2.1 Sample population
Participants were recruited from the staff and student populations of Sun Yat-sen University. The inclusion criteria for the NCLBP group were as follows: 1) age between 18 and 30; 2) diagnosis of NCLBP lasting more than 3 months; 3) pain score greater than 2 on the numerical rating scale (NRS) in both static (i.e., lying, sitting, or standing) and dynamic situations (i.e., moving or walking) was defined as the present of LBP; 4) no nerve root compression symptoms (defined as pain radiating below the knee, numbness, or paresthesia in straight-leg raise test)(26); and 5) no radiographic evidence of congenital anomalies of the lumbosacral region. The exclusion criteria for the NCLBP group were as follows: 1) presence of scoliosis as assessed by the Adam’s forward bend test (27) ; 2) history of fracture or surgery in the pelvic or spinal area; 3) history of a neurological disorder or on regular medications; 4) Montreal cognitive assessment (MoCA) score of <26;and 5) pregnancy. Healthy individuals who were matched for age, gender, and education level were recruited as controls. This study included participants with a minimum NPRS score of 2 because part of the study aimed to obtain clinical information to potentially facilitate the early identification of low back pain. Thus, a slightly lower pain score was adopted as inclusion criteria. It was also for pragmatic reason to increase the sample size and enhance the statistical power.
Ethical approval of this study was obtained from the First Affiliated Hospital at Sun Yat-sen University (ETHICS No.206). An information sheet was provided to all participants prior to enrolment of the study. Written informed consent was obtained from all of the participants.
2.3 Study settings and instruments
The present study was conducted in the postural examination room in the Rehabilitation Outpatient Department of the First Affiliated Hospital of Sun Yat-sen University. The bespoke GPS 5.0 software was adopted for photo acquisition and postural analysis. The GPS hardware comprised of two aluminium, vertical bars with rulers on the sides, a plumb line for postural reference, and an adjustable mirror on the top that was attached to a stable platform (Figure 1). Two reference lines and four footprints facing different directions were used to calibrate the platform and enable consistent positioning of the feet. To set the scale of the lines, the horizontal distance between two vertical lines on the frame with rulers was recorded as 40 cm. A 1-m-tall digital camera with 2 megapixels (Logitech Pro C920; Logitech, China) was positioned 2.5 m away from the participant. The participant stood barefoot on the platform in undergarments while his or her posture was captured by the digital camera.
2.4 Data collection procedure
Prior to commencing postural assessment, participants’ characteristics, of age, gender, education level, medication status, history of NCLBP in the past year, history of any other disease, and time spent on general physical exercise per week, were recorded in a demographic information sheet. The centres of the anatomical landmarks were first marked by the assessors using red stickers. The selected anatomical landmarks were the bilateral anterior superior iliac spine (ASIS), posterior superior iliac spine (PSIS), greater trochanter, tuberositas tibiae and midpoint of the patella. Postures were captured from the anterior, posterior and left/right lateral views (Figure 1). During the postural examination, participants were asked to keep their usual body posture with their eyes looking straight ahead.
For within-day inter-rater reliability of the GPS, healthy participants received a postural assessment by two different testers separately (tester A and tester B) during the first visit. Each tester was required to identify the anatomical landmarks and apply the red sticker at the centre of each landmark. The sequence of assessments by the two testers was randomized. Tester A then repeated the postural assessment procedure one week later to establish the intra-rater reliability. The two testers were trained by an experienced therapist who was not directly involved in the study. The two raters were not aware of the group allocation of each participant.
Pelvic postural assessment for the NCLBP group
Participants in the NCLBP group received the GPS assessment by tester A on one occasion. The pelvic postural parameters that were included in the data analysis were: 1) the left/right anterior pelvic tilt angle; 2) the left/right distance between the ASIS and the midline; 3) the left/right height of the ASIS from the platform; and 4) the left/right Q angle (see Figure 1). The photo analyser module of GPS 5.0 software required the anatomical landmark to be manually confirmed from the recorded image by the assessor in order to calculate the parameter. This procedure was repeated 3 times to obtain the mean of each parameter. The mean values were used for data analysis and the calculation of pelvic asymmetry ratios. Similar to the methods used by Gnat and Bialy (8), the pelvic asymmetry ratios of each parameter were calculated first by dividing the parameter of the left side by the parameter of the right side to obtain a relative ratio between the two sides. Then, 1 was subtracted from this ratio to normalize the ratio. The equation that was used to quantify pelvic asymmetry is as follows:
Asymmetry ratio (%)=| (left pelvic postural parameter / right pelvic postural parameter) – 1 | x 100.
The pelvic asymmetry ratios that were calculated are: 1) the Q angle asymmetry ratio (QAR); 2) the height of the PSIS from the platform asymmetry ratio (PHAR); 3) the height of the PSIS from the platform asymmetry ratio (PDAR); 4) the height of the ASIS from the platform asymmetry ratio (AHAR); 5) the distance between the ASIS and the midline asymmetry ratio (ADAR); and 6) the pelvic tilt angle asymmetry ratio in the sagittal plane (PTAR).
2.5 Data analysis
All statistical analyses were conducted in SPSS ver 20.0 software (IBM SPSS Inc. Chicago, IL, USA). Statistical significance was set at p < 0.050. The sample characteristics were analysed by descriptive statistics. For the reliability analysis, relative reliability was determined by the intraclass correlation coefficient (ICC). As repeated measurements were recorded for each parameter, this study employed the ICC(2,k) model for the inter-rater reliability and ICC(3,k) model for the intra-rater reliability. ICC levels were interpreted as follows: Excellent: > 0.75; Good to Fair: 0.74 – 0.40; and Poor: < 0.40 (28). Absolute reliability was determined by the standard error of measurement (SEM) and minimal detectable difference (MDD95) (29). The MDD95 corresponds to the upper bound of the random variation that 95% of stable patients generate when tested on multiple occasions (30). The formula for MDD95 is MDD=1.96*SEM* (30). The pelvic asymmetry ratios are the ratio of pelvic postural parameters between left and right sides, which are affected by the measured values of the pelvic postural parameters.
To identify the relationship between pelvic asymmetry and NCLBP, the differences in the demographic variables (including age, height, weight, BMI, duration of exercise per week), pelvic postural parameters and pelvic asymmetry parameters between groups were tested using an independent t-test. A stepwise logistic regression analysis with a forward conditional method was employed to explore the association between pelvic asymmetry and NCLBP. Before testing this model, a bivariate Pearson correlation was performed to test the relationships between age, BMI, height, weight and the pelvic asymmetry parameters. The relationships between the occurrence of NCLBP and the other variables (including the BMI, height, weight and pelvic asymmetry parameters) were explored by Spearman correlation with the occurrences of NCLBP as ordinal data (the participants with NCLBP were marked as 1, and the controls were marked as 0). Adapting the methods used in the previous study (31), the age, BMI, height, weight, and pelvic asymmetry parameters were used as the independent variables in the logistic regression analysis.