Welcome. This short talk explains what a lumbar disc replacement is, who is a good candidate, and what the research shows about how well it works and how long it lasts. Every number in this talk comes from a published study; the list is on the last slide and the full list with links is on this deck's reference page. Please interrupt with questions at any point.

Both operations start the same way: the painful, worn-out disc is removed. In a fusion the two bones are then joined with a cage, bone graft and screws, and over months they heal into one bone. That stops the pain generator, and it also stops the motion at that level. In a disc replacement an artificial disc goes in its place: two metal plates that grip the bone and a smooth plastic core between them, so the level keeps bending. The operation is done through a small incision in the lower abdomen, with the organs moved aside, which is why a vascular surgeon often helps with the approach.

Selection is what makes this operation work, so here is who it is for and who it is not for. The left column comes from the FDA approval labeling of the two discs on the US market, prodisc L and activL FDA 2020 FDA 2015, plus the insurers. A good candidate has one worn-out disc between L3 and the sacrum, usually L4-5 or L5-S1, with pain that comes from the disc itself: back pain more than leg pain, and MRI wear that matches the symptoms. At least six months of nonsurgical care has been tried. The bone is healthy, which we check with a bone-density scan; a T-score of minus 1.0 or worse rules it out, because the plates have to seat in solid bone. The facet joints behind the disc are healthy, and any slip of the vertebra is slight. Most insurers limit it to ages 18 to 60, the range the trials studied Aetna 2026 Cigna 2026 Carelon 2026.

Who it is not for: osteoporosis or osteopenia; arthritic facet joints, a stress fracture in the back of the vertebra or a slipped vertebra, because then the pain is not only from the disc and the joint would still move on a damaged hinge; spinal stenosis or leg-dominant nerve pain, which needs a decompression; a previous fusion at that level or more levels than the device is approved for; scoliosis, infection, or allergy to the implant materials. The insurers add severe obesity, nicotine and untreated depression or substance use.

Park's ten-year study shows why this matters: well-selected patients had 77 percent success and 87 percent satisfaction; poorly selected patients 40 and 60 percent Park 2016.

The first question everyone asks: does it work? Three large randomized trials were run in the United States to get the devices approved, each comparing a disc replacement with a fusion. All three showed that disc replacement relieves pain and improves function at least as well as fusion Blumenthal 2005 Zigler 2007 Garcia 2015. At five years in the ProDisc-L trial, 77 percent of patients in both groups were satisfied, but 82 percent of the replacement patients said they would have the surgery again compared with 68 percent of the fusion patients Zigler 2012. When the randomized trials are pooled, patients with a replacement needed about half as many repeat operations: roughly 6 percent versus 12 percent Patel 2026 Zigler 2018. One honest footnote: a Norwegian trial compared replacement with an intensive rehabilitation program rather than fusion, and at eight years the difference in function was small, although a quarter of the rehabilitation patients ended up choosing surgery Furunes 2017. Surgery is for people who have already given nonsurgical care a real try.

These are the numbers that explain why patients who qualify tend to choose a replacement. In the Swedish randomized trial, with 99 percent follow-up at five years, 38 percent of replacement patients said they were completely pain-free, compared with 15 percent after fusion; function scores were better as well Sköld 2013. In the Charité trial, 66 percent of replacement patients were working full-time at five years compared with 47 percent after fusion, and only 8 percent were on long-term disability compared with 21 percent Guyer 2009. In a series of 39 athletes, 95 percent returned to their sport, most within six months, and 85 percent said their performance improved Siepe 2007. In the activL trial, nobody in either replacement group was still taking opioids at seven years Radcliff 2021; across the pooled trials, fewer replacement patients were using narcotics than fusion patients Patel 2026.

Recovery is also faster: the Swedish trial found replacement patients had reached their best result by one year, while fusion patients took two Berg 2009. And it holds up: at a mean of nearly fourteen years in a Spanish series, 74 percent were still in their original job and no implant had been revised Vanaclocha 2025. For balance, a 17-year workers' compensation series had 81 percent back to work but 38 percent needing some further surgery Carlson 2022.

