Demographics,Mechanism of Injury,and Outcomes for Acute Upper and Lower Cervical Spinal Cord Injuries: An Analysis of 470 Patients in the Prospective,Multi-Center,North American Clinical Trials Network Registry

Abstract

There is a paucity of data comparing the demographics, mechanism of injury, and outcomes of upper versus lower cervical spinal cord injuries (cSCI). The study objective was to define different clinical manifestations of cSCI. Data were collected prospectively through centers of the North American Clinical Trials Network (NACTN). Data was collected on 470 patients (21% women, mean age 50 years). Cervical vertebral level was analyzed as an ordinal variable to determine a natural demarcation to classify upper versus lower cSCI. For continuous variable analysis, falls were associated with C3 more than C4 vertebral level injuries (60% vs. 42%; p = 0.0126), while motor vehicle accidents were associated with C4 more than C3 (40% vs. 29%; p = 0.0962). Motor International Standards for Neurological Classification of Spinal Cord Injury scores also demonstrated a natural demarcation between C3 and C4, with C3 having higher median American Spinal Injury Association (ASIA) motor scores (40 [4-73] vs. 11 [3-59], p = 0.0227). There were no differences when comparing C2 to C3 nor C4 to C5. Given the significant differences seen between C3 and C4, but not C2 and C3 nor C4 and C5, upper cSCI was designated as C1-C3, and lower cSCI was designated as C4-C7. Compared with a lower cSCI, patients with an upper cSCI were more likely to have a fall as their mechanism of injury (54% vs. 36%; p = 0.0072). Patients with an ASIA C cSCI were likely to have an upper cervical injury: 23% vs. 11% (p = 0.0226). Additionally, patients with an upper cSCI were more likely to have diabetes prior to injury: 37% versus 22%, respectively (p = 0.0084). Lower cSCI were more likely injured through sports (19% vs. 8%, p = 0.0171) and present with ASIA A (42% vs. 25%, p = 0.0186) neurological grade. Patients with lower cSCI were also significantly more likely to have complications such as shock, pulmonary embolism, and pleural effusion. In conclusion, there appears to be a natural demarcation of injury type between C3 and C4. Upper cSCI (C1-C3) was more associated with falls and diabetes, whereas lower cSCI (C4-C7) was more associated with sports, worse ASIA scores, and more complications. Further research will be needed to understand the mechanistic and biological differences between these two groups and whether different treatments may be appropriate for each of these groups.

Introduction

Spinal cord injury (SCI) can be a devastating disease, as it often presents several challenges in both the acute and chronic phases.¹ The U.S. National Statistics Center estimates the number of people living with SCI in the U.S. to be 273,000²; of these, nearly half of the injuries occur in young people (ages 16-30 years)² and half of the injuries occur in the cervical spine (C-spine).³ The increased mobility of the C-spine and its relatively smaller vertebrae set the C-spine as a primary location for SCI.⁴ While the thoracic spine has the rib cage for stabilization, the C-spine has little external support, predisposing the C-Spine to SCI.² It has been well documented that injuries of the C-spine can be the most severe as they are often associated with quadriplegia, cardiopulmonary dysfunction, muscle atrophy^5
-7 and reduced quality of life.

Currently, there is a lack of consensus on the classification of cervical SCI (cSCI) types.^8,9 Some classification systems, such as the AO Spine, focus on the morphology of the injury (compression, tension band, or translational).¹⁰ Other systems classify cervical injuries according to level. This is done by either classifying injuries as upper cervical spine if they are atlantoaxial injuries (C1-C2) or lower cervical spine if they are subaxial (C3-C7).⁹ Unfortunately, this classification system is often used to classify cervical vertebral fractures and not necessarily cord level.¹¹ There remains a paucity in the available research regarding the optimal manner to demarcate cervical spinal cord injuries with classification into upper versus lower based on mechanism of injury, demographic data, or outcomes related to the neurological level of the injury. Research available that focuses on cervical level and epidemiology does so by investigating the most common site of injury overall, not necessarily assessing the significance of groupings into upper versus lower.^12
-14 Additionally, most of these studies arise outside the U.S. and may not reflect the epidemiology within the United States.

