“…Following the initial recognition of this complication, there were several attempts to modify the Harrington instrumentation, such as with the use of a Moe squared-end rod to prevent rotation following contouring, use of Harrington compression instrumentation on the convex side of scoliotic curves, and the "rod-long, fuse-short" 15 technique. These modifications have been proven to be inadequate 4,5,10,15,56 . Therefore, the use of distraction instrumentation caudad to the level of L1 or L2 is still highly inadvisable, even with the implant modifications, as a result of the substantial risk of reducing the lumbar lordosis with variable production of a symptomatic flatback deformity.…”