[Relationship between early coagulability parameters at admission and outcome in patients with severe heatstroke].

OBJECTIVE

To discuss the relationship between early coagulability parameters at admission in patients with severe heatstroke and their outcome.

METHODS

The data from 176 patients with severe heatstroke admitted to Guangzhou General Hospital of Guangzhou Military Command from January 1st, 2002 to August 31st, 2013 were retrospectively analyzed. The patients were divided into survival group (n=150) and non-survival group (n=26) according to the outcome. The incipient values of coagulability function indexes within 24 hours after admission were collected, and prothrombin time (PT), activated partial thromboplastin time (APTT) and platelet count (PLT) were compared between two groups to assess the statistically significant indexes for the analysis of the relationship between coagulability parameters and outcome with receiver operator characteristic curve (ROC curve).

RESULTS

Compared with those in survival group, PT and APTT were significantly prolonged in non-survival group [PT: 34.0 (18.8, 45.6) s vs. 18.4 (13.8, 18.0) s, Z=-6.09, P=0.000; APTT: 79.7 (41.0, 91.2) s vs. 60.8 (33.4, 41.0) s, Z=-5.08, P=0.000]. The PLT counts were significantly lower in the non-survival group than those in survival group [ 60.8(4.7, 95.3) × 10⁹/L vs. 128.4(79.8, 180.8) × 10⁹/L, Z=-4.34, P=0.000]. ROC curve analysis showed that the area under ROC curve (AUC) for PT in predicting the death of patients with severe heatstroke was 0.874, with standard error of 0.028 and 95% confidence interval (95%CI) of 0.819-0.927 (P=0.000). The best cut-off was 18.5 s, with sensitivity of 76.9% and specificity of 20.0%. AUC for APTT in predicting the death of patients with severe heatstroke was 0.812, with standard error of 0.047 and 95%CI of 0.740-0.903 (P=0.000). The best cut-off was 46.55 s, with sensitivity of 69.2% and specificity of 14.0%. AUC for PLT in predicting the death of patients with severe heatstroke was 0.767, with standard error of 0.040 and 95%CI of 0.688-0.845 (P=0.000). The best cut-off was 86.5 × 10⁹/L, with sensitivity of 68.0% and specificity of 36.8%.

CONCLUSIONS

Early prolonged PT and APTT and reduced PLT count are associated with increased risk of death, and it can predict a poor outcome in patients with severe heatstroke.