Planning Knee Replacement When You Have Spinal Deformity: What Your Spine Surgery Could Mean for Your Knees
The first study to examine native knee alignment classification in adult spinal deformity patients finds that ASD creates a distinct knee profile — and that spinal realignment surgery can shift individual knee alignment in ways that matter for total knee arthroplasty planning.
OPENING SUMMARY
When a patient needs both spine surgery and knee replacement, the order and coordination of those procedures matters more than most people realize. A growing approach in knee arthroplasty called kinematic alignment aims to restore each patient's unique native knee geometry — but doing that accurately requires knowing what that native geometry actually is. This study, the first of its kind, examines how adult spinal deformity influences native knee alignment classification and what happens to that classification after spinal realignment surgery. The findings carry important implications for any patient — and any surgical team — navigating both conditions at once.
STUDY SNAPSHOT
Study type – Retrospective review of prospectively collected multicenter data
Number of patients – 264
Number of knees analyzed – 528
Mean age – 59 years (range 19–84)
Female – 76%
Mean BMI – 26.2
Imaging – Full-body biplanar standing radiographs preoperatively and at 1-year follow-up
Classification system – Coronal Plane Alignment of the Knee (CPAK) — 9 phenotypes based on aHKA and JLO
Exclusions – Prior hip or knee arthroplasty
OA subgroup – 83 patients (166 knees) with radiographic knee OA (Kellgren–Lawrence grade ≥3)
Published in – The Journal of Arthroplasty, Volume 41 (2026), pages S200–S206
DOI – 10.1016/j.arth.2026.03.041
Corresponding author – Emil R. Haikal, MD — Lenox Hill Hospital, Northwell Orthopaedic Institute, New York
WHY THIS MATTERS
Total knee arthroplasty (TKA) has traditionally used mechanical alignment as its standard — positioning implants to create a neutral mechanical axis regardless of the patient's natural knee geometry. A growing body of evidence has challenged this one-size-fits-all approach, giving rise to kinematic alignment (KA), which instead aims to restore each patient's individual native joint line orientation and limb alignment. The premise is that recreating what the knee naturally looked like before arthritis may produce better soft tissue balance, kinematics, and functional outcomes.
To apply kinematic alignment effectively, surgeons need to know a patient's native coronal knee alignment — their prearthritic baseline. The CPAK classification system was developed precisely for this purpose: it uses two radiographic parameters that are independent of joint deformity to estimate native knee phenotype across nine distinct categories.
What has been almost entirely unexplored, until this study, is how adult spinal deformity affects CPAK classification — and whether spinal realignment surgery changes it. This is not a niche question. ASD is an increasingly prevalent condition, and many patients with spinal deformity also have knee arthritis. If spinal deformity systematically alters coronal knee alignment, and if spine surgery can further shift that alignment, then using CPAK measured before or during spine surgery to plan knee replacement may lead to a target that no longer reflects the patient's true native anatomy after the spine is corrected.
BACKGROUND: THE CPAK CLASSIFICATION
The CPAK system categorizes native knee alignment using two measurements obtained from full-body radiographs:
The arithmetic hip-knee-ankle angle (aHKA) describes overall limb alignment — whether the leg is in varus (bow-legged), neutral, or valgus (knock-kneed).
The joint line obliquity (JLO) describes the tilt of the knee joint line itself — whether the apex is distal, neutral, or proximal.
By combining these two parameters, the CPAK system yields nine distinct knee phenotypes. Because both measurements are derived from bony anatomy rather than joint surfaces affected by arthritis, the resulting classification is intended to reflect native — prearthritic — alignment even in patients with significant joint degeneration. The CPAK system has been validated and applied across diverse populations worldwide, with knee alignment phenotypes known to vary by geography and ethnicity.
KEY FINDINGS
1. ASD patients have a fundamentally different knee alignment profile than healthy individuals. Even before surgery, the CPAK distribution in ASD patients differed significantly from that of a healthy reference population (Chi-square = 148.45, P <0.001). Specifically, CPAK types III and VI — valgus and apex-distal phenotypes — were overrepresented, while types I, II, and V were underrepresented. The authors suggest this distinct alignment profile likely reflects chronic compensatory adaptations to altered spinal posture and pelvic orientation that develop over time in patients living with spinal deformity. Existing CPAK classification frameworks, developed in healthy populations, may not adequately represent the alignment patterns seen in this population.
2. At the group level, overall CPAK distribution remained stable after spinal surgery. The cohort's overall CPAK profile did not change significantly from preoperative to one-year postoperative (Chi-square = 10.67, P = 0.058). However, the distribution remained significantly different from healthy norms at one year (Chi-square = 189.92, P <0.001) — meaning that while surgery did not destabilize the group's alignment pattern, ASD patients continued to have a distinct knee alignment profile that differs from the general population even after their spine was corrected.
3. At the individual level, 36% of knees changed CPAK classification after surgery. Despite the stable group-level distribution, 190 of 528 knees (36%) shifted to a different CPAK phenotype at one year. These shifts were most common between adjacent phenotypes in the CPAK matrix. Patients whose knees changed classification were significantly older (62.2 vs. 54.9 years; P <0.001), had higher BMI (26.8 vs. 25.4; P = 0.03), and underwent greater corrections in spinal alignment — including larger changes in coronal C7 plumb line, sagittal vertical axis, pelvic shift, and knee flexion angle — compared to those whose classification did not change.
