Key messages

• Duloxetine leads to a slight decrease in pain and improvement in functionality in osteoarthritis, but is associated with frequent side effects. 
• The benefit/risk ratio is probable not favorable. 

What is the evidence.
See evidence matrix  in Epistemonikos later

We found four systematic reviews [1],[2],[3],[4] including four primary studies, reported in six references [5],[6],[7],[8],[9],[10]. All correspond to randomized controlled trials.

What types of patients were included*

Three trials focused on osteoarthritis of the knee [5],[6],[8] and one did not specify it [7].

Average WOMAC (The Western Ontario and McMaster Universities Osteoarthritis Index) ranged between 50 and 57 in three trials [5],[6],[8] and one did not report it [7].

Average age of the participants ranged between 61 and 69 years in the different trials.

The proportion of women was between 16 and 77% in the different trials.

What types of interventions were included*

All of the trials used duloxetine in oral presentation.

One trial used 60 mg/day as starting dose, which was increased to 120 mg/day at the seventh week in patients that reported a reduction in pain less than 30% [5].

One trial used 60 mg [6].

One trial used a flexible dosage between 60 and 120 mg per day, keeping a baseline treatment with NSAIDs at therapeutical dose [7].

One trial had two active arms, with 60 mg and 120 mg per day [8].**

All trials allowed the use of concomitant analgesia.

All trials compared against placebo.

What types of outcomes
were measured

The outcomes, as classified in the identified systematic reviews, were as follows: effect on the intensity of pain (analyzed as a decrease or improvement); overall impression of the patient-reported improvement; sub-scale of physical functionality established by the Western Ontario and McMaster Universities Arthritis Index (WOMAC); and adverse side effects.

To whom this evidence does and does not apply

• While the intention of this summary was to include osteoarthritis in general, considering all of the possible affected joints, the trials that were found focused specifically on osteoarthritis of the knee. However, in the absence of direct evidence for other joints, it appears reasonable to extrapolate the conclusions of this summary.

• The outcomes included in this summary are those we consider critical for making adequate clinical decisions. Furthermore, we conducted a search of the Core Outcome Measures in Effectiveness Trials (COMET) Initiative, finding one relevant article [11] that indicated, through a consensus, the following as the most relevant outcomes: joint pain, functionality, quality of life in relation to health, work situation, mortality, reoperation, hospital readmission, and overall satisfaction with the results of the treatment.

• While the observed benefits on pain and functionality were statistically significant, the magnitude of the effect was below the minimally important difference reported in the literature [12].
• Considering there are substantial side effects, the balance between benefits and harms is probably unfavorable.

• Although duloxetine is widely available in most countries, the cost is usually high. Taking into account the reported small or irrelevant benefits, the increase in side effects, and the cost, the balance between cost and benefit is clearly unfavorable.

• Faced with the evidence presented in this summary, most patients and attending clinicians should lean against the use of this intervention. 
• However, without clearly effective therapeutic alternatives, variations could exist in the clinical decisions made by individual patients.
• Recommendations against the use of this intervention in the main clinical guidelines should minimize variability in the decisions made by physicians.

Differences between this summary and other sources

• The reviews included in this summary differ among one another. One review [1] concluded there are no differences between duloxetine and other second-line oral treatments. Two reviews [2],[3] reported duloxetine would be better than placebo at reducing pain and improving functionality, with the adverse side effects considered acceptable for treating osteoarthritis pain. Finally, another review [4] determined the effect of duloxetine in pain management is conservative and that the generated effect is likely influenced by the clinical-practice experience of the professional recommending treatment (the authors base this conclusion on the abnormal effect distribution, i.e., few patients had an average response and the average was not an appropriate descriptor).
• The conclusions of this summary partially disagree with the main international guidelines for osteoarthritis: the Osteoarthritis Research Society International (OARSI) [13] indicates duloxetine would be useful as treatment for osteoarthritis of the knee in patients without comorbidities and in patients presenting osteoarthritis in multiple joints. The American Academy of Orthopaedic Surgeons (AAOS) [14] does not mention the use of duloxetine for osteoarthritis.

• The probability is low that the conclusions of this summary change with future trials, due to the high certainty of the existing evidence. 
• At least one published trial exists that was not included in the identified systematic reviews [15]. Future reviews incorporating this trial could provide new information. 
• We did not identify any ongoing trial in the International Clinical Trials Registry Platform of the World Health Organization.