and reduced physical function [15]. Thus, considering their impact on posture, function, and surgical outcomes, it becomes imperative to study these thoracic parameters as a crucial factor in the development and severity of neck pain.

As per the available literature, no comprehensive analysis has been carried out regarding the influence of cervical spine problems on the existence or intensity of thoracic postural and inlet variables. Therefore, the objectives of this review were as follows: (1) investigate the relationship between (a) the sagittal thoracic postural and inlet parameters and (b) the measures of neck pain and the sagittal thoracic postural and inlet parameters and (2) ascertain the impact of test position and age on the thoracic spinal malalignment in subjects with nontraumatic neck pain.

Methodology

This systematic review and meta-analysis was registered on PROSPERO (PROSPERO 2022 CRD42022342274) and conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

Data sources and search strategy

A comprehensive literature search was conducted in multiple databases, including EBSCO (via CINAHL complete), PubMed, Scopus, Web of Science, and Embase in December 2022, along with a manual search of the reference lists of the available articles. The database search was rerun with the same search words in August 2023, for inclusion of any recent publications. The relevant articles published from inception until July 2023 were included.

The search was performed using the search keywords (including MeSH) “cervical pain”, “neck pain”, “posture”, “neck-shoulder pain”, “thoracic kyphosis”, “thoracic angle”, and “thoracic inlet”. Details of the search strategy are summarized in Supplementary Table 1. The longitudinal, cohort, and cross-sectional studies that examined the neck pain subjects (using standard scales to measure intensity and disability) with either or all the sagittal thoracic postural and/or inlet measures, with or without a control group for comparison, were identified. The included studies assessed the static/neutral thoracic posture of human subjects with neck pain. The studies with available English language full text were included. The studies were excluded if the full-text article could not be retrieved. Studies that assessed dynamic or working position posture; posture with external weight held; studies with neck pain pertaining to trauma, temporomandibular joint dysfunction, neurological conditions (e.g., myelopathy), migraine, etc.; and any kind of intervention studies, conference proceedings, and editorials were not considered part of this review.

Study selection

After screening the titles and/or abstracts in the primary search, all the retrieved articles were imported into the software EndNote 20.3 (Clarivate Analytics) and checked for duplicates by both authors. All full-text articles were then independently assessed by two reviewers (B. R., A. P.) for relevance, based on inclusion and exclusion criteria. Secondary/hand searching was done from the reference list of available articles. Using both reviewers’ lists of relevant studies, a comparison was drawn. Both reviewers independently then extracted data and assessed the risk of bias in the included studies. Any difference in opinion was resolved by consensus.

Data extraction

Relevant data from all the studies was included, in the form of study design, sampling methods; sample characteristics; inclusion criteria for the neck pain group and control group; outcome measures for neck pain, sagittal thoracic posture, and thoracic inlet; results; and conclusion. Corresponding authors of included studies were contacted via email, wherever additional data was required.

Quality assessment

Quality assessment for all the included studies was done through the Newcastle-Ottawa Quality Assessment Scale (NOS) (adapted for cross-sectional studies) [16] independently by two reviewers (B. R., A. P.). Any disagreement was resolved through a mutual consensus. The NOS investigates the possibility of bias in three distinct domains: selection (maximum five stars), comparability (maximum two stars), and outcome (maximum three stars), thus allowing for a total score of 10 (10 stars). All studies included in the review were assessed for quality, regardless of their inclusion in the meta-analysis.

Grading the certainty of evidence

The certainty of evidence was graded by the two authors independently using GRADE (Grading of Recommendations Assessment, Development and Evaluation) approach (http://gdt.gradepro.org). Any difference in opinion was resolved through mutual consensus. The grading was done to incorporate the following key domains: (a) risk of bias, (b) inconsistency, (c) indirectness, and (d) imprecision, along with other optional ones.

