Once Upon a (Medium-Firm) Mattress

Abstract & Commentary

Source: Kovacs FM, et al. Effect of firmness of mattress on chronic non-specific low-back pain: Randomised, double-blind, controlled, multicentre trial. Lancet. 2003;362: 1599-1604.

Right up there alongside "eating an apple a day" and "not walking outside with wet hair on a cold day" is the other widely held "medical truth" that "a firm mattress is good for your back." While the first 2 tenets are largely discounted by anyone with some medical training, the virtues of a firm mattress are generally espoused by the medical and lay community alike—that is, despite the utter lack of scientific evidence to support it. In an attempt to shed light on this poorly understood subject, Kovacs and associates assess the effect of the firmness of a mattress on low back pain.

In a double-blind, multicenter trial, 313 adults with chronic nonspecific low back pain were randomized (with 310 patients available in the intention-to-treat analysis) to use a firm mattress vs a medium-firm mattress. Inclusion criteria included adults older than 18 with nonreferred low back pain not related to an underlying disorder such as trauma, cancer, or infection. Patients taking analgesics dosed in the evening or lasting 24 hours were excluded. Mattress firmness was based upon the European Committee for Standardization rating scale. Patients were randomly assigned to a firm mattress (scored 2-3) or a medium mattress (scored 5-6). Outcomes were measured at 90 days with the primary end points including pain while lying in bed; pain upon arising; and disability. Self-assessment of pain was made using the visual analogue scale (VAS), and disability was assessed with the Roland Morris 24-item disability questionnaire.

The results at 90 days showed that both groups experienced improvements compared to baseline. As for comparisons between groups, on the "pain while lying in bed" measurement, the firm group improved 70% compared to an 80% improvement in the medium-firm group (P = ns); both improved 57% with respect to "pain upon rising" and patients in the medium-firm group had a superior improvement in disability of 30% vs 50% (P = .008). Kovacs et al note that 77% of patients in the firm group accurately perceived the firmness of their mattress, whereas 43% of the medium-firm group also believed they were on firm mattresses. If this confounding variable is removed, the improvement of the medium-firm mattress over the firm nearly doubles.


Consider that in order to do this study, Kovacs et al had to actually purchase new firm or medium-firm mattresses for each subject. At an average price of $500, this low-tech experiment ending up costing a minimum of $150,000, not counting a group purchase discount. But the point remains that such a seemingly simple study is in practice very hard to do. As with many provocative studies, more questions are raised than answered. What were the actual firmness measures of the baseline mattresses and did the degree of increase in firmness correlate best with improvement? Indeed many more issues can be raised, but for now let us consider the possibility that an apple a day doesn’t necessarily keep the doctor away, that you don’t catch a cold from being cold, and that a medium-firm mattress that feels more comfortable may in fact be more comfortable. — Jeffrey Reich, Assistant Professor of Neurology, New York Presbyterian Hospital- Cornell Campus; Assistant Editor, Neurology Alert.