A Matter of Perspective: How Much Variance Exists Among Accepted Definitions of Proximal Junction Kyphosis?

Introduction: Proximal Junction Kyphosis (PJK) is usually defined as excess in PJK angle, a Cobb angle between the upper-instrumented vertebra (UIV) and a supra-adjacent vertebra (SAV), either one (UIV+1) or two (UIV+2) levels above the UIV. No expert consensus for magnitude of excess or which SAV to use exists. Our objective was to characterize the variance among existing definitions used in research.

Methods: We reviewed radiographic and outcome data of 116 thoracolumbar fusion patients over 65 years of age between 2015 and 2024. A literature review of PubMed revealed 6 commonly used definitions of PJK.

Results: Our cohort included 79 females (68.1%). Average age and BMI were 70.9 +/- 4.3 years and 28.9 +/- 5.4.

The following definitions of PJK were identified: 1) PJK angle > 20° with UIV+2 as the SAV, 2) PJK angle > 10° with a >10° change from pre-op with UIV+2 as the SAV, 3) PJK angle > 2 standard deviations from average with UIV+1 as the SAV, 4) PJK angle > 10° with a >10° change from pre-op with UIV+1 as the SAV, 5) PJK angle > 15° with UIV+1 as the SAV, and 6) PJK angle > 30° with UIV+2 as the SAV, or displaced rod fracture, or reoperation within 2 years for junctional failure, pseudoarthrosis, or rod fracture.

PJK rates, by definition, were 1) 20.7%, 2) 36.6%, 3) 4.7%, 4) 23.9%, 5) 16.1%, and 6) 10.3%. Pearson-Chi Squared testing revealed significant variance among rates (p < 0.001). Post-hoc pairwise proportions testing revealed 5 distinct pairs, and 10 pairs as non-significantly distinct. The distinct pairs were 1&3 (p=0.006), 2&3 (p < 0.001), 2&5 (p=0.009), 2&6 (p < 0.001), and 3&4 (p < 0.001).

Conclusion : 10/15 (67%) of definition pairs were not significantly different when evaluating PJK rate. Among the 5 distinct pairs, definitions 2 and 3 were most common. Definitions 1, 4, 5, and 6 were not distinct from each other. These results suggest that definitions 2 & 3 should be used with caution when consistency across studies is desired.