BACKGROUND: Dose adjustment or laboratory monitoring of low-molecular-weight heparin (LMWH) is commonly recommended for patients with severe renal insufficiency (creatinine clearance < or =30 mL/min), but the basis for this recommendation is unclear. PURPOSE: To compare levels of anti-Xa heparin and risk for major bleeding in LMWH-treated patients with a creatinine clearance of 30 mL/min or less versus those with a creatinine clearance greater than 30 mL/min by using standard weight-adjusted therapeutic doses, empirically adjusted doses, or prophylactic doses of LMWH. DATA SOURCES: Electronic databases (MEDLINE, EMBASE, and the Cochrane Library) searched to December 2005 with no language restrictions. The authors also searched reference lists and contacted experts. STUDY SELECTION: Observational or subgroups of randomized studies that included non-dialysis-dependent patients with varying degrees of renal function who were treated with LMWH and reported creatinine clearance and anti-Xa levels or major bleeding. DATA EXTRACTION: Two reviewers independently selected studies and extracted data on patient characteristics, renal function, LMWH treatment, anti-Xa levels, and major bleeding. The pooled odds ratio of major bleeding in patients with a creatinine clearance of 30 mL/min or less was calculated by using the Peto method. DATA SYNTHESIS: Eighteen studies using 3 preparations of LMWH (15 studies using enoxaparin, 2 using tinzaparin, and 1 using dalteparin) were included. Peak anti-Xa levels measured 4 hours after a subcutaneous injection were statistically significantly higher in patients with a creatinine clearance of 30 mL/min or less compared with those with a creatinine clearance greater than 30 mL/min in studies that used a standard therapeutic dose of enoxaparin (4 studies) but not in studies of empirically dose-adjusted enoxaparin (3 studies). Data were insufficient to assess the relationship between anti-Xa and renal function for prophylactic doses of enoxaparin and therapeutic doses of tinzaparin or dalteparin. In 12 studies involving 4971 patients, LMWH was associated with a statistically significant increase in the risk for major bleeding in patients with a creatinine clearance of 30 mL/min or less compared with those with a creatinine clearance greater than 30 mL/min (5.0% vs. 2.4%; odds ratio, 2.25 [95% CI, 1.19 to 4.27]; P = 0.013). When analyzed according to LMWH preparation, major bleeding was increased when a standard therapeutic dose of enoxaparin was used (8.3% vs. 2.4%; odds ratio, 3.88 [CI, 1.78 to 8.45]) but may not be increased when an empirically adjusted dose of enoxaparin is used (0.9% vs. 1.9%; odds ratio, 0.58 [CI, 0.09 to 3.78]; P = 0.23 for heterogeneity). There were insufficient studies to assess the risk for major bleeding with tinzaparin, dalteparin, and prophylactic doses of enoxaparin. LIMITATIONS: The data for tinzaparin and dalteparin were limited. Data are observational, and the potential for confounding cannot be excluded. CONCLUSIONS: Non-dialysis-dependent patients with a creatinine clearance of 30 mL/min or less who are treated with standard therapeutic doses of enoxaparin have elevated levels of anti-Xa and an increased risk for major bleeding. Empirical dose adjustment of enoxaparin may reduce the risk for bleeding and merits additional evaluation. No conclusions can be made regarding other LMWHs.