Objective: To explore the characteristics of inpatients with suicide risk from liaison-consultation records at a general hospital.Methods: Liaison-consultation records in a general hospital from July 2009 to March 2011 were investigated.Mental disorders were diagnosed by attending doctors or professors according to DSM-Ⅳ.Suicide risk including suicide ideation, suicide attempt and suicide gesture were classified by consultant doctors.Patients with suicide risk were compared with other patients without suicide risk.Results: There were 1203 patients interviewed by professional psychiatrists, among which105 patients had suicide risk (78 of them had suicide ideation; 12 had suicide attempt; three had suicide gesture; and seven had suicide attempt history).The suicide risk was rated by consultants as mild (56), moderate (28), severe (12)and very severe (3).Bivariate analyses showed that patients with suicide risk were more likely to be elder than 65 (Odds ratio =5.256, 95% Cl =3.417-8.083), coming from Gynaecology and Obstetrics wards (Odds ratio =2.499, 95% Cl =1.005-6.216), or have depressive disorders (Odds ratio =5.256, 95% Cl =3.417-8.083).They were less likely to have past history of mental disorders (Odds ratio =0.565, 95% Cl =0.377-0.848) and be classified as mental disorders due to general medical condition (Odds ratio =0.404, 95% Cl =0.193-0.846), medically unexplained symptoms (Odds ratio =0.115, 95% Cl =0.016-0.831) and no mental disorder (Odds ratio =0.494, 95% Cl =0.245-0.998).No significant differences were found in gender, anxiety and stress related disorders or other mental disorders.More pharmacotherapy (X2=20.709, P<0.000) and psychological intervention (X2=20.672, P<0.000) were conducted to this group compared to the other.There was no significant difference about referral to psychiatric hospital/wards.Conclusion: Suicide risk is common in consultation at general hospital.Elder inpatients and patients with depression appear to be at higher risk.Limitations of the study include lack of detail information about suicide risk and the non-structured clinical interviews.