4. Changes in lumbar lordosis, pelvic shift, and knee flexion angle independently predicted changes in limb alignment. Multivariate analysis identified change in lumbar lordosis (standardized β = 0.186, P = 0.010), change in pelvic shift (standardized β = 0.157, P = 0.019), and change in knee flexion (standardized β = 0.128, P = 0.046) as independent predictors of change in aHKA — the limb axis component of CPAK. This confirms that spinal correction can drive measurable changes in coronal plane knee alignment, extending the known sagittal spine-knee relationship into the coronal plane. Notably, no variables remained significant as independent predictors of change in JLO, suggesting that joint line obliquity may be more stable or less sensitive to postural correction than overall limb axis alignment.
5. Older patients tended to shift toward valgus alignment after spinal correction. Age showed a positive correlation with change in aHKA (r = 0.264, P = 0.014), meaning older patients were more likely to shift in the valgus direction following ASD correction. This age-related pattern is consistent with broader observations about how aging affects lower extremity alignment and compensatory mechanisms.
6. Meaningful alignment changes occurred in approximately 12–14% of patients. Using a threshold of two degrees or greater to define a clinically meaningful change, 37 patients had a meaningful aHKA change and 31 had a meaningful JLO change. These patients were older, heavier, had higher BMI, and underwent substantially greater sagittal spinal correction — including larger changes in SVA, lumbar lordosis, T1PA, pelvic shift, knee flexion, and ankle dorsiflexion — than those with smaller changes.
7. The findings extended to patients with knee osteoarthritis. Among the 83 patients with radiographic knee OA at baseline, 39% of osteoarthritic knees changed CPAK phenotype after surgery — a rate consistent with the overall cohort. Correlation patterns and predictors of alignment change in this subgroup were also consistent with the overall population, suggesting that even arthritic knees remain responsive to postural correction from spinal realignment.
PRACTICAL IMPLICATIONS
CPAK classification should be interpreted cautiously in ASD patients — especially those undergoing or anticipating spinal surgery. The study's most direct clinical message is a caution: using a preoperative CPAK phenotype measured in a patient with untreated spinal deformity to plan kinematic TKA may result in an alignment target that no longer reflects the patient's native anatomy after the spine is corrected. If spine surgery shifts coronal knee alignment — as it did in 36% of knees in this cohort — then a TKA planned to that pre-correction CPAK target could be misaligned relative to the post-correction knee.
Addressing spinal deformity before knee arthroplasty may allow for more reliable CPAK characterization. The authors suggest that in patients with concurrent spinal deformity and knee arthritis requiring replacement, correcting the spinal deformity first — and then measuring CPAK — may provide a more accurate and stable basis for kinematic TKA planning. This has important implications for surgical sequencing decisions in patients who need both procedures.
Spine surgeons and arthroplasty surgeons should collaborate closely in patients with both conditions. The authors explicitly call for close collaboration between spine and knee specialists in patients with concurrent disease. Alignment strategies in TKA should account for the dynamic relationship between spinal and lower extremity alignment — and a TKA aligned to a malaligned spine may end up in a suboptimal position if the spine is subsequently corrected.
The CPAK system itself may have limitations in this population beyond the spine-knee interaction. The authors note that CPAK classification is based on discrete cutoffs applied to continuous variables, making it sensitive to small measurement shifts. Additionally, radiographs in this study were obtained as part of routine spinal surgery protocols rather than dedicated lower-limb imaging, which may introduce positioning variability in patients with severe deformity. Recent literature has also questioned CPAK's ability to reliably capture extra-articular deformities in complex cases. These limitations do not invalidate the system's utility, but they reinforce the need for cautious interpretation in ASD patients specifically.
Future research should focus on surgical sequencing and comprehensive alignment planning tools. The authors identify several important next questions: How does spinal realignment lead to changes in CPAK classification mechanistically? Which patients are most likely to experience meaningful knee alignment shifts? And what is the optimal sequencing and coordination of spine and knee surgery for patients who need both? Comprehensive alignment strategies that account for both spinal and lower extremity parameters — and planning tools that incorporate spino-pelvic measurements into TKA planning — represent the logical next frontier.
ABOUT THE CPAK CLASSIFICATION SYSTEM
The coronal plane alignment of the knee (CPAK) classification was originally proposed by MacDessi et al. and published in The Bone & Joint Journal in 2021. It has since been validated and applied in diverse populations worldwide. For readers who want to understand the classification system in more depth before reading this study, the original publication is available at: https://doi.org/10.1302/0301-620X.103B2.BJJ-2020-1050.R1
CONCLUSION
This study breaks new ground by examining, for the first time, how adult spinal deformity influences native coronal knee alignment classification — and what happens to that classification when the spine is surgically corrected. The findings establish that ASD patients have a distinct knee alignment profile that differs from healthy individuals, and that spinal realignment shifts individual knee alignment in a clinically meaningful proportion of patients. For a field increasingly interested in personalized, kinematically aligned knee replacement, these findings are a critical reminder that the spine and knee are not independent systems — and that surgical planning for one must account for the other.
READ THE FULL STUDY