Statistical analysis

Mean differences (MDs) with 95% confidence intervals (CIs) for continuous outcomes (T1S, TIA, NT, HTA, TKA) were used to estimate the pooled effects. Using the data extracted from the eligible studies, the continuous variables were expressed as the mean value and standard deviation of the outcome measures assessed in pain and pain-free groups, along with the number of participants for which the variables were measured in each group. The I2 test assessed the statistical heterogeneity between studies, whereas heterogeneity across the studies was analyzed using Cochran’s Q test and then transformed into I2 percent with its p-value. The Review Manager (RevMan v5.4.1, The Cochrane Collaboration, Software Update, Oxford, UK) was used for meta-analyses. The analysis for outcome measures of all thoracic postural and inlet parameters was done using RevMan, for a comparison among neck pain and asymptomatic subjects. The random-effects model was used owing to the heterogeneity of participants. The heterogeneity of I2 value above 25%, 50%, and 75% is regarded as low, moderate, and high, respectively [17]. In case of high heterogeneity for any variable, sensitivity analysis was performed, using sequential and combinatorial algorithms as suggested by Patsopoulos et al. [18]. The positional variation (upright vs. lying position) for assessment of thoracic inlet parameters was analyzed using Stata 16 (Stata Corp. LLC, 4905 Lakeway Drive College Station, TX 77845-4512, USA). As there were a limited number of included studies per variable (< 5), publication bias was not assessed.

Results

Study selection and characteristics

Out of a total of 311 studies identified in the primary search, finally, 15 studies were found relevant for the review as per the eligibility criteria. However, owing to a lack of appropriate data, only 12 studies were included in the meta-analysis for association of sagittal thoracic postural and inlet parameters with the presence and severity of neck pain and related disability or for test position difference among the inlet parameters. PRISMA flowchart is described in Fig. 1. Relevant data extracted from individual studies are described in Tables 1, 2, and 3.

Fig. 1
figure 1

Prisma flow chart for systematic review and meta-analysis of included studies

Full size image
Table 1 Data extraction for thoracic postural studies
Full size table
Table 2 Data extraction for thoracic inlet studies
Full size table
Table 3 Demographic details of included studies
Full size table

Thoracic inlet parameters

The data for thoracic inlet parameters from 710 subjects across 5 studies [12, 25,26,27, 29] presented the pooled mean difference between neck pain and asymptomatic subjects to be significant for TIA [2.12 (0.48, 3.75); p = 0.01] but insignificant for T1S [−0.86 (−5.73, 4.01); p = 0.73] and NT [2.37 (−1.74, 6.49); p = 0.26]. There was a significantly high heterogeneity for NT and T1S (I2 = 92, 97% respectively), while it was low for TIA (I2 = 44%) (Figs. 2, 3 and 4). The single study removal using a sequential algorithm did not help much with reducing the between-study heterogeneity for T1S and NT to sub-threshold. Therefore, the combinatorial algorithm was applied. The omission of two studies [25, 29] effectively cut down the heterogeneity to I2 = 65% and 48% for T1S and NT respectively. For TIA, however, a single study [27] omission successfully brought down the heterogeneity to nil (I2 = 0%) and indicated significantly higher TIA for neck pain group subjects (Supplementary Fig. 1).

Fig. 2
figure 2

Forest plot for TIA in neck pain subjects compared to asymptomatic subjects

Full size image
Fig. 3
figure 3

Forest plot for T1 slope in neck pain subjects compared to asymptomatic subjects

Full size image
Fig. 4
figure 4

Forest plot for neck tilt in neck pain subjects compared to asymptomatic subjects

Full size image

Sagittal thoracic postural parameters

The outcome measures used to assess the sagittal thoracic curvature included photographically measured high thoracic angle (HTA) [7, 8] and its complementary measure called upper thoracic angle (UTA) [22], thoracic kyphosis angle (TKA) [9, 21] and index (TKI) [20], and midthoracic curve (MTC) [19].

The two studies [7, 8] measuring high thoracic angle (N = 161) reported an insignificant pooled mean difference of 4.42 (−1.50, 10.34); p = 0.14 (Fig. 5). There was a high statistical heterogeneity of I2 = 88% and chi-square = 8.51. The complementary measure of HTA, i.e., upper thoracic angle (UTA), was examined by Kanda et al. in young and elderly female subjects [22]. The study reported that subjects with neck-shoulder pain had larger UTA than control, and young subjects had smaller UTA than elderly.