The goal of the current study is to determine if there is a meaningful difference in which cervical injuries should be demarcated into upper and lower based on presentation, management, and hospital course of cSCI patients in North America. To do this, the registry of the North American Clinical Trials Network (NACTN) was used. This is a U.S.–Canadian prospective registry that collects information regarding demographics, past medical history, circumstances involving the injury, clinical presentation at a NACTN hospital, and the timing of treatment. Long-term data are also collected at 6 and 12 months follow up to determine how treatment and neurological status may have evolved.¹⁵ Findings from such a study may help determine if classification of upper/atlantoaxial versus lower/sub-axial classification is arbitrary or has clinical significance regarding surgical and medical management. With the granular nature of the data available in the NACTN registry, it is our hypothesis that there will be a natural demarcation point between upper and lower cervical spine injuries, although not necessarily in an atlantoaxial and subaxial distribution.

Methods

The NACTN SCI registry was queried for all patients who sustained a cSCI. The methodology regarding the registry, as well as the data analysis, are detailed in the accompanying article in this edition. Demographics, characteristics of injury, clinical presentation, and treatment timing were then obtained from the demographics and surgery files. Clinical presentation included variables such as diastolic blood pressure, systolic blood pressure, mean arterial pressure, and Acute Physiology and Chronic Health Evaluation II (APACHE II) scores. The APACHE II is a revised classification system of the APACHE scoring system. It collects information on 12 separate patient physiologic characteristics to provide a measurement of the severity of acute disease.¹⁶ Increasing APACHE II scores correlate with the subsequent risk of hospital death.¹⁶ For surgical timing, dates of injury, arrival, and surgery (for those who underwent surgery) were used to calculate trajectory times (injury to arrival, arrival to surgery, and injury to surgery). Cervical vertebral level was analyzed as an ordinal variable (1 = C1 to 7 = C7) to determine a natural demarcation. Based on this, patients were grouped into upper cSCI or lower cSCI and compared.

Participants' characteristics (age, sex, race), mechanism of injury (motor vehicular accidents, fall, sport/recreation, assault, other), comorbidities (high blood pressure, diabetes, heart attack, pulmonary, malignancy, smoker, drug abuse), and initial clinical status upon presentation were evaluated. The latter includes the initial documented American Spinal Injury Association (ASIA) examination to include the ASIA Impairment Scale (AIS) grade, motor, pin prick, and light touch scores. Initial ASIA measurements were those taken prior to surgery or within 7 days of injury if surgery was not performed. Treatment course, hospital length of stay, complications (cardiac, gastrointestinal and genitourinary, pulmonary, and skin), mortality, and discharge disposition (rehabilitation hospital, home, other) for the included patients also were analyzed.

Statistical analysis

Continuous variables were summarized using mean with associated standard deviation (SD), and median with 1st and 3rd quartiles which were compared with Brown-Mood test. The latter non-parametric test was used because the normality assumption of the two-sample t-test was not satisfied per the Kolmogorov-Smirnov test. Categorical variables were summarized with frequency count and percentage and compared with the chi-square test or Fisher's exact test as appropriate. All tests were two-sided and the significance level was set to 5%. The false discovery rate (FDR) was used to adjust form multiple testing. Data preprocessing and analyses were performed in SAS 9.4 (SAS Institute Inc., Cary, NC).

