Influence of prone, supine, and lateral positions during spine surgery on vascular, abdominal, and postural anatomy : a comprehensive review and Bayesian meta-analysis

Abstract

Background

Patient positioning alters the three-dimensional relationship between the spine and surrounding neurovascular and visceral structures, thereby influencing both the technical feasibility and safety of lumbar procedures. Quantitative estimates of these positional shifts remain heterogeneous. Objective

To determine, across contemporary imaging studies, how prone, supine, and lateral decubitus positions alter the displacement of great vessels and retroperitoneal organs, the location of the psoas/lumbar plexus, and segmental lumbar lordosis. Methods

MEDLINE, Embase, and CENTRAL were searched from 2015 to 2025. Eligible studies compared at least two positions in adults and reported millimetre or degree differences for the outcomes of interest. Random‑effects (REML) subgroup meta‑analyses, a graph‑theoretical network meta‑analysis (netmeta), leave‑one‑out diagnostics, and Bayesian sensitivity models were performed. Risk of bias was assessed with ROBINS‑I. Results

Nine studies (41 independent comparisons; n = 1,248) met inclusion criteria. Retro‑peritoneal organs moved posteriorly by a pooled + 6.34 mm (95% CI 1.87–10.80; p = 0.007) when patients were turned from lateral decubitus to the prone position, narrowing the anterior working corridor at L2–L4. No significant pooled displacement was detected for major vessels (+ 1.26 mm, 95% CI −2.43–4.94), psoas/plexus (+ 0.94 mm, 95% CI −3.58–5.46) or segmental lordosis (+ 1.55°, 95% CI −4.62–7.73°). Direct contrasts showed that the supine-to-prone transition increased combined displacement/lordosis by + 3.64 mm / °(95% CI 0.53–6.76). Network ranking favoured the supine position for anatomical stability, but inconsistency was high (I2 = 89%). Two studies were low, three moderate, three serious and one critical risk of bias; removing serious/critical studies did not change the effect direction. Conclusions

Turning a patient prone produces a reproducible posterior migration of the colon and kidney (6 mm) and a modest increase in lumbar lordosis (3–4°). Vascular and psoas positions are highly patient-specific and cannot be assumed based on supine imaging alone. Preoperative planning should therefore incorporate position-matched imaging or intraoperative navigation, especially for anterior or anterolateral approaches at L2–L4. Further high-quality, multi-positional imaging studies are warranted to clarify the sources of the marked heterogeneity observed.