Usual interbody cages at the lumbar spine are made of titanium or carbon fiber-polyetheretherketone (PEEK). Pure PEEK cages have more recently been proposed for its lower elasticity modulus. The goal of our study was to investigate a series of patients with circumferential fixation using anterior PEEK cages for degenerative lumbar spine disorders with a specific interest in the local lordosis. Fifty-seven consecutive patients aged 54.6 years (29 to 75) were reviewed. The level of arthrodesis varied from L2L3 to L5S1. The clinical status and the radiologic variations in local lordosis at the level of arthrodesis were measured. Decrease in lordosis at follow-up was tested in a multivariate analysis regarding age, obesity, spinal level, bone graft amount, type of posterior instrumentation, postoperative lordosis increase, and cage height. The average follow-up was 5.7 years (4 to 8). Clinical outcomes were excellent or good in 49 cases. Fusion was definite in 56 cases. Although 47 patients had no change in lordosis after surgery, 10 cases showed lordosis increase (8.2 degrees; 5 to 12). At follow-up, local lordosis decreased in 13 cases (5.6 degrees; 4 to 8). The linear model was significant (P<0.001; R=0.590) showing that loss in lordosis was related with postoperative lordosis increase (P=0.01), cage height (P<0.001), posterior instrumentation rigidity (P=0.026), age (P=0.047), and low level (P=0.013). Lumbar circumferential arthrodesis using PEEK cages provided good clinical results and fusion rate. However, lordosis correction was not maintained at follow-up, especially at lower levels, using high cages, in older patients, and when associated with a rigid primary posterior instrumentation. Regarding the last point, this is likely that the order of the instrumentation (posterior first, then anterior) played a role in the loss of lordosis in case of rigid posterior fixation.