Condition: Ischias

Relief of leg pain and the rate of perceived recovery were faster with early microdiscectomy than with prolonged conservative care (recovery hazard ratio 1.97, 95% CI 1.72 to 2.22). There was no significant overall difference in disability during the first year (P=0.13), and about 95% in both groups reported recovery at one year. Measured in: 283 adults with severe sciatica for 6 to 12 weeks, randomized to early surgery or conservative care with surgery if needed. 39% of the conservative group ended up having surgery within the year, so this compares early surgery with a strategy of delay-then-operate-if-needed, not with avoiding surgery altogether.

Most sciatica caused by a herniated lumbar disc improves over weeks to a few months as the disc material shrinks and the nerve irritation settles, and imaging studies show that herniated disc fragments often regress spontaneously over that time. Surgery is reserved mainly for severe or persistent cases and for progressive neurological signs. Measured in: Clinical review synthesizing natural history, imaging and management evidence for lumbar disc sciatica. This is a narrative review, so it summarises the field, not pooling results to a single number, and the pace of recovery varies widely from person to person.

Both groups improved substantially over two years. In the intention-to-treat analysis the differences between surgery and non-operative care were small and not statistically significant for the primary outcomes, though very high crossover in both directions blurred the comparison. Measured in: 501 surgical candidates (mean age 42, 42% women) with imaging-confirmed lumbar disc herniation and radiculopathy for at least 6 weeks. Only about half the surgery group had surgery promptly and many in the non-operative group crossed over to surgery, so the intention-to-treat result understates the effect of the operation itself; as-treated analyzes favored surgery.

Pooling cohort studies, people who had surgery for sciatica still reported meaningful leg and back pain and disability an average of five years later, with pain scores that had improved from baseline but not resolved. Recovery is real but often incomplete over the long run. Measured in: 39 cohort studies (40 records; 13,883 patients) of people followed up to 5 years after surgery for sciatica. What could explain it instead: Selection by severity. People who come to surgery tend to have started with worse or more stubborn sciatica, so higher residual symptoms may reflect who gets operated on, not the operation.. These are single-arm cohorts without a comparison group, so they describe long-term status after surgery, not proving surgery caused the residual symptoms or that non-surgical care would differ.

At 8 weeks the leg-pain intensity score was 3.7 with pregabalin and 3.1 with placebo (adjusted mean difference 0.5, 95% CI -0.2 to 1.2, P=0.19), with no difference at 52 weeks and no benefit on disability or other secondary outcomes. Adverse events, dizziness in particular, were more common with pregabalin. Pain was rated on a 0 to 10 scale. Measured in: 209 adults with moderate-to-severe sciatica, randomized to pregabalin up to 600 mg/day or matching placebo. This tested pregabalin specifically at up to 600 mg/day over 8 weeks; it does not speak to every possible dose or duration, but it is a well-conducted placebo-controlled trial of the standard use.

Across pooled trials, gabapentinoids and other anticonvulsants gave no clinically meaningful reduction in low back or lumbar radicular pain versus placebo, with high-certainty evidence of increased adverse events such as dizziness, fatigue and difficulty concentrating. Measured in: 9 randomized trials of anticonvulsants for chronic low back pain or lumbar radicular pain. The included trials varied in the exact drug, dose and pain definition, and most were of moderate size, so the pooled estimate is more solid for the harms than for ruling out every small benefit.

Epidural corticosteroid injections produced a small reduction in leg pain and disability in the short term compared with placebo, an effect that was not maintained at longer follow-up. Serious harms were uncommon in the trials. Measured in: 25 randomized trials of epidural corticosteroid injections for lumbosacral radicular pain. The short-term average benefit was small and below what many would call clinically important, and it did not last, so it eases severe pain briefly and is not a lasting fix.

At week 4, leg pain on a 100 mm scale fell 30.8 mm with acupuncture versus 14.9 mm with sham (between-group difference -16.0 mm, 95% CI -21.3 to -10.6), and disability improved more with acupuncture (Oswestry difference -8.1 points). The separation appeared by week 2 and persisted through week 52. No serious adverse events occurred. Measured in: 216 adults with chronic sciatica from a herniated disc, randomized to 10 sessions of acupuncture or sham over 4 weeks. This was a single trial conducted in Chinese hospitals with experienced acupuncturists, so the size of the effect may not transfer to every setting, and blinding of a needling sham is imperfect.

In a network meta-analysis spanning disc surgery, epidural injection, non-opioid drugs, biological agents and conservative care, no single strategy stood out as clearly superior across outcomes, and much of the evidence base was of low quality with wide uncertainty. Measured in: 122 comparative studies (90 of them randomized or quasi-randomized trials) of treatments for sciatica, analyzed by network meta-analysis. Indirect comparisons across trials that differed in patients, severity and outcome timing carry real uncertainty, so the lack of a standout reflects both actual similarity and weak data.

Pooled trials found non-steroidal anti-inflammatory drugs no more effective than placebo for overall pain in sciatica, with only very low to low certainty evidence and small, inconsistent effects on global improvement. Measured in: 10 randomized trials of NSAIDs in people with sciatica. The trials were mostly old and small with a high risk of bias, so this reflects thin evidence, not a strong finding, and short-term symptom relief for some individuals is not excluded.

Pooling randomized trials, acupuncture was associated with greater pain reduction and higher overall effectiveness for sciatica than control treatments, but most included trials were small and at high risk of bias. Measured in: Randomized trials of acupuncture for sciatica pooled in a systematic review and meta-analysis. The included trials were largely small, single-country and methodologically weak, so the pooled benefit should be read as promising, not settled.

Across cohort studies, more intense or longer-lasting leg pain at baseline, greater psychological distress and worse general health predicted a poorer recovery from sciatica managed without surgery, while findings for the size of the disc herniation itself were inconsistent. Measured in: 14 original cohorts (reported in 23 articles) of prognostic factors in non-surgically treated sciatica. The primary studies measured different factors in different ways and were of variable quality, so these are consistent signals, not a validated prediction rule.

Pooling diagnostic studies, the straight-leg-raise test had a sensitivity of 0.92 (95% CI 0.87 to 0.95) but a specificity of only 0.28 (95% CI 0.18 to 0.40) for lumbar disc herniation in surgical populations, so a negative test makes a herniation unlikely while a positive one is weak confirmation. The crossed straight-leg-raise test reversed the pattern, with specificity 0.90 (95% CI 0.85 to 0.94) and sensitivity 0.28 (95% CI 0.22 to 0.35). Most other single physical tests performed poorly on their own. Measured in: 16 cohort studies and 3 case-control studies comparing physical-examination tests against imaging or surgical findings for lumbar disc herniation. Most of the studies were done in surgical populations where disc herniation was very common, so the numbers may read differently in primary care, and combining tests performs better than any one alone.