BACKGROUND: Data describing the coagulopathy of Crimean-Congo haemorrhagic fever are scarce. We did rotational thromboelastometry (ROTEM) and conventional coagulation testing in patients with Crimean-Congo haemorrhagic fever to increase our understanding of the coagulopathy of this infectious disease. METHODS: We did a prospective observational cohort study of adults aged 18 years and older and admitted to hospitals with PCR-confirmed Crimean-Congo haemorrhagic fever in Samsun and Tokat, Turkey. Demographic, clinical, and laboratory data were collected and blood samples for ROTEM analysis and coagulation testing were drawn at admission and during hospital admission and convalescence (up to 30 days after onset of illness). For the ROTEM analysis we recorded the following extrinsically activated ROTEM (EXTEM S) variables, with normal ranges indicated: clotting time (38-79 s), clot formation time (34-159 s), amplitude at 10 min after clotting time (43-65 mm), maximum clot firmness (50-72 mm), and maximum lysis (>15% at 1 h). The following fibrin-specific ROTEM (FIBTEM S) variables were also recorded: amplitude at 10 min after clotting time (normal range 7-23 mm) and maximum clot firmness (9-25 mm). Disease severity was assessed by Swanepoel criteria, severity grading score (SGS), and the severity scoring index (SSI), with mild disease defined as meeting no Swanepoel criteria, graded mild by SSI, and graded low risk by SGS. FINDINGS: Between May 27, 2015, and Aug 2, 2015, 65 patients with confirmed Crimean-Congo haemorrhagic fever were recruited and had blood taken at 110 time points. Most were male (40 [62%] of 65) with mild disease (49 [75%] of 65). Haemorrhage occurred in 13 (20%; 95% CI 11·1-31·8) of 65 patients and 23 (35%) of 65 received blood products (15 received fresh frozen plasma and eight received red blood cell concentrates), and 21 patients received platelet transfusions. At admission, the following EXTEM S variables differed significantly between mild cases and moderate to severe cases: median clotting time 56 s (range 42-81; IQR 48-64) versus 69 s (range 48-164; IQR 54-75; p=0·01); mean amplitude at 10 min after clotting time 45·1 mm (SD 7·0) versus 33·9 mm (SD 8·6; p<0·0001); median clot formation time 147 s (range 72-255; IQR 101-171) versus 197 s (range 98-418; IQR 156-296; p=0·006); and maximum clot firmness 54·4 mm (SD 7·2) versus 45·1 mm (SD 12·5; p=0·003). The EXTEM S variables were compared at different time points; maximum clot firmness (p=0·024) and amplitude at 10 min after clotting time (p=0·090) were lowest on days 4-6 of illness. We found no significant differences in FIBTEM variables between mild and moderate to severe cases (median amplitude at 10 min, 13 mm [range 8-20; IQR 11-15] vs 12 mm [range 6-25; IQR 10-15; p=0·68]; and median maximum clot firmness, 15 mm [range 9-60; IQR 13-21] vs 17 mm [range 7-39; IQR 13-23; p=0·21]); and no hyperfibrinolysis (maximum lysis >15%). INTERPRETATION: Coagulopathy of Crimean-Congo haemorrhagic fever is related to defects in clot development and stabilisation that are more marked in severe disease than in mild disease. The combination of normal and slightly deranged coagulation screens and FIBTEM results with the absence of hyperfibrinolysis suggests that the coagulopathy of Crimean-Congo haemorrhagic fever relates to platelet dysfunction. FUNDING: Wellcome Trust, UK Ministry of Defence, and National Institute for Health Research Health Protection Research Unit.