Results

Contiguous variable analysis

Of the 989 patients in the registry available for analysis, 470 met inclusion criteria of having a documented cSCI. Analysis was performed for demographics, mechanism of injury, in association with ASIA scores for C1-C7 vertebral level to determine a natural demarcation point to classify upper vs lower cSCI. Each level was considered as a separate category in this evaluation. For example, when comparing C3 to C4, solely these two levels were assessed, not C3 and caudally nor C4 and distally. For the mechanism of injury, C3 was more associated with falls (60% vs. 42%), while C4 was more associated with motor vehicle accidents (MVAs; 40% vs. 29%; p = 0.0126 and p = 0.0962, respectively), thus indicating a potential demarcation point. Motor International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI) score also demonstrated a demarcation between C3 and C4, with C3 having higher median ISNCSCI motor scores (40 [4 - 73] vs. 11 [3 - 59], p = 0.0227). Regarding demographics, there were no statistically significant differences between C2 and C3, C3 and C4, or between C4 and C5 (Table 1). There were no differences when comparing C2 to C3 nor C4 to C5 for demographics, falls, MVA, or ASIA scores. Therefore, for the remainder of the study, upper cSCI was designated as C1-C3, and lower cSCI was designated as C4-C7.

Table 1.

Comparative Analysis of C2 vs. C3 vs. C4 vs. C5

			C2	C3	C4	C5	4 groups
			(n = 33)	(n = 110)	(n = 90)	(n = 120)	C3 vs. C2	C4 vs. C2	C5 vs. C2	C4 vs. C3	C5 vs. C3	C5 vs. C4
Injury characteristics	Mechanism of injury		n = 33	n = 108	n = 90	n = 116
		MVA, n (%)	12 (36%)	31 (29%)	36 (40%)	36 (31%)	0.4046	0.7144	0.564	0.0962	0.7037	0.1823
		Fall, n (%)	15 (45%)	65 (60%)	38 (42%)	47 (41%)	0.1383	0.7486	0.6122	0.0126	0.0036	0.8054
		Sport/recreation, n (%)	2 (6%)	7 (6%)	12 (13%)	30 (26%)	0.9311	0.2735	0.0268	0.1107	0.0003	0.0298
		Assault, n (%)	2 (6%)	0 (0%)	3 (3%)	3 (3%)	0.971	0.5039	0.3428	0.9724	0.9729	0.7526
		Other, n (%)	2 (6%)	5 (5%)	1 (1%)	0 (0%)	0.7414	0.1602	0.97	0.1861	0.9706	0.9738
ASIA measure	AIS (initial within the first 7 days)		n = 22	n = 86	n = 72	n = 95
		A, n (%)	5 (23%)	22 (26%)	24 (33%)	34 (36%)	0.7829	0.3499	0.2475	0.2872	0.1402	0.7415
		B, n (%)	1 (5%)	8 (9%)	11 (15%)	18 (19%)	0.4815	0.2162	0.1329	0.2552	0.0706	0.5364
		C, n (%)	8 (36%)	19 (22%)	9 (13%)	12 (13%)	0.1735	0.0153	0.0114	0.1209	0.096	0.9798
		D, n (%)	8 (36%)	37 (43%)	28 (39%)	31 (33%)	0.5728	0.8313	0.7383	0.5994	0.1513	0.4032
	Motor score		n = 18	n = 61	n = 53	n = 61
		Mean ± SD	48 ± 37	41 ± 34	31 ± 33	37 ± 35
		Median [Q1-Q3]	39 [10 - 90]	40 [4 - 73]	11 [3 - 59]	18 [10 - 77]	0.7061	0.1608	0.2553	0.0227	0.021	0.4957
	Pin prick score		n = 15	n = 58	n = 49	n = 49
		Mean ± SD	75 ± 41	60 ± 39	50 ± 39	55 ± 37
		Median [Q1-Q3]	100 [48 - 108]	61 [16 - 100]	41 [16 - 90]	54 [20 - 98]	0.0969	0.0075	0.0368	0.279	0.7323	0.3723
	Light touch score		n = 15	n = 59	n = 52	n = 54
		Mean ± SD	78 ± 40	63 ± 39	56 ± 40	62 ± 38
		Median [Q1-Q3]	100 [60 - 108]	68 [20 - 101]	56 [16 - 103]	62 [22 - 105]	0.1358	0.0722	0.1139	0.4967	0.7249	0.772

MVA, motor vehicle accident; AIS, American Spinal Injury Association Injury Score; ASIA, American Spinal Injury Association; SD, standard deviation; Q, quartile.

