Which is better? Surgery or nonsurgical care for lumbar disc herniation (LDH)? Surely by now enough studies have been done to identify the preferred choice. But, in fact, a review of five of the largest randomized controlled trials (RCTs) still couldn’t answer the question.
And the reason for that is poor study designs, high crossover rates, and large differences in the way results were measured. Crossover refers to patients who were in the surgery group and decided to switch to the nonsurgical group and vice versa.
In this report, orthopedic surgeons reviewed those five randomized controlled trials (RCTs). RCTs are supposed to offer the highest level of evidence that a specific treatment is effective. Levels of evidence are reported from Level I through Level IV. RCTs are usually level I and II studies. This means they are high-quality and offer conclusive (Level I) evidence or lesser quality but strong (Level II) research evidence.
Each of the five studies reviewed were published sometime between 1966 and 2007. Over 1,000 patients were included. The studies were all comparative. That means they compared the results of lumbar discectomy (removal of the disc) against nonoperative care.
As suggested by the name (randomized trials) each person was randomly assigned to one of these two groups. Patients in each group could have been treated either way. Anyone who needed emergency treatment (especially surgery) was not included. Everyone had painful symptoms for at least six to 12 weeks before entering the study.
The biggest red flag in these studies was the high rate of crossover. As much as 40 per cent of the surgical group decided not to have surgery and ended up in the nonsurgical group. And a full third of the nonsurgical group crossed over and had a discectomy. In one of the five studies, more than half of the patients in the nonoperative group had surgery.
When that happens, the researchers can analyze the results and report them based on the initial treatment (before they crossed over). This is called intent-to-treat (ITT) analysis. Three of the five studies approached the problem in this way. One study reported the outcomes both ways: as treatment assigned and treatment received.
Clinical outcomes and complications varied among the studies. The complication rate varied from 1.6 per cent up to five per cent. Problems during or after surgery included bleeding, sciatica, dural tears, and recurrent herniation.
The general trend was for improved results with surgery during the early follow-up. Long-term results were the same between the two groups for all measures. Measures included pain, function, and level of disability. Despite millions of dollars and multiple centers’ involvement, we still don’t have a clear answer about the best way to treat lumbar herniated disc (LDH).
The authors suggest rethinking the use of randomized controlled trials (RCTs) for patients with LDH. Improving care for patients with this diagnosis may require a different research approach. For one thing, RCTs only include patients who sign up and agree to be in the study. What about patients who don’t enroll? Would they have the same or different results? Can we apply what we learn to everyone (those who join studies and those who don’t)?
Can we really say patients are equally good candidates for either group? That’s a value judgment without evidence to back it up. And the authors mention one other point of contention. Patients who have the worst symptoms tend to improve the most no matter what group they are in. And the more severe the symptoms, the more likely that patients in the nonsurgical group would cross over to the surgical side. These factors can affect the study results.
The authors conclude by saying that the validity of these studies was seriously compromised. Randomized controlled studies may not be the best way to identify optimal treatment for lumbar disc herniation.