Opioid Systematic Review

Methods

There were 21 papers which included 41 unique study groups. All papers were observational studies. As result, this meta analysis provides evidence of association but not causation. We had three primary research questions:

What is the prevalence of opioid use after surgery?
Is pre-surgery opioid use associated with post-op opioid use?
Does surgery influence opioid use?

Research question one included 11 papers with 24 study groups. We used a random/mixed effects model to calculate the post-operation prevalence of opioid use for patients receiving cervical fusions, lumbar fusions or non-fusions.

Research question two included four papers with four study groups. We used a random/mixed effects model to calculate the odds ratio for post operation opioid use depending if the patient was a pre-operation opioid user or non-user.

Research question three included 12 papers with 19 study groups. We used a random/mixed effects model to calculate the risk difference in opioid before and after surgery.

Results

Research question 1

What is the prevalence of opioid use after surgery?

There were 11 papers and 24 unique study groups with sufficient data to answer this research question. There were three papers and five study groups for non-fusions, five papers and 12 study groups for lumbar fusions, and four papers and seven study groups for cervical fusions. For all studies, the prevalence range of post-operation opioid use was from 0.06 to 0.85 (median: 0.31). For non-fusion patients, the prevalence range of post-operation opioid use was from 0.06 to 0.58 (median: 0.22). For lumbar fusion patients, the prevalence range of post-operation opioid use was from 0.29 to 0.85 (median: 0.32). For cervical fusion patients, the prevalence range of post-operation opioid use was from 0.12 to 0.43 (median: 0.14)

Figure 2: Forest plot of the prevalence of post-operation opioid use for non-fusion, lumbar fusion and cervical fusion procedures

Research question 2

Is pre-surgery opioid use associated with post-op opioid use?

There were four papers and four unique study groups with sufficient data to answer this research question. The only surgery type in this analysis were non-fusions.

Figure 3: Forest plot of the odds of being an opioid user after surgery if patient was a pre-operation opiod user compared to pre-operation non-opioid user

Research question 3

What is the impact of surgery on opioid use?

There were 12 papers and 19 unique study groups which had sufficient data to answer this research question. There were three papers and six unique study groups for cervical fusions, three papers and seven unique study groups for lumbar fusions, and six papers and six unique study groups for non-fusions. Figure 4 provides a summary of all the studies used in this section. The summary statistic in figure 4 is the risk difference between opioid users pre and post operation.

Figure 4: Forest plot of opioid prevalence pre and post operation for cervical fusion, lumbar fusion and non-fusion cases

Figure 5 compares the prevalence of opioid use prior to surgery to the prevalence of opioid use after surgery based on surgery type. The surgeries included are cervical fusions, lumbar fusions and non fusions. Prior to surgery, the min, median and max for proportion of opioid used in patients was 0.17 , 0.34, and 0.49 for cervical fusions, 0.18 , 0.22, and 1 for lumbar fusions, and 0.33 , 1, and 1 for non fusions. After surgery, the min, median and max for proportion of opioid used in patients was 0.12 , 0.22, and 0.43 for cervical fusions, 0.29 , 0.3, and 0.33 for lumbar fusions, and 0.13 , 0.31, and 0.95 for non fusions.

Figure 5: Boxplots Option A

Figure 5: Boxplots Option B