Novel approach to an early assessment of a patient’s potential for neurological remission after acute spinal cord injury: Analysis of hemoglobin concentration dynamics

Abstract

Context/objective: Examining hemoglobin (Hb) dynamics with regard to the potential of neurological remission in patients with traumatic spinal cord injury (TSCI).

Design: Prospective Clinical Observational Study.

Setting: BG Trauma Centre Ludwigshafen, Department of Paraplegiology, Rhineland-Palatinate, Germany.

Methods: From 2011 to 2017 a total of 80 patients with acute spinal injury were enrolled and divided into three groups: initial neurological impairment either with (G1; n = 33) or without subsequent neurological remission (G0; n = 35) and vertebral fractures without initial neurological impairment as control group (C; n = 12). Blood samples were taken for 3 months at 11 time-points after injury. Analyses were performed using routine diagnostics.

Outcome measures: Multiple logistic regression was used to determine the prognostic value of Hb regarding neurological remission respecting clinical covariates.

Results: Data showed elevated mean Hb concentrations in G1 from the third day to 1 month compared to G0, Hb levels were significantly higher in G1 after 3 days (P = 0.03, G1 > G0). The final multiple logistic regression model based on this data predicting the presence of neurological remission resulted in an AUC (area under the curve) of 80.5% (CI: 67.8%–93.2%) in the ROC (receiver operating characteristic) analysis.

Conclusion: Elevated Hb concentrations are associated with a higher likelihood of neurological remission. Elevated concentrations of Hb in G1 compared to G0 over time might be linked to both a better initial oxygen supply response and a decreased ECM (extracellular matrix) degradation highlighting the role of Hb as a valuable biomarker for neural regeneration after TSCI.

Introduction

Acute traumatic spinal cord injury (TSCI) still is a devastating consequence of severe accidents with an annual incidence of about 40 per million in industrial countries and only slightly lower one in developing countries.^1–3 Despite considerable progress, causal treatment is still missing. Complex cellular and biochemical processes characterize tissue regeneration processes occurring after TSCI. The first hours after acute TSCI are determined by micro bleedings, axonal lesions, inflammation, edema, and ischemia initiating neuronal cell death at the lesion side.⁴ While this initial phase of injury is caused by the mechanic trauma to the spinal cord and the surrounding tissues, a secondary injury phase involving vascular dysfunction, amongst others, leads to a protracted period of tissue destruction.⁵

Besides the severe neurological impairment patients are systemically affected by substantial blood loss due to both the initial trauma and the subsequent necessary surgeries, often resulting in substantial hypoperfusion.⁶ After the injury, perfusion in the gray matter is compromised due to hemorrhage, and general hypoperfusion further deteriorates existing ischemia. Over time ischemia can extend to the white matter. A particularly severe consequence of this local hypoperfusion is a slowed or blocked conduction of action potentials along axons leading to spinal shock.^1,2 Additionally, ischemia after TSCI leads to hypoxia activating reactive oxygen species (ROS) that are known to be both cyto- and neurotoxic. Among other things, ROS cause irreversible secondary lesions to the spinal cord and ongoing neurological impairment of patients.^3,4 Via both caspase-dependent and caspase-independent programmed motor neuron death, ROS can induce cell death in up to 50% of the neurons after a primary lesion.⁶ Recent studies have outlined a reduction of cell death of motor neurons by inhibiting oxidative stress after SCI.³ Despite these adverse effects, hypoxia has also been reported to be involved in the regulation of growth and differentiation mechanisms in the nervous system.⁷ On a molecular level, the hypoxia-induced factor-1α (HIF-1α) regulates gene expression by binding hypoxia response elements (HREs) in the corresponding promoter region.⁸ On the one hand controlled acute intermittent hypoxia induces an increased expression of growth and neurotrophic factors in non-respiratory motor neurons and therefore induces recovery of limb function in mammals;⁹ on the other hand uncontrolled and prolonged hypoxia, has a detrimental effect on neuronal tissue after SCI.¹⁰