Fig. 5
figure 5

Forest plot for high thoracic angle in neck pain subjects compared to asymptomatic subjects

Full size image

The use of a spinal mouse to measure thoracic kyphosis had an insignificant pooled mean difference of 4.91 ((−3.72, 13.54); (p = 0.26)) among the neck pain and asymptomatic subjects [21, 22] (Fig. 6). TKA was reported by two studies (N = 438) having a high statistical heterogeneity (I2 = 94%, chi-square = 16.99). Quek et al. used flexicurve to analyze the thoracic kyphosis index in elderly neck pain subjects, without comparison to a control group [20]. Increased thoracic kyphosis was significantly correlated with age, but not with neck pain-related disability.

Fig. 6
figure 6

Forest plot for thoracic kyphosis angle in neck pain subjects compared to asymptomatic subjects

Full size image

Though Helgadottir et al. reported no significant difference in MTC (p = 0.99) between the neck pain and control groups, the exact data was not reported [19]. We tried communicating with the authors, but no response was received.

Positional difference for inlet parameters

Six studies [12, 23, 24, 26,27,28] were included in the analysis for the positional variation (lying vs. upright) in thoracic inlet parameters. Data from 516 subjects revealed a 3.14° higher TIA, 4.12° higher NT, and 2.26° lower T1 slope in lying position (relative to upright) for neck pain subjects. There was no significant between-study heterogeneity (I2 = 0%; p = 0.9) for positional variation as depicted in Figs. 7, 8 and 9.

Fig. 7
figure 7

Meta-analysis for test-position difference (lying vs upright) in TIA, for neck pain subjects

Full size image
Fig. 8
figure 8

Meta-analysis for test-position difference (lying vs upright) in neck tilt, for neck pain subjects

Full size image
Fig. 9
figure 9

Meta-analysis for test-position difference (lying vs upright) in T1 slope, for neck pain subjects

Full size image

Correlation of inlet and postural parameters with neck pain

In a sample of 101 office workers examined at the desk between the 4th and 5th h of work, Nejati et al. reported a significant association between HTA and neck pain when in the working posture but not when in the neutral position. The authors, however, failed to measure it statistically in terms of correlation coefficient [8]. Lau et al. observed HTA to be positively correlated with neck pain intensity (r = 0.43) and disability (r = 0.44) [7]. Similar results were reported by Kaya and Celenay who found a positive correlation between TKA and neck pain intensity (r = 0.391) [9]. However, Tsunoda et al. did not ascertain an association between TKA and neck-shoulder pain [21], for there were negligible differences in the two groups. In the study on elderly female subjects with neck pain, Quek et al. reported a negligible negative correlation of TKI with neck disability index (NDI) (r = −0.05) [20].

While three included studies [6, 24, 26] reported a significant correlation among the inlet parameters (T1S, TIA, NT), no study analyzed for correlation of inlet parameters with neck pain measures. Also, the authors did not find any study to date looking for a correlation between sagittal thoracic postural and inlet parameters.

Neck pain predictors

The thoracic kyphosis was a significant predictor for the presence of neck pain [7, 9] but insignificant for neck pain intensity and disability [7]. As reported by Kaya and Celenay [9], the cutoff for sagittal thoracic curvature and mobility to detect neck pain was 45.5° and 30°, respectively. Using multiple logistic regression with age and gender adjustment, HTA was found to be a good predictor for the presence of neck pain (OR = 1.37, p < 0.01) [7]. T1 slope (but not TIA) was reported to be a significant risk factor for degenerative neck pains [25, 29]. A > 22° of T1S (pre-operative) was shown to be of significant diagnostic value for degenerative cervical spondylolisthesis. However, age as a neck pain predictor was quite contrarily reported, with an insignificant relation outlined by two studies [25, 29]. On a conflicting note, Tsunoda et al. reported both age and gender to be significant predictors of neck-shoulder pain [21].

Risk-of-bias assessment

The risk of bias in the included studies was assessed using the NOS, the results of which are shown in Table 4. All the studies were assessed for quality grouped into three domains (selection, comparability, and outcome). The quality scores of included studies varied from 4 to 9, with a median score of 6. The quality of the studies was categorized based on the method described in a previous study [30]. The study quality ranged from poor (n = 12) to fair (n = 1) to good (n = 2). The score did not affect the inclusion or exclusion of any study from the review, it instead dictated the strength of the reported results.

