Poor visualization limits diagnosis of proximal junctional kyphosis in adolescent idiopathic scoliosis

BACKGROUND CONTEXT: Multiple methods are used to measure proximal junctional angle (PJA) and diagnose proximal junctional kyphosis (PJK) after fusion for adolescent idiopathic scoliosis (AIS); however, there is no gold standard. Previous studies using the three most common measurement methods, upper-instrumented vertebra (UIV)+1, UIV+2, and UIV to T2, have minimized the difficulty in obtaining these measurements, and often exclude patients for which measurements cannot be recorded. PURPOSE: The purpose of this study is to assess the technical feasibility of measuring PJA and PJK in a series of AIS patients who have undergone posterior instrumented fusion and to assess the variability in results depending on the measurement technique used. STUDY DESIGN/SETTING: A retrospective cohort study was carried out. PATIENT SAMPLE: There were 460 radiographs from 98 patients with AIS who underwent posterior spinal fusion at a single institution from 2006 through 2012. OUTCOME MEASURES: The outcomes for this study were the ability to obtain a PJA measurement for each method, the ability to diagnose PJK, and the inter- and intra-rater reliability of these measurements. METHODS: Proximal junctional angle was determined by measuring the sagittal Cobb angle on preoperative and postoperative lateral upright films using the three most common methods (UIV+1, UIV+2, and UIV to T2). The ability to obtain a PJA measurement, the ability to assess PJK, and the total number of patients with a PJK diagnosis were tabulated for each method based on established definitions. Intra- and inter-rater reliability of each measurement method was assessed using intra-class correlation coefficients (ICCs). RESULTS: A total of 460 radiographs from 98 patients were evaluated. The average number of radiographs per patient was 5.3 +/- 1.7 (mean +/- standard deviation), with an average follow-up of 2.1 years (780 +/- 562 days). A PJA measurement was only readable on 13%-18% of preoperative films and 31%-49% of postoperative films (range based on measurement technique). Only 12%-31% of films were able to be assessed for PJK based on established definitions. The rate of PJK diagnosis ranged from 1% to 29%. Of these diagnoses, 21%-100% disappeared on at least one subsequent film for the given patient. ICC ranges for intra-rater and inter-rater reliability were 0.730-0.799 and 0.794-0.836, respectively. CONCLUSIONS: This study suggests significant limitations of the three most common methods of measuring and diagnosing PJK. The results of studies using these methods can be significantly affected based on the exclusion of patients for whom measurements cannot be made and choice of measurement technique. (C) 2015 Elsevier Inc. All rights reserved.