LAUSR

Editor We read with great interest the systematic review and meta-analysis by Haisley et al.1 investigating perioperative risk factors associated with the development of pressure injuries (PIs). Here, we offer some discussion points regarding their conclusions. Haisley et al. concluded that the use of corticosteroids had no statistically significant association with development of PI. We disagree. Our previous research found that perioperative corticosteroid use was an independent risk factor for PI in surgical patients (odds ratio (OR) 2⋅808, 95 per cent c.i. 1⋅062 to 11⋅769)2. In this meta-analysis, the authors included only three studies with 1218 patients; our study2 was not included in the pooled result. When incorporated into our study2, we reperformed a meta-analysis and found a significant association (risk ratio (RR) 1⋅09, 95 per cent c.i. 1⋅02 to 1⋅16) between the use of corticosteroids and PI (Fig. S1, supporting information). Therefore, we think perioperative corticosteroid administration was an important factor for development of PI. The authors concluded that cardiovascular disease and respiratory disease should be taken into consideration for the development of PI. We propose that some other surgical types might be as important, such as hepatobiliary surgery. The results of our recent research confirm this view; 8⋅6 per cent (95 per cent c.i. 4⋅4 to 14⋅9) of 128 patients undergoing hepatobiliary surgery developed PI3. Similarly, the meta-analysis also revealed hepatobiliary surgery was prone to PI (Chen et al.26 in Haisley et al.). We agree with the authors’ view that longer duration of surgery was an important risk factor for PI. Our previous research also confirmed this through propensity score matching (95 versus 240 min; P = 0⋅003)4. Further, our other meta-analysis (Fig. S2, supporting information) revealed a significant dose–response association between duration of surgery and PI (model χ2 = 9⋅29; P = 0⋅000)5. Nevertheless, we should emphasize we found that surgery-related PIs are associated with duration of surgery but have no significant relationship with length of cardiopulmonary bypass (P = 0⋅830)4.