Characteristics of the spinopelvic parameters of patients with sacroiliac joint pain

To evaluate the characteristics of the spinopelvic parameters on radiography in patients with sacroiliac joint pain (SIJP). Two hundred fifty patients were included and divided into the SIJP group (those diagnosed with SIJP based on physical findings and response to analgesic periarticular injections; n = 53) and the non-SIJP group (those with low back pain [LBP] because of other reasons; n = 197). We compared their demographic characteristics and spinopelvic parameters using radiography. All differences found in the patients’ demographic characteristics and spinopelvic parameters were analyzed. More female participants experienced SIJP than male participants (P = 0.0179). Univariate analyses revealed significant differences in pelvic incidence (PI) (P = 0.0122), sacral slope (SS) (P = 0.0034), and lumbar lordosis (LL) (P = 0.0078) between the groups. The detection powers for PI, SS, and LL were 0.71, 0.84, and 0.66, respectively. Logistic regression analyses, after adjustment for age and sex, revealed significant differences in PI (P = 0.0308) and SS (P = 0.0153) between the groups, with odds ratios of 1.03 and 1.05, respectively. More female participants experienced SIJP than male participants. Higher PI and SS values were related to SIJP among LBP patients.


Methods
Subjects. The study was performed at one of the core hospitals in our city. The hospital has surgical and outpatient departments. The out-patient department adopts a complete referral system; therefore, only patients who have a referral letter go to the hospital for consultation. A total of 281 consecutive LBP patients were registered consecutive LBP patients from June 2015 to December 2019 during their first visit. LBP was defined as a pain localized between the costal margin and the inferior gluteal folds. Radiography of the lumbar spine and pelvis at a standing position were evaluated for all patients. Lumbar MRI examinations were added and evaluated for LBP cases with buttock and/or lower limb pain as well as severe LBP cases without buttock or lower limb pain. Following these radiological evaluations, patients who had an infection or tumor in the lumbar spine or pelvis were excluded. We excluded 24 cases of metastatic lumbar spine tumor and 7 cases of lumbar infectious spondylitis/ discitis. Finally, a total of 250 patients were included in the study. Subsequently, an analgesic periarticular SIJ injection was administered in order to diagnose and evaluate patients who indicated buttock pain in the region from the posterior iliac crest to the inferior gluteal fold and with positive result on either the Patrick's test or the SIJ shear test. We defined the SIJP group (n = 53) as those who experienced unilateral buttock pain and fulfilled the criteria described here, as defined in our previous report 10 . The non-SIJP group (n = 197) was defined as those who had experienced LBP due to reasons other than SIJP. We used a self-written questionnaire to collect patient background information that included age, sex, height, weight, and smoking habits. The body mass index (BMI) was calculated using height and weight data. The first author measured spinopelvic parameters such as PI, PT, SS, and lumbar lordosis (LL) using the radiographs. We compared and analyzed the demographic characteristics and spinopelvic parameters of both groups. The medical/ethics review board of Kanto Rosai Hospital approved the study, including any relevant details, and confirmed that all experiments were performed in accordance with relevant guidelines and regulations and that informed consent was obtained from all participants.

