Abstract
Introduction
The PACE trial was the first randomized controlled trial (RCT) investigating the efficacy of paracetamol in acute low back pain. Non-compliance to study medication was considered to be a limitation of this RCT. In contrast to conventional statistical methods, complier average causal effects (CACE) analysis may provide unbiased estimates of the effects for participants compliant to paracetamol.
Methods
Intention to treat (ITT), as-treated, propensity weighted CACE and joint modeling CACE estimates were calculated for pain intensity, disability, global perceived effect and function at two weeks of follow up with compliance defined as an average of at least four tablets per day during the first two weeks of the trial. For pain intensity, exploratory analyses were conducted using additional time points and definitions of compliance.
Results
547 participants had been randomized to placebo and 550 to regular paracetamol; of the latter group, 72% of patients was classified as compliers. Mean differences in pain intensity between paracetamol and placebo using the main time point and definition of compliance were non-significant (ITT 0.11, p = 0.49; as-treated 0.29, p = 0.12; propensity weighed CACE −0.12, p = 0.51; joint modeling CACE 0.28, p = 0.13); similar results were obtained for disability, global perceived effect and function and for additional time points and definitions of compliance.
Conclusion
In compliers, paracetamol had no significant effect on pain intensity when compared to placebo; this supports the conclusions from the original analysis. However, these calculations are based on patient-reported compliance, which may not perfectly represent actual medication consumption.
Conflicts of interest: AM has received funding for a postgraduate research scholarship from GlaxoSmithKline. CM has received funding to review teaching materials prepared by GlaxoSmithKline. The other authors declare no competing interests.
Sources of funding: The PACE trial was an investigator-initiated study funded by a project grant from National Health and Medical Research Council of Australia. GlaxoSmithKline Australia provided subsequent supplementary funding and the paracetamol and matched placebo. CM is supported by Australian Research Council Future Fellowships FT-100100603. CL is supported by a Career Development Fellowship from the National Health and Medical Research Council, Australia (APP1061400). This secondary analysis of the PACE data has been supported by a program grant of the Dutch Arthritis Foundation and by the Foundation “De Drie Lichten” in The Netherlands.
