Surgery vs. observation for liver hemangioma: a systematic review and meta-analysis.

To conduct a systematic review of observational studies to evaluate effectiveness of surgery for liver hemangioma.Related studies were identified using different searching engines. Two reviewers independently extracted data on mortality, morbidity and symptoms recurrence and/or aggravation.Sixteen studies with a total of 1485 patients (402 in surgery and 1085 in observation group) were included in the analysis. Two deaths in surgical group (8.0%, 2/25) and two deaths in observation group (1.4%, 2/143) were reported. The RRs for mortality were not homogeneous (χ2=3.40, 1 d.f., P=0.07, I2=71 per cent). The RRs for morbidity were homogeneous across studies (x2=5.55, 12 d.f., P=0.94, I2=0 per cent). Morbidity in surgery group was significantly higher than that in observation group (RR=14.7, 95 per cent c.i. 9.56 to 45.63). Eight studies reported the symptom aggravation and RRs were heterogeneous (x2=31.03, 7 d.f., P<0.0001, I2=77 per cent), However, showed no statistical difference.The currently involved retrospective cohort studies of surgical series were likely to imply that surgery may take more risks than the benefits for non-emergency hemangioma patients.