Oxygen delivery depends on hematocrit, hemoglobin concentration, and blood viscosity.¹¹ Recent studies have shown a favorable role of elevated Hb levels in the prevention of brain hypoxia. Furthermore, Hb concentrations of more than 9 g/dL in patients with poor-grade aneurysmal SAH (Subarachnoid Hemorrhage), are strongly associated with a decreased incidence of brain hypoxia and cell energy dysfunction.¹² The relevance of sufficient Hb levels is further confirmed by similar findings in patients with supratentorial intracerebral hemorrhage.¹³

Red-cell transfusions are frequently used in clinical routine to avoid hypoperfusion and resulting oxygen debt.¹⁴ Despite their benefits for some patients, the use of red-cell transfusions above a Hb concentration of 7 g/dL is highly contentious due to adverse effects as immunosuppressive and microcirculatory complications.^15,16 Even in critically ill patients, restrictive transfusion protocols seemed to be equivalent to more liberal ones.¹⁴ However, since the introduction of leucocyte depletion for blood transfusion, the risk of leukocyte-mediated adverse reactions has significantly decreased.¹⁷

In patients with ischemic brain and spinal cord injuries, the optimal Hb transfusion threshold remains undefined. Yet, it might be higher than restrictive protocols suggest. Kurtz et al. have investigated brain tissue oxygenation in patients after severe brain injury. Their findings revealed that patients with Hb levels between 9.1 and 10 mg/dL were less likely to undergo brain tissue hypoxia (BTH) and metabolic crises (MC). Patients with both higher and lower Hb concentrations had a higher risk of suffering from BTH and MC, whereas Hb concentrations of <8 mg/dL were associated with a higher mortality rate.^18,19

In the current study, we sought to firstly investigate the role of anemia (assessed by Hb dynamics) during the early phase of TSCI. Secondly, we sought to determine whether analysis of Hb could serve as a potential early biomarker for neurological remission after TSCI.

Material and methods

Study design

This study was designed as a prospective clinical observational study.²⁰ Patient enrollment started after the local ethics committees of both the University of Heidelberg (S-514/2011) and the Landesärztekammer Rheinland-Pfalz (837.188.12 / 8289-F), Germany had approved upon it. Moreover, the study was conducted in accordance with the latest version of the Declaration of Helsinki.

Inclusion and exclusion criteria

The study included all adult patients with acute TSCI. The criteria for patient exclusion were: non-traumatic SCI, cancerous diseases in medical history, patients with chronic inflammatory diseases such as Morbus Bechterev and Spondylolysis, traumatic brain injuries, severe abdominal trauma, traumatic amputation of extremities, coma and any additional major life-threatening trauma apart from the SCI. No patient received methylprednisolone sodium succinate at no time during the study. Patients could choose to leave the study voluntarily at any time and for any reason. All patients gave their informed consents.

Setting

All treatments and analyses were performed at the BG Klinik Ludwigshafen in Germany (level 1 trauma center).

Participants

From 2011 to 2016 a total of 80 patients with acute TSCI was enrolled in the current study.

Assignment to groups

First, based on initial impairment and injuries, patients were divided into two groups; patients with neurological impairment (TSCI) where assigned to the study group G and patients with acute spinal injury without initial neurological impairments were assigned to the control group (C, n = 12). Depending on the outcome, patients in group G were further divided into two subgroups 3 months after injury (Fig. 1):

1. patients with neurological remission subsequent to treatment (G1, n = 33)

2. patients without neurological remission after treatment (G0, n = 35)

Figure 1.

Collective identification scheme.

Outcome measurements and analysis

Blood samples were drawn at the following points in time: On admission (0 h), and after 4, 9, 12 h, 1 and 3 days and 1, 2 weeks and 1, 2 and 3 months after trauma (Fig. 2).^21–25 Blood samples were obtained at the same points in time for all participating patients. Hb dynamics were immediately determined via blood gas analyses as part of the routine diagnostics. If more than one sample was obtained at one of the defined points in time, the average was calculated according to the scheme presented in Table 1.

Figure 2.

Blood samples drawing scheme. At each time point four vials of serum (each 7.5 ml) were obtained, that is, on admission, 4, 9, 12 h, 1 and 3 days as well as 1, 2 weeks and 1, 2 and 3 months after SCI.

Table 1. Averaging scheme for Hb when more than one value was available within the given timeslot corresponding to our sampling scheme.

