Prevalence, predictors and outcome of hypofibrinogenaemia in trauma: a multicentre observational study

Abstract

Introduction: Exsanguination due to trauma-induced coagulopathy is a continuing challenge in emergency trauma care. Fibrinogen is a crucial factor for haemostatic competence, and may be the factor that reaches critically low levels first. Early fibrinogen substitution is advocated by a number of authors. Little evidence exists regarding the indications for fibrinogen supplementation in the acute phase. This study aims to estimate the prevalence of hypofibrinogenaemia in a multi-center trauma population, and to explore how initial fibrinogen concentration relates to outcome. Also, factors contributing to low fibrinogen levels are identified.

Methods: Patients arriving in hospital less than 180 minutes post-injury requiring full trauma team activation in four different centers were included in the study. Time from injury, patient demographics, injury severity scores (ISS) and 28 days outcome status were recorded. Initial blood samples for coagulation and blood gas were analyzed. Generalized additive regression, piecewise linear regression, and multiple linear regression models were used for data analyses.

Results: Out of 1,133 patients we identified a fibrinogen concentration ≤1.5g/L in 8.2%, and <2 g/L in 19.2%. A non-linear relationship between fibrinogen concentration and mortality was detected in the generalized additive and piecewise linear regression models. In the piecewise linear regression model we identified a breakpoint for optimal fibrinogen concentration at 2.29 g/L (95% confidence interval (CI): 1.93 to 2.64). Below this value the odds of death by 28 days was reduced by a factor of 0.08 (95% CI: 0.03 to 0.20) for every unit increase in fibrinogen concentration. Low age, male gender, lengthened time from injury, low base excess and high ISS were unique contributors to low fibrinogen concentrations on arrival.

Conclusions: Hypofibrinogenaemia is common in trauma and strongly associated with poor outcome. Below an estimated critical fibrinogen concentration value of 2.29 g/L a dramatic increase in mortality was detected. This finding indicates that the negative impact of low fibrinogen concentrations may have been previously underestimated. A number of clinically identifiable factors are associated with hypofibrinogenaemia. They should be considered in the management of massively bleeding patients. Interventional trials with fibrinogen substitution in high-risk patients need to be undertaken.

Figure 1

Multivariable generalised additive model and piecewise linear model for relationship between fibrinogen concentration and 28-day survival. Results from the multivariable generalised additive model (GAM) and the piecewise linear model for the relationship between fibrinogen concentration and 28-day survival, adjusted for Injury Severity Score, age, time from injury, mechanism of injury, base excess, International Normalized Ratio, platelet count and gender. The functional relationship is clearly nonlinear (a), resulting in a corresponding nonconstant odds ratio across the observed range of fibrinogen values (b). For the piecewise linear model, the breakpoint (95% confidence interval (CI)) is estimated at 2.29 (1.93, 2.64).