Upper versus lower cSCI: Demographics and baseline patients characteristics

Given the above results, patients were divided into two groups representing those with upper (C1-C3) and lower (C4-C7) C-spine injuries. The two sub-groups were then compared. The mean age in the entire cohort was 50 (Table 2A). There was no significant difference in age, gender, or race for patients suffering from lower cervical injuries than those with upper cervical.

Table 2A.

Characteristics of the Participants with Cervical Injuries

				2 groups
			All cervical	C1-C3	C4-C7		FDR-adjusted
			(n = 470)	(n = 158)	(n = 312)	p value	p value
Demographics	Age		n = 451	n = 153	n = 298
		Mean ± SD	50 ± 35	54 ± 15	46 ± 18
		Median [Q1-Q3]	50 [35 - 61]	54 [47 - 65]	47 [30 - 60]	0.0554	0.1173
	Sex		n = 439	n = 146	n = 293
	Sex	Female, n (%)	90 (21%)	22 (15%)	68 (23%)	0.0466	0.1116
	Race		n = 428	n = 141	n = 287
		White, n (%)	298 (70%)	95 (67%)	203 (71%)	0.478	0.5433
		Black, n (%)	91 (21%)	33 (23%)	58 (20%)	0.4477	0.5372
		Other, n (%)	39 (9%)	13 (9%)	26 (9%)	0.9567	0.9567
Injury characteristics	Mechanism of injury		n = 463	n = 156	n = 307
		MVA, n (%)	175 (38%)	48 (31%)	127 (41%)	0.0262	0.0857
		Fall, n (%)	193 (42%)	84 (54%)	109 (36%)	0.0002	0.0072
		Sport/recreation, n (%)	69 (15%)	12 (8%)	57 (19%)	0.0019	0.0171
		Assault, n (%)	10 (2%)	2 (1%)	8 (3%)	0.3543	0.4724
		Other, n (%)	16 (3%)	10 (6%)	6 (2%)	0.0131	0.0524
Initial clinical state	SBP		n = 465	n = 154	n = 311
		Mean ± SD	123 ± 106	130 ± 30	121 ± 29
		Median [Q1-Q3]	123 [106 - 142]	128 [113 - 145]	121 [102 - 139]	0.8427	0.8668
	DBP		n = 465	n = 154	n = 311
		Mean ± SD	72 ± 62	79 ± 21	74 ± 19
		Median [Q1-Q3]	72 [62 - 85]	77 [64 - 88]	71 [61 - 84]	0.6622	0.7012
	MAP		n = 465	n = 154	n = 311
		Mean ± SD	91 ± 79	96 ± 18	89 ± 19
		Median [Q1-Q3]	91 [79 - 104]	95 [82 - 107]	90 [76 - 102]	0.3259	0.4512
	APACHE II		n = 297	n = 97	n = 200
		Mean ± SD	6 ± 3	7 ± 5	7 ± 6
		Median [Q1-Q3]	6 [3 - 10]	7 [4 - 9]	6 [3 - 10]	0.015	0.054
Comorbidities	High BP		n = 455	n = 152	n = 303
	High BP	Yes, n (%)	151 (33%)	44 (29%)	107 (35%)	0.1738	0.272
	Diabetes		n = 453	n = 152	n = 301
	Diabetes	Yes, n (%)	122 (27%)	56 (37%)	66 (22%)	0.0007	0.0084
	Heart attack		n = 453	n = 151	n = 302
	Heart attack	Yes, n (%)	18 (4%)	5 (3%)	13 (4%)	0.6099	0.6653
	Pulmonary		n = 453	n = 151	n = 302
	Pulmonary	Yes, n (%)	108 (24%)	33 (22%)	75 (25%)	0.4829	0.5433
	Malignancy		n = 451	n = 151	n = 300
	Malignancy	Yes, n (%)	6 (1%)	1 (1%)	5 (2%)	0.3796	0.4881
	Smoker		n = 452	n = 150	n = 302
	Smoker	Yes, n (%)	8 (2%)	5 (3%)	3 (1%)	0.0756	0.1512
	Drug abuse		n = 453	n = 151	n = 302
	Drug abuse	Yes, n (%)	42 (9%)	19 (13%)	23 (8%)	0.0858	0.1626