Time point	Average of all values
0 hours	Admission -12 hours
1 day	12–36 hours
3 days	2.5–3.5 days
7 days	6.5–7.5 days
2 weeks	13–15 days
1 month	25–31 days
2 months	50–62 days
3 months	75–93 days

Demographical and clinical patient data was collected for each patient, and all patients were assessed through the ASIA impairment scale (AIS) (Table 2). This assessment was conducted with patients who were awake and responsive at the time of admission and after 3 months by the head physical therapist at the BG Trauma Centre, according to the International Standards for Neurological Classification of SCI (ISNCSCI; Table 2).^23,26,27 Initial examinations were performed within the first 3 days after the injury occurred, as suggested by Herbison et al.²⁸ Neurological remission was characterized as a positive AIS conversion within 3 months after the injury.

Table 2. Asia impairment scale with a short explanation of the different grades of impairment.

AIS grade	Clinical state
A	Complete – No motor or sensory function is preserved in the sacral segments S4–S5
B	Incomplete – Sensory but not motor function is preserved below the NLI and includes the sacral segments S4–S5
C	Incomplete – Motor function is preserved below the NLI, and more than half of key muscles below the NLI have a muscle grade less than 3
D	Incomplete – Motor function is preserved below the NLI, and at least half of key muscles below the NLI have a muscle grade of 3 or more
E	Normal – Motor and sensory function is normal

Statistics

All statistical calculations were performed with R version 3.4.4 using ‘ggplot2’ to create figures, ‘pROC’ for receiver operator characteristics (ROC) analysis and ‘corrplot’ for correlation matrices.^29–32 To detect location shifts between groups, the non-parametric Mann–Whitney U-test for independent samples was used. The Chi-square-test was used to assess differences within categorically distributed variables. Univariate logistic regression models were used to assess the predictive power of Hb regarding the criterion of neurological remission. Multiple logistic regression was utilized to estimate the adjusted odds ratio of covariates. All points of time with datasets that contained at least of two-thirds of all 80 cases were included in the statistical modeling process. Model comparison was based on the Akaike information criterion (AIC).^33,34 Model selection was performed via stepwise backward AIC compairison.^33,35 The primary measure for the predictive performance of the logistic regression model was the area under the curve (AUC) of the ROC analysis. All p values quoted are to be interpreted descriptively as they were not adjusted for multiple testing, and this is an exploratory analysis. Serum levels are expressed in absolute mean concentrations ± SEM (standard error of the mean). Statistical significance was assessed with P < 0.05.

Results

Patient demographics in G1 and G0

At admission, the average age of patients was 44.06 years (CI: 40.02–48.11). 68 patients (54 male, 14 female) were affected by TSCI. There was no significant difference of age, sex or etiology between G1 and G0. Yet, the neurological level of injury (NLI) differed significantly (P = 0.04) between the groups G1 and G0. The group G0 entailed a higher number of patients with thoracic impairment in G0. Results from the initial and final AIS differed greatly between the two groups with more AIS A patients in G0 than in G1. All 80 patients (G1/ G0/ C) had undergone surgery. Initial and final AIS score distribution of patients in G0 and G1 is given in Fig. 3. For further information on the entire collective please consult Table 3.

Figure 3.

Distribution of initial AIS and final AIS in G0 and G1. Colored connections indicate the individual configuration of each patient with initial neurological impairment with respect to the initial and final AIS as well as the group assignment to either G0 or G1.

Table 3. Descriptive statistics are given concerning the whole collective.