Definition of SIJP.
We used our previously published definition for SIJP 10 . We precisely defined and evaluated the region where buttock pain occurred as pain at the area from the posterior iliac crest to the inferior gluteal fold 10 . We defined SIJP as unilateral buttock pain that met the following criteria: positive results on either the Patrick's test or the SIJ shear test, which are direct stress tests of the SIJ region 1 ; positive response to the analgesic periarticular SIJ injection; no findings of nerve root or cauda equina compression on lumbar MRI; no hip osteoarthritis on radiography of the hip. We evaluated physical findings, including Patrick's and SIJ shear test results, and radiographs of the lumbar spine and pelvis. When we suspected SIJP, we evaluated lumbar MRI findings in order to rule out the possibility of lumbar diseases and performed an additional examination using an analgesic SIJ injection, which was a 3-mL injection of 1% lidocaine, into the periarticular space of the SIJ with fluoroscopic control in reference to previous studies 3,8 . We evaluated the degree of buttock pain both before and after the injection. We considered a reaction to the analgesic injection as positive when the pain had improved more than 70% 15 min after the injection.
Statistical methods. Descriptive statistics were determined and presented. Using the Anderson-Darling test, we evaluated whether each item was normally distributed or not. Between-group differences in demographic characteristics were evaluated using the Fisher's exact test for categorical variables. Continuous variables were shown as means ± standard deviation for parametric items and medians as and first/third quantiles for nonparametric items. The P values were calculated by Student's t-test and the Wilcoxon rank sum test for parametric and non-parametric items, respectively. The post hoc detection power was calculated as 1 − beta and represents the probability that we reject the null hypothesis when it was false 13 . In addition, we performed logistic regression analyses adjusted by age and sex for the spinopelvic parameters. One guideline suggested that accurate estimation of the discriminant function parameters required a sample size of at least 20 cases for each independent variable 14 . The number of the case group was 53; thus, three variables, including age and sex, were adopted in the logistic regression analyses for each parameter. Correlation coefficient analysis was performed for each of the spinopelvic parameter items. We performed receiver operating characteristic (ROC) curve analysis and calculated the area under the ROC curve (AUC), sensitivity, and specificity for each item. Statistical analysis was performed using JMP version 14.0 (SAS Institute, Cary, NC, USA). A P value < 0.05 was considered significant.

Results
There were no significant differences in the demographic characteristics (age, BMI, and smoking habits) between the groups except for sex. The ratio of female participants was significantly higher in the SIJP group than that in the non-SIJP group (P = 0.0179). Among the continuous variables, PI and SS were normally distributed, whereas age, BMI, PT, and LL were not normally distributed. There were significant differences in PI (P = 0.0122), SS (P = 0.0034), and LL (P = 0.0078) between the groups ( Table 1). The detection powers for PI, SS, and LL were 0.71, 0.84, and 0.66, respectively. The mean PI values were significantly higher for female than for male participants (51.6° ± 12.0° vs. 47.9° ± 9.3°; P = 0.0089). The mean SS values for female and male participants were not significantly different (34.0° ± 9.5° vs. 33.0° ± 8.1°; P = 0.3789). Logistic regression analyses of 250 radiographs, after an adjustment for age and sex, revealed significant differences in PI (P = 0.0308) and SS (P = 0.0153) between the groups, with odds ratios of 1.03 and 1.05, respectively (Table 2). Among each of the spinopelvic parameters, SS and LL showed a strong correlation coefficient of 0.764, followed by a correlation coefficient of 0.676 for PI and SS (

Discussion
Higher PI and SS values were significantly related to SIJP. We defined SIJP in the study as unilateral buttock pain that met the following criteria: positive results on direct stress tests of the SIJ region 1 ; positive response to the analgesic periarticular SIJ injection; no other lumbar spine or hip diseases. The diagnosis of SIJP has not been clearly confirmed and remains challenging 1-3 . Laslett 15 demonstrated that three or more positive pain provocation SIJ tests have sensitivity and specificity of 91% and 78%, respectively. However, other studies showed that widely used provocation tests do not show high specificity 7,8 . Dreyfuss et al. 16 concluded that SIJP is difficult to identify using historical and physical examination data. Although these studies were published more than 10 years ago, the diagnosis of SIJP using only physical examination has remained challenging. Therefore, we did not consider several physical examinations as gold standard for diagnosing SIJP in our study; instead; therefore, we defined SIJP in detail as aforementioned. A literature review recommended using the term "pelvic girdle pain, " which included pain originating from the intraarticular region and the posterior ligamentous region of the SIJ 3 . Other studies have demonstrated that the posterior ligament region is a significant source of SIJP 17,18 . We had also adopted the same criteria using periarticular block for SIJP following lumbar spine surgery in our previous study 12 . The incidence was also as appropriate and acceptable as those reported in previous works regarding SIJP following lumbar spine surgery. Therefore, we adopted periarticular SIJ injection to evaluate the response to analgesic injections. We met the criteria using lumbar MRI and hip radiography for all patients in the SIJ group for exclusion of lumbar and hip causes, respectively. Although the lumbar MRI findings always reveal the causes of LBP, we performed lumbar MRI in our work to exclude specific findings, which could be associated with buttock pain development, such as lumbar disc herniation or lumbar spinal stenosis. Therefore, our definition of SIJP is appropriate.
The SIJP group included approximately 20% of all patients in the current study, which were consistent with those of previous studies that reported 10-30% in their SIJP groups 2,3 . No significant differences were observed Table 1. Demographic characteristics and spinopelvic parameters by univariate analyses. Data are presented as means ± standard deviation for parametric items and medians and first/third quantiles for non-parametric items, or as numbers of participants (%). SIJP sacroiliac joint-related pain. *P < 0.05. The P values were calculated by Student's t-test and the Wilcoxon rank sum test for parametric and non-parametric items, respectively.  Table 2. Logistic regression analyses for odds ratios and 95% confidence intervals of the spinopelvic parameters for sacroiliac joint-related pain. Data are calculated by logistic regression analysis after adjustment for age and sex on 250 radiographs. The number of the case group was 53; thus, three variables, including age and sex, were adopted in the logistic regression analyses for each parameter. *P < 0.05.  Table 3. Correlation coefficient of each item derived from the univariate analyses.

