Fibrinogen; a predictor of injury severity and mortality among patients with traumatic brain injury in Sub-Saharan Africa: a prospective study

Introduction Fibrinogen levels drop quicker than any other factors in severe trauma such as Traumatic Brain Injury (TBI). Contemporaneous studies show that fibrinogen concentrations < 2 g/L are strongly related to mortality. However, little is known regarding fibrinogen levels and TBI severity as well as mortality in sub-Saharan Africa. We therefore set out to determine whether fibrinogen levels are associated with TBI severity and seven days outcomes. Objectives To determine the sensitivity and specificity of fibrinogen levels and the association with severity and mortality among TBI patients at Mulago Hospital. Methods We prospectively enrolled 213 patients with TBI aged between 13 and 60 years of age and presenting within 24hrs of injury. Patients with pre-existing coagulopathy, concurrent use of anticoagulant or antiplatelet agents, pre-existing hepatic insufficiency, diabetes mellitus and who were pregnant were excluded. Fibrinogen levels were determined using the Clauss fibrinogen assay. Results Majority of the patients were male (88.7%) and nearly half were aged 30 or less (48.8%). Fibrinogen levels less than 2g/L were observed in 74 (35.1%) of the patients while levels above 4.5 g/L were observed in 30(14.2%) of the patients. The average time spent in the study was 3.7 ± 2.4 days. The sensitivity and specificity using fibrinogen < 2g/L was 56.5% and 72.9% respectively. Fibrinogen levels predict TBI severity with an AUC = 0.656 (95% CI 0.58–0.73: p = 0.000) Fibrinogen levels < 2g/L (hypofibrinogenemia) were independently associated with severe TBI. (AOR 2.87 CI,1.34–6.14: p = 0.007). Levels above 4.5g/L were also independently associated with injury severity (AOR 2.89, CI 1.12–7.48: p < 0.05) Fibrinogen levels more than 4.5g/L were independently associated with mortality (OR 4.5, CI;1.47–13.61, p < 0.05). Conclusions The fibrinogen level is a useful tool in predicting severity including mortality of TBI in our settings. We recommend the routine use of fibrinogen levels in TBI patient evaluations as levels below 2g/L and levels above 4.5g/L are associated with severe injuries and mortality


Introduction
Fibrinogen levels drop quicker than any other factors in severe trauma such as Traumatic Brain Injury (TBI). Contemporaneous studies show that brinogen concentrations < 2 g/L are strongly related to mortality. However, little is known regarding brinogen levels and TBI severity as well as mortality in sub-Saharan Africa. We therefore set out to determine whether brinogen levels are associated with TBI severity and seven days outcomes.

Objectives
To determine the sensitivity and speci city of brinogen levels and the association with severity and mortality among TBI patients at Mulago Hospital.

Methods
We prospectively enrolled 213 patients with TBI aged between 13 and 60 years of age and presenting within 24hrs of injury. Patients with pre-existing coagulopathy, concurrent use of anticoagulant or antiplatelet agents, pre-existing hepatic insu ciency, diabetes mellitus and who were pregnant were excluded. Fibrinogen levels were determined using the Clauss brinogen assay.

