Fibromyalgia and Duloxetine: A Cluster Analysis

October 18, 2009
Bradley Schmidt

An international research team believes that its cluster analysis tool can be used to identify effective treatment modalities for a host of maladies using numerous different therapies.

During an afternoon abstract session, an international team of researchers laid out the methodology behind a cluster analysis that they used to determine the efficacy of duloxetine for the improvement of pain in fibromyalgia patients (“Typology of Patients with Fibromyalgia: Cluster Analysis of Duloxetine Study Patients”).

More important than the conclusions that they reached, say the authors, is the means used to arrive there. They believe that this tool can be used to identify effective treatment modalities for a host of maladies using numerous different therapies.

Using a hierarchical clustering analysis, the researchers separated individuals according to 18 variables, with a particular focus on domains of pain, mental impairment, global impression rating and overall functioning because the criteria were used in all five studies from which they culled data, allowing researchers to assess change without data loss.

Researchers scrutinized data assessing variability from baseline to post-treatment within and across clusters, over a 52-week period.

Using the Pain Interference Scale and Fibromyalgia Impact Questionnaire to rank subject pain scores, the authors categorized patients into five clusters based on pain severity:

  • Cluster 1 patients produced the highest pain scores across the board
  • Cluster 2 patients were characterized by moderately high pain scores but had the highest score on the SF-36 physical impairment sub scale
  • Cluster 3 was defined by high mental impairment scores on measures of depression and anxiety (SF-36), and moderate levels of pain, with moderate scores on global depression and overall functioning
  • Cluster 4 displayed moderate outcomes, but produced better outcome scores than patients in the first three clusters across scales
  • Cluster 5 displayed the very best scores of all groups

Definition of Score

  • The best-scoring subjects were defined as those who had scores of less than 3.29 on the pain interference scale and scores below two on the FIQ work interference scale were categorized.
  • Moderate categories scored higher than the best groupings on the Work Interference.
  • Individuals in the worst category had scores on the Pain Interference Scale above 3.29, and scores above 7.14 on the Interference Scale.
  • Those categorized as mentally and physically impaired scored less than 7.14 on the interference scale, but had either high scores on the FIQ depression scale (mentally impaired), or low scores (physically impaired).

OutcomesSubjects within each of the worst three categories showed greater reductions in group membership on duloxetine compared with those taking placebo.

Compared to those taking placebo, those in the duloxetine arm ended up transitioning in greater numbers to the physically and mentally impaired categories.

Those defined by the physical impairment category displayed lower stability when taking duloxetine, and saw a greater transition to moderate and good categories than those taking placebo.

Those in moderate categories on duloxetine had greater stability compared to placebo, and fewer regressed from moderate to mentally impaired.

Overall, the greatest positive change came for patients on duloxetine who were in the worst, physically impaired, and mentally impaired categories. Researchers did not witness much change in either the moderate and best categories between duloxetine and placebo.

A duloxetine regimen resulted in higher improvement rates compared to placebo, and duloxetine either maintained or improved health status for subjects in all categories.

Practical implicationsResearchers indicate that the greatest benefit that clinicians can derive from their findings is through applying the same methods in different studies to achieve similar results

Clinicians should be able to use the cluster classifications in different and evolving ways to measure changes in outcomes across different classifications, allowing them to enhance treatment planning for their fibromyalgia patients.