Pelvic incidence Pelvic tilt Sacral slope Lumbar lordosis
Pelvic incidence - www.nature.com/scientificreports/ regarding the demographic characteristics of the two groups except for sex. An epidemiological report showed that more female participants presented with SIJP than male participants 19 ; however, reports about sex differences are scarce. Our result was consistent with this report and provides findings about sex differences related to SIJP. A study reported that the ligaments around the SIJ in female participants were not as strong as the ligaments in male participants, which could cause instability of the SIJ 20 . A finite element study showed that female participants diagnosed with SIJP had higher mobility, stresses, loads, and pelvis ligament strains than male participants, which could account for the higher incidence of SIJP in female participants 21 .
In the logistic regression analyses after adjustment for age and sex, we found that PI and SS were the significant items for the diagnosis of SIJP. A study showed that high PI and SS were related to LBP, especially in female participants 22 , whereas another report showed that low PI, SS, and LL were related to LBP in general 23 . Although both studies did not define the area of the LBP in detail and mentioned only SIJP, our result corresponded with the findings of the former study. PI is a sum of the PT and SS. In the current study, the PI and SS values were moderately correlated, and the difference in the PI values between the two groups could be attributed to that of the SS values, whereas there was not much difference in the PT values between the two groups.
A study showed that a posterior shift in the shearing force line that resulted from a high PI might place a load on the posterior components of the low back, which was followed by an increase in LBP 24 . That mechanism may account for our results; however, the PI value of the SIJP group in our study was similar to the Japanese normative value of 52.3°2 5 . Therefore, our results could be interpreted as a low PI value of the non-SIJP group, rather than the value for the SIJP group being high. A small LL is often seen in patients with lumbar diseases such as spinopelvic imbalance 26 . Although the LL value was excluded after the logistic regression analyses after adjustment for age and sex, the LL value was highly correlated with the SS value and moderately correlated with the PI value in the current study. Therefore, the small SS and PI values could correlate with the small LL value in the non-SIJP group, in which the patients did not experience SIJP but rather LBP caused by spinopelvic imbalance. As the result, the SIJP group demonstrated significantly higher PI values than the non-SIJP group, which indicated that buttock pain without spinopelvic imbalance or findings of nerve root or cauda equina compression on lumbar MRI could indicate SIJP.
There were some limitations to this study. First, there could be false-positive responses to the analgesic injection. A false-positive response rate to the analgesic SIJ injection was found to be as high as 10-20% 8 . This limitation could result in an overestimation in the SIJP group. Second, the SIJP group may be underestimated depending on interpretation of lumbar MRI examinations. Asymptomatic lumbar disc herniation and lumbar spinal stenosis findings sometimes occur 27 . Moreover, a report demonstrated that SIJP often co-existed with other lumbar diseases 28 . Therefore, we could have excluded a possible cause of SIJP. Third, there was selection bias among our patients. Patients from one hospital cannot represent all patients with SIJP. However, an advantage of our study is that all consecutive patients seeking care for low back or buttock pain who met study eligibility were completely evaluated, including all the required examinations, tests, and imaging, which could decrease selection bias.
In conclusion, our study found that more female participants presented with SIJP than male participants. This study is the first to identify that higher PI and SS values were significantly related to SIJP among LBP patients. These results could be of value in diagnosing SIJP.

Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.