Introduction
Trauma accounts for 11% of the world's disability adjusted life years (DALYs) with 90% of these occurring in Low and Middle Income Countries.(1) Traumatic Brain Injury (TBI) per se is a major cause of disability globally with an incidence rate of 200 per 100 000 people per year (2) In Uganda, head injuries with TBI are the commonest type of injuries accounting for 44% of trauma admissions at hospitals in Kampala (3) and mortality rate of 220/100,000.(4) The morbidity and mortality due to TBI is higher in low-income and middle-income countries (5,6) despite advancements in the clinical evaluation of patients with TBI using standardized protocols such as the Advanced Trauma Life Support (ATLS) protocols. (7,8) Evidence from prior studies shows that deaths from trauma can be prevented if adequate and timely identi cation of the problem is done and the appropriate line of management is decided early. (9) In low and middle income settings where there is limited access to prompt investigation modalities for TBI victims, clinicians often nd themselves relying on trauma algorithms, trauma assessment tools and clinical examination ndings to diagnose and direct TBI management. Some of the trauma assessment tools employed in the evaluation of TBI patients include; Abbreviated Injury Score(AIS), Trauma Injury Severity Score (TRISS) and the Glasgow Coma Scale (GCS) speci cally for TBI.(10) These assessment tools have been found to have considerable limitations and that they may not correlate well with severity of injury.(11) Among these, the GCS remains the commonest tool used to assess TBI severity in Sub-Saharan Africa. (12,13) Despite its wide utility, the GCS does not provide speci c parametric clinical information about the pathophysiologic abnormalities in TBI which are the targets of our interventions. (14) One example of a pathophysiologic event is intracranial bleeding which is also associated with poor clinical outcomes such as mortality and disability. (15) Fibrinogen, which is a positive acute phase protein(16) as well as a haemostatic protein (17) has been retrospectively studied as a prognostic indicator among TBI patients as well as a predictor of in hospital mortality (18)(19)(20). Following trauma, brinogen levels deteriorate more frequently and earlier than other routine coagulation parameters. (21,22) Additionally hypo brinogenemia has been described as a common occurrence in TBI possibly due to trauma induced coagulopathy. (19,21,23) A recent study showed that brinogen concentrations less than 2g.L were associated with poor outcomes including mortality in contrast to concentrations above 2.5g/L that are associated with favourable outcomes.(18, 19) Current guidelines also emphasize that brinogen concentrations be maintained over 1.5-2.0 g/L in severe trauma patients (22) Fibrinogen, therefore has a pivotal role in TBI; however, little is known regarding its sensitivity and speci city in diagnosis of severe TBI in sub-Saharan Africa.
We therefore set out to study the predictive ability of brinogen levels in determining TBI severity and predicting clinical outcomes in TBI as a step in improving prompt diagnosis and management of TBI victims. The aims of the study were to determine the sensitivity and speci city of low brinogen levels in predicting severity of traumatic brain injuries, to describe the association of brinogen levels with TBI severity and 7-day outcomes among TBI patients at Mulago Hospital. We hypothesized that plasma brinogen levels are associated with severity and short-term clinical outcomes in TBI patients.