	All (N = 80)	G0 (N = 35)	G1 (N = 33)	C (N = 12)	P value
Sex					0.31
Female	19 (24)	5 (14)	9 (27)	5 (42)
Male	61 (76)	30 (86)	24 (73)	7 (58)
Age					0.90
Min	15	17	15	27
Max	82	77	82	71
Median (IQR)	43.50 (26.50, 59.00)	43.00 (27.00, 53.50)	48.00 (23.00, 59.00)	41.00 (32.00, 60.00)
Mean (95% CI)	44.06 (40.02, 48.11)	43.43 (37.46, 49.40)	43.88 (36.99, 50.77)	46.42 (37.48, 55.35)
Etiology					0.85
Fall	51 (64)	21 (60)	21 (64)	9 (75)
Traffic	26 (32)	12 (34)	11 (33)	3 (25)
Other	3 (4)	2 (6)	1 (3)	0 (0)
NLI					0.04
No SCI	10 (12)	0 (0)	0 (0)	10 (83)
Cervical	28 (35)	13 (37)	15 (45)	0 (0)
Thoracic	30 (38)	19 (54)	9 (27)	2 (17)
Lumbar	12 (15)	3 (9)	9 (27)	0 (0)
Decompression					0.48
No	21 (26)	6 (17)	9 (27)	6 (50)
Yes	59 (74)	29 (83)	24 (73)	6 (50)
Decompression Number of segments					0.22
Min	0	0	0	0
Max	4	4	3	2
Median (IQR)	1.00 (0.00, 2.00)	2.00 (1.00, 2.00)	1.00 (0.00, 2.00)	0.50 (0.00, 1.00)
Mean (95% CI)	1.29 (1.06, 1.51)	1.57 (1.22, 1.92)	1.24 (0.90, 1.58)	0.58 (0.21, 0.96)
Spinal fusion					0.93
No	10 (12)	3 (9)	4 (12)	3 (25)
Yes	70 (88)	32 (91)	29 (88)	9 (75)
Spinal fusion Number of segments					0.45
Min	0	0	0	0
Max	6	6	5	3
Median (IQR)	1.00 (1.00, 2.00)	1.00 (1.00, 2.00)	2.00 (1.00, 2.00)	1.50 (0.75, 2.00)
Mean (95% CI)	1.57 (1.33, 1.82)	1.57 (1.18, 1.97)	1.67 (1.29, 2.04)	1.33 (0.78, 1.89)
Instrumentation					0.34
Conservative	2 (2)	0 (0)	0 (0)	2 (17)
Ventral	23 (29)	9 (26)	13 (39)	1 (8)
Dorsal	55 (69)	26 (74)	20 (61)	9 (75)
EK					0.57
No	51 (64)	20 (57)	21 (64)	10 (83)
Yes	29 (36)	15 (43)	12 (36)	2 (17)
AIS initial					< 0.01
A	37 (46)	28 (80)	9 (27)	0 (0)
B	13 (16)	2 (6)	11 (33)	0 (0)
C	12 (15)	0 (0)	12 (36)	0 (0)
D	6 (8)	5 (14)	1 (3)	0 (0)
E	12 (15)	0 (0)	0 (0)	12 (100)
AIS final					< 0.01
A	28 (35)	28 (80)	0 (0)	0 (0)
B	5 (6)	2 (6)	3 (9)	0 (0)
C	9 (11)	0 (0)	9 (27)	0 (0)
D	25 (31)	5 (14)	20 (61)	0 (0)
E	13 (16)	0 (0)	1 (3)	12 (100)

AO, AO classification; AIS, American Spinal Injury Association (ASIA) Impairment Scale; NLI, neurological level of injury. Neurological remission was defined as improvement in AIS. P values show differences between G0 and G1.

Evaluation of Hb

All three groups (G1/G0/C) started with comparable Hb concentrations and presented a remarkable drop within 24 h after injury. Hb concentrations of Group G1 and C decreased only slightly within the following 48 h, while group G0 continued to decrease steeply. These findings are suggesting a link between the probability of neurological remission and the Hb concentration 72 h after trauma.

At admission, slightly higher Hb concentrations were observed in G1 than in G0. Initial Hb concentrations were 13.16 ± 0.33 g/dl in G1, 12.97 ± 0.33 g/dl in G0 and 13.51 ± 0.46 g/dl in C. Three days after admission, Hb concentration levels dropped to a minimum in all groups (G1: 9.93 ± 0.31 g/dl; G0: 9.18 ± 0.22 g/dl; C: 10.50 ± 0.88 g/dl). Interestingly, we found that levels were lowest in G0, followed by those in G1 while patients in C showed the highest Hb concentrations. Over time, analysis of Hb concentrations in G1 and G0 showed a slow but steady increase with final concentrations (at 3 months) comparable to those observed at admission (G1: 12.31 ± 0.51 g/dl; G0: 11.99 ± 0.36 g/dl). Due to loss to follow up, continuous data for C was only available for the first 2 weeks after injury. However, within 2 weeks, a more similar pattern was observed to G1 than to G0 with the final concentration of 10.62 ± 0.69 g/dl. Hb concentration levels of G1 and G0 were significantly different 3 days after injury (Hb 3d, P = 0.03, G0 < G1). The relative Hb concentrations did not reveal significant differences between G0 and G1 when compared to the value at admission (Hb percentage). The corresponding observations are presented in Fig. 4(A) and (B). Table 4 provides detailed statistical results of the group comparisons. In addition it also shows the utilized test based on the presence or absence of normal distribution.