FDR, false discovery rate; SD, standard deviation; MVA, motor vehicle accident; SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; APACHE II, Acute Physiology and Chronic Health Evaluation II.

Table 2B.

ASIA Scores

			2 groups
		All cervical	C1-C3	C4-C7
		(n = 470)	(n = 158)	(n = 312)	p value	FDR-adjusted p value
AIS (Initial within the first 7 days)		n = 363	n = 118	n = 245
	A, n (%)	132 (36%)	30 (25%)	102 (42%)	0.0026	0.0186
	B, n (%)	45 (12%)	10 (8%)	35 (14%)	0.1156	0.1892
	C, n (%)	55 (15%)	27 (23%)	28 (11%)	0.0044	0.0226
	D, n (%)	131 (36%)	51 (43%)	80 (33%)	0.0496	0.1116
Motor Score		n = 255	n = 88	n = 167
	Mean ± SD	30 ± 8	43 ± 35	37 ± 32
	Median [Q1-Q3]	30 [8 - 70]	40 [5 - 77]	26 [10 - 60]	0.0056	0.0252
Pin Prick Score		n = 228	n = 80	n = 148
	Mean ± SD	55 ± 20	61 ± 40	53 ± 36
	Median [Q1-Q3]	55 [20 - 99]	62 [16 - 102]	42 [20 - 90]	0.0031	0.0186
Light Touch Score		n = 238	n = 82	n = 156
	Mean ± SD	60 ± 21	65 ± 40	58 ± 38
	Median [Q1-Q3]	60 [21 - 104]	69 [20 - 105]	56 [22 - 104]	0.1056	0.181

ASIA, American Spinal Injury Association; FDR, false discovery rate; AIS, American Spinal Injury Association Injury Score; SD, standard deviation.

Patients who had a fall were more likely to have higher cervical injuries: 54% versus 36% (p = 0.0072). Patients injured while involved in sports were conversely more likely to have a lower cervical injury: 19% versus 8% (p = 0.0171). Patients injured through MVA were not statistically more likely to have lower cervical injuries: 41% versus 31% (p = 0.0857; Table 2A).

There was no significant difference in systolic blood pressure, diastolic blood pressure, or mean arterial pressure, or APACHE II scores between upper and lower cSCI groups. The only statistically significant different comorbidity between the two groups was diabetes being higher in the upper cSCI compared with lower: 37% versus 22%, respectively (p = 0.0084; Table 2A). The other baseline characteristics regarding comorbidities were nonsignificant and are summarized in Table 2A.

Patients who were ASIA A were likely to have lower cervical injury: 42% versus 25% (p = 0.0186; Table 2B). No statistically significant difference was found in ASIA B or ASIA D between the two sub-groups. Patients with ASIA C were more likely to have an upper cervical injury: 23% versus 11% (p = 0.0226). Overall, patients with higher cervical injuries tend to have a better median motor score at presentation 40 (5-77) versus 26 (10-60; p = 0.0252), respectively. The same can be said for sensory (pinprick only): 62 (16-102) versus 42 (20 -90; p = 0.0186). There was no significant difference for light touch.

Upper versus lower cSCI: Surgical timing

The mean time from injury to arrival at the hospital, arrival to surgery, and injury to surgery was 5, 29, and 40 h respectively (Table 3A). There was no significance in any of the times between upper and lower cSCI (Table 3B).

Table 3A.