Table 4 Quality scores for risk-of-bias assessment
Full size table

Certainty of evidence

We found certainty of evidence to be very low to low for all outcome measures, except TKA which had moderate certainty of evidence for differentiating between neck pain and asymptomatic subjects. Supplementary Table 2 summarizes the certainty of evidence using GRADE.

Discussion

This review aimed to investigate the association between thoracic spine dysfunction and neck pain. We identified 15 studies that met our eligibility criteria, and subsequently, 12 of them were subjected to meta-analysis to examine the impact of nontraumatic neck pain on sagittal thoracic postural and inlet parameters. Low- and very low-certainty evidence indicates that findings of the available literature on thoracic inlet and postural parameters must be viewed with caution, given that number of studies per postural parameter ranged from 1 to 2.

The included studies examined the thoracic inlet variables in subjects with various types of nontraumatic neck pain (including non-specific, cervical degenerative disc conditions, spondylolisthesis, spondylosis, and others). These findings suggest that cervical sagittal balance and postural alterations are present across different types of neck pain. Various thoracic postural measures determining the increased thoracic kyphosis included higher thoracic kyphosis angle and index and greater UTA (or lower HTA). The studies focusing on analyzing these parameters also associated neck pain with sagittal thoracic posture and other factors like age and gender. However, it is worth noting that there is a lack of high-quality research to support these associations. Another parameter of interest to authors was the correlation of the inlet parameters (TIA, T1S, NT) with the thoracic postural and neck pain measures. Surprisingly, none of the studies investigated these crucial connections.

All included studies reported postural and inlet variables recorded during the neutral position in neck pain subjects, with or without comparison to a control group. Our review suggests that inlet parameters like neck tilt and T1 slope, as well as the sagittal postural parameters for thoracic kyphosis, did not significantly differ in neck pain subjects relative to asymptomatic individuals. However, TIA was significantly higher in the neck pain subjects, compared to the controls, corroborating previous observations [32, 33], cervical lordosis as an independent risk factor for the effective conservative treatment, and a higher T1S and cervical lordosis in the effective treatment group [34]. TIA has been assumed to be a constant morphologic parameter like pelvic incidence; hence, TIA can be the basis of planning surgical treatment to restore spinal alignment [6, 35].

The first thoracic vertebra at the cervicothoracic junction is of paramount importance, as being at the base of the neck it reacts to the tilting of the neck in any direction or under any stress. Thus, T1 responds to the physiological changes in the cervical spine above, and NT responds to maintain a horizontal gaze [

Conclusion

This review highlights the limited and heterogenous evidence of low to fair quality, available with regard to the relationship between the sagittal thoracic postural and inlet parameters to pain variables in nontraumatic cervical dysfunction. TIA was the only thoracic inlet variable to be significantly different for symptomatic and asymptomatic subjects. With insufficient evidence, if thoracic posture and neck pain are associated, it is minuscule. Test-position difference reflected marginally lower T1 slope, and higher TIA and neck tilt in lying compared to upright, for neck pain patients. Also, only thoracic kyphosis and T1 slope could predict the presence of neck pain. There is a lack of evidence for associating the inlet and postural parameters in subjects with neck pain.

Defining the key terms

Thoracic inlet parameters [12, 23,24,25,26,27,28,29]

  • Thoracic inlet angle (TIA)—Angle formed by a line perpendicular to the superior end plate (SEP) of T1 and a line connecting the center of SEP of T1 and the upper end of the sternum

  • Neck tilt (NT)—Angle formed by reference vertical line drawn in the upper end of the sternum and a line connecting the center of the SEP of T1 and the upper end of the sternum

  • T1 slope (T1S)—Angle formed between the reference horizontal line and the SEP of T1

Thoracic postural parameters

  • High thoracic angle (HTA)—Angle between a line connecting C7 to T7 and a horizontal line from T7 [7, 8]

  • Upper thoracic angle (UTA)—Angle between a line connecting C7 to T5 and a vertical line from T5 [22]

  • Thoracic kyphosis angle (TKA)—The sum of the angles T1/2 to T11/12 [21]

  • Thoracic kyphosis index (TKI)—Thoracic width/horizontal thoracic length × 100 [20]

  • Midthoracic curve (MTC): 4 × [arctan (2 × thoracic height/thoracic length)] [19]