Figure 4.

Hb concentration analysis. Part A indicates the serum concentration pattern in mean ± SEM, part B presents the relative Hb concentration compared to the initial concentration at admission, part C shows the number of cases included in the corresponding multivariate logistic regression model regarding the criterion neurological remission with respect to the patients age and sex. Part C indicates the corresponding AUC of the models.

Table 4. Specifics of the regression models I and II.

				Confidence interval
	Variables	P values	Adj. Odds ratio	2.50%	97.50%
Model I	Hb 3d	0.06	1.81	1.01	3.59
Model II	Hb 3d	0.05	2.06	1.04	4.71
	NLI	0.04	4.18	1.14	19.81
	EK (yes / no)	0.04	0.22	0.04	0.83
	Spinal fusion (yes / no)	0.10	0.12	0.01	1.4
	Spinal fusion Number of segments	0.08	1.80	0.95	3.8
	Instrumentation	0.02	0.11	0.01	0.61

The adjusted odds ratios of variables included in the logistic regression model, P values and the corresponding 95% confidence interval of the adjusted odds ratio in the analysis.

Logistic regression and ROC analysis

Logistic regression modeling was performed for data at admission as well as 24 h, 3 days and 1 week after an injury. All other points in time did not provide a sufficient number of cases in each dataset and were thus excluded from regression modeling. Figure 4 presents the number of cases included in each model (Fig. 4C) and the resulting AUC (Fig. 4D). All included variables were standardized. Based on the AUC, the best performing model was Model I based on Hb 3d. Due to loss to follow up a total was of 53 cases included in this calculation (Fig. 4C). A ROC analysis using the Model I to predict neurological remission, resulted in an AUC of 67.7 (CI: 52.9%–82.6%) (Table 4, Fig. 5).

Figure 5.

ROC (Receiver Operator Characteristic) curve analysis of Model I indicating its AUC (Area under the Curve) and the corresponding confidence interval (CI) for predicting remission as a function of the Hb concentration 3 days after injury (Table 4).

For an improved prediction of neurological remission based on Hb 3d, further modeling was set up, including the potential covariates age, sex, NLI, Hb 3d, Hb 3d percentage, type of instrumentation, presence and number of decompressed segments, and presence as well as the number of segments that underwent spinal fusion. A backward selection of variables was then performed based on the Akaike Information Criterion (AIC) to select the model with the best characteristics relative to each of the other models. The resulting final Model II included patient’s NLI, Hb 3d, the application of erythrocyte concentrates, the presence or absence of spinal fusion, the number of segments that underwent spinal fusion as well as the type of instrumentation. Any variance of variable selection led to a deterioration of the AIC. The corresponding ROC analysis based on Model II revealed an Area under the Curve of 80.5% (CI: 67.8%–93.2%) (Fig. 6). Table 4 provides an overview of the model specifications used in the analysis including the 95% confidence intervals and P values for the adjusted odds ratios.

Figure 6.

ROC (Receiver Operator Characteristic) curve analysis of the final Model II indicating its AUC (Area under the Curve) and the corresponding confidence interval (CI) for predicting remission as a function of the Hb concentration 3 days after injury with respect to covariates (Table 4).

Discussion

Complex cascades involving cytokines like MMPs,^22,36–38 IL-10 or IL-8 play an essential role in neural regeneration.²³ Hypoxia is considered an essential factor of both neural cell impairment and as a stimulus for remodeling processes. Thus, hypoxia might be a potential key factor regarding the outcome of TSCI. In this explorative and prospective clinical observational study, we sought to determine the role of Hb during neural regeneration after TSCI. The data of the current study revealed that different concentrations of Hb might be related to the neurological outcome. Low Hb concentrations within 3 days after injury seem to be associated with poor neurological outcome 3 month after TSCI.

