Clinically significant proximal junctional kyphosis (PJK) occurs in 20% of children treated with posterior distraction-based growth friendly surgery. In an effort to identify modifiable risk factors, it has been theorized biomechanically that low radius of curvature (ROC) implants (i.e., more curved rods) may increase post-operative thoracic kyphosis, and thus may pose a higher risk of developing PJK. We sought to test the hypothesis that EOS patients treated with low ROC (more curved rods) distraction-based treatment will have a greater risk of developing PJK as compared to those treated with high ROC (straighter) implants.

This is a retrospective review of prospectively collected data obtained from a multi-centre EOS database on children treated with rib-based distraction with minimum 2-year follow-up. Variables of interest included: implant ROC at index (220 mm or 500 mm), patient age, pre-operative scoliosis, pre-operative kyphosis, and scoliosis etiology. In the literature, PJK has been defined as clinically significant if revision surgery with superior extension of the upper instrumented vertebrae was performed.

In 148 scoliosis patients, there was a higher risk of clinically significant PJK with low ROC (more curved) rods (OR: 2.6 (95%CI 1.09-5.99), χ2 (1, n=148) = 4.8, p = 0.03). Patients had a mean pre-operative age of 5.3 years (4.6y 220 mm vs 6.2y 500 mm, p = 0.002). A logistic regression model was created with age as a confounding variable, but it was determined to be not significant (p = 0.6). Scoliosis etiologies included 52 neuromuscular, 52 congenital, 27 idiopathic, 17 syndromic with no significant differences in PJK risk between etiologies (p = 0.07). Overall, patients had pre-op scoliosis of 69° (67° 220mm vs 72° 500mm, p = 0.2), and kyphosis of 48° (45° 220mm vs 51° 500mm, p = 0.1). The change in thoracic kyphosis pre-operatively to final follow up (mean 4.0 ± 0.2 years) was higher in patients treated with 220 mm implants compared to 500 mm implants (220 mm: 7.5 ± 2.6° vs 500 mm: −4.0 ± 3.0°, p = 0.004).

Use of low ROC (more curved) posterior distraction implants is associated with a significantly greater increase in thoracic kyphosis which likely led to a higher risk of developing clinically-significant PJK in EOS patients.

Proximal junctional kyphosis (PJK) is defined as adjacent segment kyphosis >10° between the upper instrumented vertebrae and the vertebrae 2 levels above following scoliosis surgery. There are few studies investigating the predictors and clinical sequelae involved with this relatively common complication. Our purpose was to determine the radiographic predictors of post-op PJK and to examine the association between PJK and pain/HRQOL following surgery for AIS.

The Post-Operative Recovery after Scoliosis Correction: Home Experience (PORSCHE) study was a prospective multicenter cohort of AIS patients undergoing spinal fusion surgery. Pre-op and minimum 2 year f/u scoliosis and sagittal spinopelvic parameters (thoracic kyphosis–TK, lordosis–LL, pelvic tilt-PT, sacral slope-SS, pelvic incidence-PI) were measured and compared to numeric rating scale for pain (NRS) score, SRS-30 HRQOL and to the presence or absence of PJK (proximal junctional angle >100). Continuous and categorical variables were assessed using logistic regression and binomial variables were compared to binomial outcomes using chi-square.

163 (137 females) patients from 8 Canadian centers met inclusion criteria. At final f/u, PJK was present in 27 patients (17%). Pre-op means for PJK vs No PJK: Age 14.1 vs 14.7yr; females 85 vs 86%; scoliosis 57±22 vs 62±15deg; TK 28±18 vs 19±16deg ∗, LL 62±11 vs 60±12deg, PT 8±12 vs 10±10deg, SS 39±8 vs 41±9deg, PI 47±14 vs 52±13deg, SVA −9±30 vs −7±31mm. Final f/u for PJK vs No PJK: Scoliosis 20±11 vs 18±8deg, final TK 26±12 vs 19±10deg∗, LL 60±11 vs 57±12deg, PT 9±12 vs 12±13deg, SS 39±9 vs 41±9deg, PI 48±17 vs 52±14deg, SVA −23±26 vs −9±32mm∗. Significant findings: Pre-op kyphosis >40deg has an odds ratio (OR) of 4.41 (1.50–12.92) for developing PJK∗. The presence of PJK was not associated with any significant differences in NRS or SRS-30. ∗denotes p<0.05.

This prospective multicenter cohort of AIS patients demonstrated a 17% risk of developing PJK. Pre-op thoracic kyphosis >40deg was associated with the development of PJK; however, the presence of PJK was not associated with increased pain or decreased HRQOL.

Radiostereometric analysis (RSA) allows for precise measurement of interbody distances on X-ray images, such as movement between a joint replacement implant and the bone. The low radiation biplanar EOS imager (EOS imaging, France) scans patients in a weight-bearing position, provides calibrated three-dimensional information on bony anatomy, and could limit the radiation during serial RSA studies. Following the ISO-16087 standard, 15 double exams were conducted to determine the RSA precision of total knee arthroplasty (TKA) patients in the EOS imager, compared to the standard instantaneous, cone-beam, uniplanar digital X-ray set-up.

At a mean of 5 years post-surgery, 15 TKA participants (mean 67 years, 12 female, 3 male) were imaged twice in the biplanar imager. To reduce motion during the scan, a support for the foot was added and the scan speed was increased. The voltage was also increased compared to standard settings for better marker visibility over the implant. A small calibration object was included to remove any remaining sway in post-processing.

The 95% confidence interval precision was 0.11, 0.04, and 0.15 mm in the x, y, and z planes, respectively and 0.15, 0.20, and 0.14° in Rx, Ry, and Rz. Two participants had motion artifacts successfully removed during post-processing using the small calibration object.

With faster speeds and stabilization support, this study found an in vivo RSA precision of ≤ 0.15 mm and ≤ 0.20° for TKA exams, which is within published uniplanar values for arthroplasty RSA. The biplanar imager also adds the benefits of weight bearing imaging, 3D alignment measurements, a lower radiation dose, and does not require a reference object due to known system geometry and automatic image registration.