Timing of Injury to Arrival to Surgery

				Surgery
			All (n = 470)	Unknown (n = 16)	No (n = 17)	Yes (n = 3 91)	p value
Timing	Injury to arrival		n = 406	n = 16	n = 17	n = 373
		Mean ± SD	5 ± 7	4 ± 6	3 ± 4	5 ± 7
		Median [Q1-Q3]	2 [1 - 6]	2 [1 - 5]	1 [1 - 3]	2 [1 - 6]	0.5094
	Arrival to surgery					n = 411
		Mean ± SD				29 ± 31
		Median [Q1-Q3]				16 [8 - 41]
	Injury to surgery					n = 392
		Mean ± SD				40 ± 42
		Median [Q1-Q3]				23 [13 - 54]

SD, standard deviation; Q, quartile.

Table 3B.

Timing in Those who Underwent Surgical Treatment

				2 groups
			All cervical	C1-C3	C4-C7		FDR-adjusted
			(n = 470)	(n = 158)	(n = 312)	p value	p value
Treatment	Surgery		n = 452	n = 147	n = 305
	Surgery	Yes, n (%)	434 (96%)	136 (93%)	298 (98%)	0.0496	0.1116
	Time of surgery	Underwent surgery	n = 410	n = 127	n = 283
		< 24 h, n (%)	201 (49%)	46 (36%)	155 (55%)
		Over 24 h, n (%)	209 (51%)	81 (64%)	128 (45%)	0.0005	0.0084

FDR, false discovery rate.

Of the original cohort of 470 patients, 452 had data regarding their surgical management and 434 (96%) underwent surgery (Table 3B). Of those who underwent surgery, 201(49%) had surgery within 24 h of admission. Patients with lower cSCI were more likely to have surgery in <24 h: 55% versus 36% (p = 0.0084). There were no significant differences in time to surgery between the two groups.

Upper versus lower cSCI: Hospital outcomes

The mean length of stay (LOS) during the acute time after injury in the entire cohort was 14 days. There was no difference in LOS, hospital mortality, or disposition between upper and lower cSCI groups (Table 4). The highest reported complications were categorized as pulmonary (44%), with the most common issue being respiratory failure (29%). There was a significant difference between the two groups, with pleural effusion and pulmonary embolism: 13% versus 4% (p = 0.0438) and 6% versus 1% (p = 0.047), respectively. There were also more gastrointestinal and genitourinary (GIGU) and skin complications within the lower cSCI groups (29% vs. 15%, p = 0.037 and 19% vs. 7%, p = 0.037, respectively). As far as cardiac complications, there was a significant difference between the two groups regarding shock. Lower cSCI was more associated with shock (14% vs. 2%, p = 0.0074). Although shock was reported as a cardiac complication, it is unclear from the data extraction if a distinction between cardiogenic and neurogenic shock was made.

Table 4.