The hospital side of the story comes from a national database of US admissions between 2010 and 2019: about 8,000 lumbar disc replacements compared with 1.1 million fusions, matched for age, sex and health Upfill-Brown 2022. Any in-hospital complication occurred in 7 percent of replacement patients versus 13 percent of fusion patients; blood transfusion in 3 percent versus 8 percent; the stay was about two-thirds of a day shorter; and the hospital cost was about 4,500 dollars lower. Another analysis using insurance claims found a slightly longer stay for arthroplasty than for one type of anterior fusion, so I would not promise a specific number of days Khalid 2023; what is consistent is fewer complications and less blood loss.

The reasons are mechanical: no bone has to heal, so there is no brace and no waiting; no screws go through the back muscles; and the front approach spares the muscles that hold you upright. The Swedish economic analysis found the health-care cost of replacement was lower than fusion over two years, with similar overall value Fritzell 2011.

This is the reason to choose a replacement over a fusion: the level keeps moving. In the Charité trial, the operated level bent about six degrees at five years after a replacement, compared with about one degree after a fusion Guyer 2009. In the ProDisc-L trial, 94 percent of patients had a normal range of motion at two years and it held steady through year five Zigler 2007 Zigler 2012. The longest series are reassuring: David followed 106 patients for an average of thirteen years and found 9 in 10 levels still moving, with about ten degrees of bend David 2007; Lemaire's hundred patients at ten years showed about the same Lemaire 2005. The honest caveat is that over many years some levels slowly stiffen, and bone can bridge around the implant; in one fifteen-year series that happened in most segments, yet 93 percent of patients still had a good result Lu 2018. Stiffening is common; trouble from it is not.

How long does it last? Two different questions hide in there. The first is whether the implant itself fails. At one center that has done these operations for twenty years, only 1 in 80 implants, 1.3 percent of 2,141 patients, was ever removed or exchanged Guyer 2024. In the longest published follow-up, thirty patients at an average of fifteen years, 93 percent had a good or excellent result and no implant had failed Lu 2018; David's 106 patients at thirteen years were similar David 2007. The second question is whether you might need any further spine surgery. That is more common: across the ten-year studies about 1 in 10 patients had another operation, usually a fusion added at the same level or surgery at a neighbouring level Lemaire 2005 David 2007 Park 2016, and in one Danish series it was one in three Laugesen 2017; a systematic review of studies with at least five years of follow-up put reoperation at about 14 percent Wen 2024. I would rather you hear that from me than from the internet. And what we do not have is twenty-year data. The devices have been in use since the late 1980s, but published follow-up stops at about fifteen years. So I can tell you it has lasted fifteen years in the patients studied; I cannot promise twenty.

Patients who know someone with a knee replacement ask whether this will wear out the same way. A modern knee replacement is a very good operation: 82 percent are still in place at 25 years Evans 2019, and when a knee does fail today it is usually from loosening or infection rather than the plastic wearing through Sharkey 2014. A disc replacement uses the same idea, a metal-on-plastic bearing, but it carries far smaller loads and moves through a much smaller arc than a knee. Studies of implants that were removed for other reasons show measurable but small wear, about a third of a millimetre after up to sixteen years, with some cracking at the rim of older designs Kurtz 2007; bone loss from wear particles has been reported but is rare Punt 2012. So the honest answer is: it is not immune to wear, it is a lightly loaded joint, and at fifteen years of follow-up wear has not been a common reason for removal. Removal for any reason runs about 1 in 80 Guyer 2024.