Materials And Methods
Study design and setting.
We prospectively studied 213 randomly selected TBI patients admitted to the Casualty unit at Mulago National Referral Hospital (MNRH) between December 2021 and May 2022. MNRH is the biggest public hospital in Uganda at approximately 5 kilometres from the city centre and it receives 75% of injured victims in Kampala. (24) The Casualty unit of the hospital is the entry point for all trauma cases presenting to the hospital.
Study population and sampling.
The inclusion criteria were as follows: patients aged 13 to 60 years with a clinical diagnosis of TBI documented using Computed Tomography (CT) or Glasgow Coma Scale (GCS) score by clinician and admitted within 24Hrs of TBI occurrence. The age range of 13 to 60 years was used in consideration of the altered metabolism of brinogen that occurs at the young and elderly extremes of age. (25)(26)(27) TBI in this study was de ned as any alteration of brain function or presence of other evidence of brain pathology based on the GCS or head CT scan in a patient, caused by an external force such as accidents, assault, falls and burns.(28) Patients on concurrent use of anticoagulant or antiplatelet agents, medical diagnosis of liver disease, hypertension, and Diabetes mellitus, patients admitted after 24hrs of the injury occurrence and pregnant women were excluded.
To achieve our rst objectives, we used the proportion of patients with low brinogen from a prior study (19) and level of precision of 7% at 95% con dence interval to determine a sample size of 186 patients using the Kish and Leslie formula. (29) To study the relationship of brinogen with outcomes, we calculated the sample size using formula for cohort studies based on comparison of two proportions representing the event rates in both the exposed and the non-exposed groups. (30) Using proportions from the study by Lv et al.(31), with 95% CI and power of 80%, we determined a sample size of 140 patients for the cohort. Adjusting upwards for losses to follow up, we estimated the sample size to be 200 patients. We therefore enrolled a total of 213 patients using systematic random sampling to answer our objectives.
All patients were evaluated and treated according to the local protocol. Informed consent was obtained from the patients included in the study and for the unconscious patients, waiver of informed consent was obtained from the Research Ethics Committee of Makerere University and Mulago Hospital Ethics committee.
Study procedure and data Collection.
Data obtained included demographic information such as age, sex, occupation, time of injury, level of education, mechanism of injury and type of head injury. Clinical data including, blood pressure, pulse oximetry, temperature, pupillary reaction, CT scan results, GCS score and brinogen levels taken at time of admission were also obtained. The severity of injury was determined by the GCS score obtained by the neurosurgical team. The GCS score of ≤ 8 was categorized as severe TBI and scores 9-15 as non-severe TBI. (Fig. 1) Fibrinogen levels were measured by the Clauss brinogen assay using the "Yumizen G FIB 5" reagent. The test was carried out on fresh decalci ed venous blood obtained from the participants. On admission, 5 milliliters of venous blood were drawn from each of the participants into 3.2% sodium citrate vacutainers and transported to the laboratory within 60 minutes of collection for analysis. Samples were centrifuged to obtain plasma that was prepared for analysis as a 1:10 dilution with Yumizen G IMIDAZOL buffer. The prepared sample was then analysed using an automated analyser and results recorded in g/L. We obtained levels < 1g/L as well as those above 5g/L that were retested at 1:5 dilution and 1:20 dilution respectively to obtain nal results. The brinogen levels were categorized as: normal brinogen levels between 2 and 4.5g/L.(32) A brinogen level of < 2 g/L was considered as being low and a level of > 4.5 g/L as high according to standard laboratory reference values. The patients were followed up daily for 7 days and outcomes documented. The clinical outcome studied was in hospital mortality within 7 days of admission. The 11 patients lost to follow up were not analysed for outcomes. (Fig. 1) They were however included in the analysesis for the association between brinogen and injury severity on admission to the hospital.
Statistical analysis.
All study data collected was entered in Epidata version 4.6 software, cleaned and exported to STATA version 14 for analysis. Continuous variables were summarised as means with standard deviation.
Categorical variables are expressed as percentages. Bivariate analyses of categorical variables were performed using Pearson's chi test and presented as p values. A Receiver Operating Characteristic (ROC) curve is used to describe the predictive ability of brinogen levels in TBI. The sensitivity ± positive predictive value and speci city ± negative predictive value of brinogen levels were calculated using a 2 2 table. Binary logistic regression models were used to describe the relationship between categorical variables between the patient groups with GCS ≤ 8 and that with GCS ≥ 9. Following bivariate analyses, logistic regression multivariate models were used to evaluate the association between brinogen and TBI severity as well as in-hospital mortality. All patients with missing data were excluded from the analyses.
The models were tested for multiple collinearities, goodness of t and all independent variables with correlation coe cient above ± 0.4 were excluded from the logistic regression model. The relationship is presented as odds ratio with 95% con dence intervals. The sensitivity using brinogen <2g/L was SN =56.5%, PPV=64.9%, (p< 0.001) The speci city using brinogen <2g/L was SP =72.9%, NPV=61.7% (p< 0.001). (Table 2) Fibrinogen levels predict TBI severity with an AUC = 0.656, 95% CI (0.58-0.73) p=0.000. (Figure 2) [1] Missing 2 [2] Crude Odds ratio: Unadjusted for other variables. [3] Statistical signi cance determined by chi square test. Values in Bold are statistically signi cant. On bivariate analysis, brinogen levels were signi cantly associated with injury severity. SBP and pulse oximetry were also signi cantly associated with TBI severity. Without adjustment, TBI patients with brinogen levels < 2g/L were 3 times more likely to have severe TBI than in participants with brinogen levels of 2-4.5g/L. Further multivariate analysis with logistic regression showed a strong association between brinogen levels and TBI severity. When adjusted for age, systolic blood pressure, temperature and oximetry; brinogen levels < 2g/L (hypo brinogenemia) were independently associated with severe TBI. (AOR 2.87 CI,1.34-6.14: p = 0.007) Levels above 4.5g/L were also independently associated with injury severity (AOR 2.89, CI 1.12-7.48: p < 0.05) SBP > 140mmHg, oxygen saturation < 90% and temperature > 38º C were also associated with severe TBI. (Table 3) On bivariate analysis, brinogen levels were signi cantly associated with mortality. Other factors viz SBP, temperature and oximetry were also signi cantly related to mortality. (Table 4) After adjusting for patient age, hypoxia levels, and hemodynamic status on multivariate analysis, brinogen levels more than 4.5g/L were independently associated with mortality (OR 4.5, CI;1.472-13.607, p < 0.05). Temperature > 38ºC, SBP and oximetry were also associated with mortality. (Table 4)

