In their meta-analysis of 9 studies on the effect of antidepressive agents on the symptom of "chronic back pain," meaning "discomfort for 2 or more months" in the "back," sometimes in the "low back" and sometimes not, Salerno et al1 leave unstated the diagnoses or causes of the pain. Are we not mixing apples and oranges when collectively included are, for example, possible instances of disk disease, osteoarthritis, subacute physical or athletic strain, malignancy, partial vertebral collapse, osteoporosis, osteomalacia, rheumatoid spondylitis, or even the referred pain of visceral disease? Perhaps the therapy, which resulted in a small reduction of pain, was very effective in some but never effective in others. They conclude with a recommendation for further studies. I would suggest that such studies include diagnoses.