The second reason to keep motion is the discs next door. When a level is fused, its share of bending moves to the levels above and below, and over the years those discs wear faster. In the ProDisc-L trial, at five years 9 percent of replacement patients had new wear in a neighbouring disc on X-ray compared with 29 percent of fusion patients Zigler 2012. In a thousand replacement patients, 2 percent needed surgery at a neighbouring level, against about 4.5 percent reported after fusion Rainey 2012, and the pooled randomized trials show adjacent-level disease in about 4 percent versus 8 percent Patel 2026; an older systematic review found a similar gap Harrop 2008. I want to be precise here. The X-ray difference is consistent. The difference in actual neighbouring-level operations within a single trial is small, and a Norwegian study found neighbouring discs wore at the same rate as in patients who had no surgery at all Furunes 2018. So: less wear next door is well supported; a lifetime with fewer operations is likely but not yet proven beyond ten years.

The risks. The operation is done from the front, through the abdomen, which is what makes the recovery quick, but it means the large blood vessels have to be moved aside; injury to one of them happens in about 2 in 100 cases and is why an access surgeon is often part of the team Ahn 2024. In men, the nerves that control ejaculation run across the front of the spine; retrograde ejaculation, where semen flows backward into the bladder, was reported in up to 1 in 10 men in one series, and it is usually but not always temporary Lindley 2012. The implant can settle into the bone or shift, mostly in the first year, and if the level stays painful it can be converted to a fusion later. And the usual risks of any spine surgery apply: infection, blood clots, nerve irritation. Recovery: patients walk the day of surgery and go home in one or two nights; there is no brace and no waiting for bone to heal; desk work in two to four weeks; driving once off strong pain medicine; lifting limits for about six weeks, then therapy toward full activity by three months. Those are typical figures for our practice and I will tailor them to you.

Insurers approve this operation, but they check a list, and it helps to know the list before we start. Age 18 to 60, one level between L3 and the sacrum, and an FDA-approved disc: Aetna, Cigna and Carelon all require it Aetna 2026 Cigna 2026 Carelon 2026; UnitedHealthcare covers a single level with its own criteria UnitedHealthcare 2026. Original Medicare does not cover lumbar disc replacement after age 60 at all CMS 2007.

Six consecutive months of documented nonsurgical care: medication and a supervised exercise program or physical therapy, with the dates; Aetna specifically wants at least three months of in-person physical therapy within the past year, so a home program alone will not satisfy it. Recent imaging: standing and bending X-rays and an MRI, within six months for Aetna, showing a single worn disc and none of the things that rule it out, such as stenosis, a slip over three millimetres, a pars defect or scoliosis. A bone-density scan with a T-score better than minus 1.0; every policy that lists criteria requires it. Pain and function scores in the chart: Aetna wants pain of at least five out of ten, and all of them want a disability score and a description of what you cannot do. And health optimised: a haemoglobin A1c under 8 percent if you are diabetic, depression, anxiety or substance use under treatment, and no nicotine, which Carelon strongly recommends and which matters for healing in any case.

In New Jersey, Horizon sends these requests to a review company called TurningPoint, whose checklist is similar but not public Horizon 2020. Our office checks your plan before we set a date, and most of these items are things we collect during the months of nonsurgical care, so by the time we decide on surgery the file is ready.

Where we go from here. First we confirm you are a candidate: a recent MRI, standing and bending X-rays, and sometimes a CT scan or a bone-density test, to look at the disc, the facet joints, stability and bone quality. Second, the nonsurgical trial has to be finished and documented: six months of therapy, medication and, when appropriate, injections, because it is good medicine and because every insurer asks for it. Third, coverage. Most commercial plans cover a single-level disc replacement when the criteria are met. Original Medicare, however, does not cover lumbar disc replacement after age 60, and for patients 60 and under it leaves the decision to the regional contractor CMS 2007. Our office checks this before we set a date. Fourth, we decide together: replacement, fusion, or more time without surgery. Bring your questions.

These are the studies the numbers came from. Every one was checked on PubMed in October 2026. The reference page of this deck has the full list with links.

Leave this slide up during questions. The link opens this deck with the notes and the full reference list; the PowerPoint file is also available for download from the same site.