Discussion
This study set out to determine the speci city and sensitivity of brinogen levels and the association with severity and mortality among TBI patients at Mulago Hospital. There are no studies addressing this topic from sub-Saharan Africa and this is the rst study in Uganda to describe this relationship. Despite numerous evidence available regarding the predictability of brinogen in the prognosis of TBI outcomes from prior studies (19,33), little is known concerning its predictive ability in the diagnosis of severe TBI.
The sensitivity using brinogen < 2g/L(hypo brinogenemia) was 56.5% with a positive predictive value of 64.9%. The speci city was 72.9% with a negative predictive value of 61.7%. In addition, hypo brinogenemia ( brinogen levels < 2g/L) was common in TBI patients occurring in 35.07% of TBI patients on admission. This is similar to 38.6% found in a previous study done by Lv,et al.(19) Our study also found that 20.4% of patients with TBI had high levels of brinogen(> 4.5g/L).
Recent research showed that for patients admitted with severe TBI, brinogen levels < 2g/L on admission are strongly related to increased mortality. (19) In addition, studies have demonstrated that in severe trauma, brinogen is reduced to critical levels. (36). By contrast, high brinogen levels have been described as being protective in patients with multiple trauma. (35) Furthermore, previous studies have demonstrated that TBI is associated with abnormalities in clot formation due to differences in brinogen levels among victims.(37) Fibrinogen could therefore be used as a marker or predictor of TBI severity.
The study found that low brinogen levels (< 2 g/L) were fairly predictive of TBI severity with an AUC = 0.656, sensitivity of 56.5% and speci city of 72.9%. Therefore, absence of hypo brinogenemia in TBI patients found in this study was associated with milder forms of TBI. This is consistent with what previous studies have described in severe trauma.
This study shows that the likelihood of having a severe form of TBI increases with low levels of brinogen(< 2g/L). Possible explanations for the above relationship stem from the presence of intracranial bleeding which is a common occurrence in TBI as noted in the CRASH trial.(38) Trauma induced coagulopathy is a crucial element in severe TBI especially when compounded with intracranial bleeding. The consumption of clotting factors and platelets in response to intracranial bleeding further 13. (39) TBI is also associated with systemic hyper brinolysis which occurs in as much as 20% of critical trauma patients hence lowering brinogen levels further. (40,41) In addition, prior studies have shown that brinogen levels drop drastically and most rapidly during haemorrhage. (34) Therefore, the association of low brinogen levels with severe TBI is possibly due to a combination of haemorrhage and trauma induced coagulopathy that occurs in severe TBI. (42) Much as there is a stronger association of severity with low brinogen levels, high brinogen levels (> 4.5g/L) were also associated with severe TBI. This could be due to the in ammation that accompanies major trauma and disruption of the blood brain barrier with release of procoagulant molecules. (37,43) A study done by Samuels et al found that TBI patients commonly presented with a spectrum ranging from hypocoagulability to hypercoagulability. (37) It is therefore likely that severe TBI without intracranial haemorrhage leads to high brinogen levels while severe TBI with haemorrhage lowers the brinogen levels.
Unlike prior research ndings, (19) this study showed that brinogen levels severe systemic in ammatory response due to a disrupted blood brain barrier causes early organ dysfunction and later multiple organ failure which leads to death. (47,48). This possibly explains the high brinogen levels found to be associated with mortality.
While one of the strengths of our study is that it was carried out prospectively in a high-volume trauma center, some limitations need to be acknowledged. This was a single centre study with a rather small sample size and with limited duration allocated to conduct the study due to speci ed time frames for research activities. Secondly, brinogen is not a routine test in Mulago Hospital for TBI patients and replacement therapy is not currently part of the management protocols in patients with TBI; hence, despite the identi cation of abnormalities among the participants, correction therapy with concentrate was not possible for the participants. Patients with coagulopathies however received other supplements such as tranexamic acid and Fresh frozen plasma. Additional prospective studies with larger sample size and longer study duration are needed to con rm the predictability of TBI severity and clinical outcomes using brinogen levels.

Conclusions
In conclusion, we established that brinogen is a useful tool in predicting severity of TBI and mortality. The study reveals that the sensitivity of brinogen levels < 2g/L is 56.5% and the speci city is 72.9%.
Fibrinogen fairly predicts TBI severity with an AUC of 0.656.
Fibrinogen levels may be used as an additional tool to screen TBI patients for injury severity. Low brinogen levels (< 2g/L) are predictors of TBI severity. High brinogen levels > 4.5g/L are also predictors of TBI severity. A brinogen level of > 4.5g/L is a strong predictor of mortality in TBI patients. Integrating brinogen as a biomarker in TBI management could therefore provide critical information about trauma physiology and ultimately in uence clinical decisions. Additional larger prospective studies are needed to con rm these ndings.

Declarations
Ethics approval and consent to participate The School of Medicine Ethics and Research Committee and Mulago Hospital Ethics and Research Committee approved the study in October 2021 with reference number; Mak-SOMREC-2021-126. Informed consent was obtained from each patient or legally acceptable representative of the patient before participation in the study. All methods were carried out in accordance with the Declaration of Helsinki.

Consent for publication
Not Applicable.

Availability of data and materials
The datasets generated and/or analyzed during the current study are not publicly available due to con dentiality agreements but are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests