Amoxicillin-clavulanate did not reduce pain more than placebo at 12 months among patients with chronic low back pain and disc herniation, according to findings from a double-blind randomized clinical trial published in JAMA Network Open.
The community-based, placebo-controlled, two-arm superiority trial included 170 patients aged 18 to 60 years, with a mean age of 44 years. Eligible patients had chronic low back pain lasting more than 3 months and disc herniation identified on magnetic resonance imaging. Patients were randomly assigned to receive amoxicillin-clavulanate, 500 mg/125 mg, twice daily for 90 days or matching placebo. All patients continued usual care with their treating health practitioners.
The trial was conducted through Monash University, with remote recruitment and follow-up by telemedicine. Recruitment occurred from January 2016 to May 2021, and outcome data were collected through July 2022. Trial procedures were modified during COVID-19 restrictions in Victoria, Australia, including pauses in recruitment during periods of severe restriction.
Patients were excluded if they had recent antibiotic therapy, a contraindication or allergy to antibiotics, active infection, kidney disease, planned surgery within 6 months, pregnancy or breastfeeding, major coexisting illness that could confound functional assessment, or certain other conditions.
The primary outcome emphasized in the report was pain intensity at 12 months on the Low Back Pain Rating Scale, which ranges from 0 to 10 points. Pain was also assessed with a 100-mm visual analog scale. Researchers considered a 1.5- to 2-point difference on the Low Back Pain Rating Scale clinically meaningful.
At 12 months, adjusted mean pain reduction on the Low Back Pain Rating Scale was 1.0 point with amoxicillin-clavulanate vs 1.1 points with placebo, for an adjusted between-group difference of 0.06 points. Pain scores on the visual analog scale also did not differ between groups at 3 or 12 months.
Disability, measured with the Roland Morris Disability Questionnaire, did not differ between groups at 12 months. The placebo group had numerically greater improvement in disability and some psychological measures at 3 months, but those differences were not sustained at 12 months and were not significant after imputation. Researchers also reported no between-group differences at 12 months in work absence, work hindrance, global improvement, general health status, depression, or fear of movement or reinjury.
The findings add to mixed evidence on antibiotic therapy for chronic low back pain. Earlier studies differed by antibiotic regimen and patient subgroup, including whether patients were selected for Modic changes—bone changes detected on magnetic resonance imaging. Two earlier trials of amoxicillin-clavulanate reported benefit in patients with chronic low back pain and Modic changes, whereas a trial of amoxicillin alone found no benefit.
In the current trial, prespecified subgroup analyses did not show a pain benefit among patients with type 1 or type 2 Modic changes. Modic changes were present at baseline in 40 patients assigned to amoxicillin-clavulanate and 52 patients assigned to placebo, an uneven distribution in the subgroup of interest.
However, the trial was not primarily powered to test efficacy in patients with Modic changes and did not specifically select for those changes. Most patients with Modic changes had type 2 changes rather than isolated type 1 changes—the subgroup that helped drive earlier interest in antibiotic therapy. As a result, the findings weigh against routine antibiotic use in this population but do not fully resolve whether a narrower group of patients with type 1 Modic changes could respond differently.
Adherence was another important consideration. Among 152 patients who completed 12-month outcomes, 128 provided adherence data through a mix of pill counts and self-report. By pill count, 83% of patients in the amoxicillin-clavulanate group and 80% of patients in the placebo group took more than 80% of the prescribed study medication. Self-reported adherence was higher. Researchers also conducted a complier average causal effect analysis, which estimated the treatment effect among patients who took the medication as prescribed, and the results still showed no between-group differences in pain intensity, disability, work absence, or work hindrance.
Adverse events were reported by 34 patients receiving amoxicillin-clavulanate and 20 patients receiving placebo. Three serious adverse events occurred: allergic reaction and radiculopathy in the amoxicillin-clavulanate group and diarrhea in the placebo group. The allergic reaction and diarrhea—one serious adverse event in each group—were assessed as possibly related to study treatment, and both patients discontinued medication but completed follow-up. The radiculopathy event was deemed unrelated. Medication was discontinued because of adverse events by 14 patients receiving amoxicillin-clavulanate and 2 patients receiving placebo.
The researchers noted that most patients were recruited from the community through social media, advertising, and other sources rather than from hospitals or specialty spine centers. This may limit generalizability to patients with more complex chronic low back pain who are seen in specialty settings and for whom clinicians may be more likely to consider antibiotics.
Other limitations included missing 12-month primary outcome data for 18 patients, discontinuation of planned antimicrobial resistance sampling because of recruitment challenges, and refinement of the primary outcome from the original unpublished protocol. The researchers noted that the published protocol specified a single primary outcome.
The findings may be clinically relevant because antibiotic use for chronic low back pain has drawn interest despite limited and conflicting evidence, and broader prescribing could contribute to antimicrobial resistance. The results support caution against routine antibiotic use for chronic low back pain with disc herniation, while leaving some uncertainty about narrower subgroups such as patients with type 1 Modic changes.
Disclosures can be found in the study.
Source: JAMA Network Open