The role of Hb concentration after TSCI within 7 days after injury

In 2006, Furlan et al. described a reduction of hemoglobin after TSCI in a case–control study including a total of 30 patients.³⁹ Three days after injury Hb concentrations in the present study were significantly higher in patients in G1 compared to G0 and only exceeded by patients in the group C. Partly, these differences might be linked to more severe trauma in the TSCI group, leading to higher blood loss, along with injury to the spinal cord, than in the control group. While oxygen delivery to tissues is determined by the coordinated activity of the heart, vasculature, and microcirculation,⁴⁰ the relative oxygen delivery depends on hematocrit, hemoglobin concentration, and blood viscosity.¹¹

The developed regression model was able to predict the outcome based on Hb 3d in addition to clinical covariates with an AUC of 80.5% (CI: 67.8%–93.2%). This indicates, that in 80.5% of cases a higher model-based score is assigned to patients who show neurological improvement.

Further controlled trials are needed to evaluate the impact of the covariates “instrumentation” and “application of erythrocyte concentrates”.

In 1999 Hebert et al. discovered that restrictive transfusion of red blood cell concentrates in critically ill patients, except for patients with acute myocardial infarction and unstable angina.¹⁴ This equals to a liberal transfusion strategy. In 2014 they confirmed their findings in patients with septic shock, by demonstrating a similar overall mortality within 90 days of septic shock regardless of the applied threshold for administering red blood cell concentrates.⁴¹ Data from Holst et al. supported this conclusion.⁴² Yet, another study by Carson et al. found an improved overall survival rate in patients who underwent surgery with a preoperative hemoglobin level of 12 g/dL or higher compared to patients with a preoperative hemoglobin level of less than 6 g/dL.⁴³ In TSCI, anemia is frequently observed in the acute phase,⁴⁴ thus patients with severe anemia, regardless of its cause, are likely to be affected by more severe associated secondary complications.

Whilst, preoperative anemia, even in a mild degree, is directly associated with poor clinical outcome intraoperative transfusion of packed red blood cells was found to be linked with an increase mortality rate,^45–47 as well as secondary complications such as surgical-site infection, pneumonia, and sepsis.⁴⁸ While it is evident that patients with physiological Hb concentrations recover better from surgery, it is not yet clear whether the higher mortality in patients who received intraoperative transfusion is due to the adverse effects of blood transfusion or is, instead, the result of increased blood loss in the patients receiving blood.⁴⁷ Our data suggests that a higher Hb concentration within the first 3 days after trauma leads to an improved outcome 3 months after TSCI. Therefore, a more liberal erythrocyte concentrate application in the early phase after TSCI might help improve the neurological impairment and should be investigated separately, especially if surgery is intended. To our knowledge, to this day, no study has yet investigated the effects of a liberal versus a restrictive erythrocyte concentrate application focusing on the overall outcome and survival in patients with TSCI.

The potential role of erythropoietin (EPO) concentration dynamics after TSCI

Besides its role as a causal factor of neuronal death hypoxia is also a potent inductor of erythropoietin (EPO) release. EPO is secreted in response to cellular hypoxia resulting in erythropoiesis as a crucial part of the body’s response to compensate for the reduced blood oxygen levels. Anti-apoptotic and neuroprotective characteristics of EPO have been described and confirmed in various animal studies after acute TSCI.^7–10,49,50 Recently, positive effects of treatment with EPO for various ischemic conditions that led to neuronal cell death have been reported.⁵¹ As one can easily monitor and adjust Hb concentrations and pO₂ in the emergency room, the intensive care unit or on the ward, both are of crucial interest for future monitoring options in the early phase after injury. Our data suggest lower Hb concentration and consecutively diminished tissue oxygenation to be trigger for ischemia in the spinal cord after injury. A neuroprotective and regenerative effect of EPO after TSCI has already been reported.^52–56 Utada et al. investigated the treatment of experimental spinal cord ischemia using either EPO, IGF-1 or both (EPO + IGF-1).⁵⁷ The treatment with both IGF-1 and EPO resulted in an improved functional and neuroanatomical outcome compared to the control group.⁵⁷ By contrast, a large animal study with 87 rats that tried to verify the effect of EPO did not conclude a positive effect of EPO after induced SCI.⁵⁸ Thus, the effect of EPO levels on TSCI remains highly contentious. Further studies with a high temporal resolution and long term follow up are needed to investigate the predictive potential of EPO concentrations in patients with TSCI over time focusing on Hb dynamics.

Oxygen depth hypothesis

The positive correlation between NLI and thorax trauma in the patient collective is highly interesting. A higher NLI is generally related to thorax and the main impact to the spinal cord is in general corresponding to the NLI. Also, an overall incidence of pneumo-/ hemothorax is associated with a worse outcome compared to patients without affection of the thorax. This might be related to the observed significant difference in NLI between G1 and G0, as higher lesions in the spinal cord also resulting in more probably in thorax trauma. There is compelling evidence showing that early surgical management within 24 h after the injury has a positive effect on motor recovery. Early surgical management also reduces the incidence of adverse pulmonary events.⁵⁹ Patients with thorax trauma suffer from a diminished oxygen saturation. While elevated Hb concentrations are known to be linked to lower sizes of nerval hypoxia,^{12, 13} the optimal transfusion protocol still needs to be determined.¹⁸ The findings here presented indicate that further studies are necessary to investigate hemoglobin concentration dynamics with regard to significant covariates such as instrumentation and erythrocyte concentrate application in particular.

Limitations

TSCI remains a severe, yet rare injury. Therefore, clinical studies on its treatment are complex, and the available sample size and follow-up of our patient collective is limited. This explorative prospective study still involves a substantial number of cases (n = 80). Differences in the distribution of the initial AIS between G0 and G1 confirm actual findings, suggesting that AIS A patients be associated with lower rates of remission.⁵⁹ Larger studies with a higher sample size in each AIS subset are needed to investigate specific characteristics against the background of clinical covariates and potential biomarkers. Differences in Hb levels are multifactorial. Patients in G1 might have suffered from a trauma associated with minor blood loss or they might have undergone an operation that led to lower blood loss compared to G0. Differing infusion therapies contributed to temporarily differing Hb concentrations, and possible secondary bleeding might have decreased the patient’s concentration of Hb. All of these are possible reasons for differences in the final Hb concentration. Yet, the severity of trauma and the blood loss due to trauma and operation are considered the main factors.

Interestingly, the variable Hb 3d, concerning clinical covariates, outnumbered the patient’s age and sex in the multiple regression modeling process and emphasized predictive relevance regarding neurological remission. Taking the thoracic spinal cords exquisite sensitivity to manipulation and injury into account, further studies with a larger sample size are needed to investigate the potential interference of thoracic impairment with clinical covariates and potentially with specific neuroinflammatory response. Univariate differences in instrumentation, spinal fusion, decompression, as well as variances due to infusion therapy are considered to have little influence on the present collective. Due to the highly standardized treatment at the BG Trauma center and its well-established research protocol, possible biases were minimized. Concerning the variance of values and the limited number of participating patients, results must nevertheless be evaluated critically.

Conclusion

The results of the current study indicate that an elevated concentration of Hb in peripheral blood is associated with higher odds of neurological remission after TSCI, whereas an early decrease in Hb concentrations might support unfavorable tissue oxygenation and thus is associated with a higher probability of absence of remission. This elevated Hb concentration in patients with neurological remission after TSCI (G1) might be linked to both a better initial oxygen supply response and a decreased ECM degradation during the glial scar formation at the lesion site during the chronic state of SCI. The absence of neurological remission correlates with an initial configuration of diminished Hb values. In this study, we were able to highlight the role of Hb as a biomarker in neural regeneration after TSCI against the background of clinical covariates. The study provides evidence for the development of new prognostic, scoring and monitoring techniques.

Disclaimer statements

Contributors None.

Funding None.

Conflicts of interest None.

Ethics approval The study has been approved by the ethics committee of the University of Heidelberg (S-514/2011) and the Landesärztekammer Rheinland-Pfalz (837.188.12 / 8289-F), Germany.