Hospitalization Outcomes for Participants Included with Cervical Injuries

				2 groups
			All cervical	C1-C3	C4-C7		FDR-adjusted
			(n = 470)	(n = 158)	(n = 312)	p value	p value
Hospital outcomes	LOS		n = 449	n = 151	n = 298
		Mean ± SD	14 ± 8	20 ± 22	24 ± 36
		Median [Q1-Q3]	14 [8 - 24]	12 [8 - 25]	15 [8 - 24]	0.8099	0.8609
	Died		n = 470	n = 158	n = 312
	Died	Yes, n (%)	14 (3%)	4 (3%)	10 (3%)	0.6849	0.8609
	Discharge disposition		n = 432	n = 148	n = 284
		Rehabilitation hospital, n (%)	334 (77%)	115 (78%)	219 (77%)	0.8895	0.8963
		Home, n (%)	60 (14%)	21 (14%)	39 (14%)	0.8963	0.8963
		Other, n (%)	38 (9%)	12 (8%)	26 (9%)	0.7154	0.8609
Complications	Cardiac		n = 415	n = 136	n = 279
		Bradycardia, n (%)	67 (16%)	23 (17%)	44 (16%)	0.7668	0.8609
		Other dysrhythmia, n (%)	18 (4%)	5 (4%)	13 (5%)	0.6445	0.8609
		Cardiac arrest, n (%)	19 (5%)	4 (3%)	15 (5%)	0.2653	0.5441
		MI, n (%)	2 (0%)	0 (0%)	2 (1%)	0.3223	0.5441
		Shock, n (%)	42 (10%)	3 (2%)	39 (14%)	0.0002	0.0074
		CHF, n (%)	2 (0%)	1 (1%)	1 (0%)	0.6028	0.8609
		Other, n (%)	72 (17%)	20 (15%)	52 (19%)	0.3208	0.5441
		One of the above, n (%)	169 (41%)	46 (34%)	123 (44%)	0.0458	0.1541
	GIGU		n = 396	n = 131	n = 265
		GI hemorrhage, n (%)	11 (3%)	0 (0%)	11 (4%)	0.018	0.0832
		Hematuria, n (%)	2 (1%)	1 (1%)	1 (0%)	0.6102	0.8609
		Ileus, n (%)	11 (3%)	4 (3%)	7 (3%)	0.8144	0.8609
		Pancreatitis, n (%)	2 (1%)	1 (1%)	1 (0%)	0.6102	0.8609
		Cholecystitis, n (%)	0 (0%)	0 (0%)	0 (0%)
		Acute renal failure, n (%)	7 (2%)	2 (2%)	5 (2%)	0.7981	0.8609
		Other, n (%)	76 (19%)	15 (11%)	61 (23%)	0.006	0.0438
		One of the above, n (%)	96 (24%)	20 (15%)	76 (29%)	0.0034	0.037
	Pulmonary		n = 415	n = 136	n = 279
		ALI ARDS, n (%)	14 (3%)	4 (3%)	10 (4%)	0.7334	0.8609
		Respiratory failure, n (%)	122 (29%)	31 (23%)	91 (33%)	0.0392	0.145
		PE, n (%)	19 (5%)	1 (1%)	18 (6%)	0.0089	0.047
		Pleural effusion, n (%)	42 (10%)	6 (4%)	36 (13%)	0.0071	0.0438
		Lobar collapse, n (%)	24 (6%)	4 (3%)	20 (7%)	0.0833	0.2202
		Mucus plug, n (%)	28 (7%)	7 (5%)	21 (8%)	0.3643	0.586
		Pneumothorax, n (%)	13 (3%)	1 (1%)	12 (4%)	0.0503	0.1551
		Hemothorax, n (%)	5 (1%)	0 (0%)	5 (2%)	0.1163	0.2689
		Other, n (%)	47 (11%)	10 (7%)	37 (13%)	0.0746	0.2123
		One of the above, n (%)	182 (44%)	46 (34%)	136 (49%)	0.004	0.037
	Skin		n = 395	n = 129	n = 266
		Sacral, n (%)	35 (9%)	7 (5%)	28 (11%)	0.0944	0.2329
		Heel, n (%)	8 (2%)	3 (2%)	5 (2%)	0.768	0.8609
		Scapular, n (%)	2 (1%)	0 (0%)	2 (1%)	0.3235	0.5441
		Occipital, n (%)	2 (1%)	0 (0%)	2 (1%)	0.3235	0.5441
		Trochanter, n (%)	0 (0%)	0 (0%)	0 (0%)
		Operative Wound, n (%)	8 (2%)	1 (1%)	7 (3%)	0.2193	0.4773
		Other, n (%)	33 (8%)	5 (4%)	28 (11%)	0.0251	0.1032
		One of the above, n (%)	59 (15%)	9 (7%)	50 (19%)	0.0022	0.037

FDR, false discovery rate; LOS, length of stay; SD, standard deviation; MI, myocardial infarction; CHF, congestive heart failure; GIGU, gastrointestinal and genitourinary; GI, gastrointestinal; ALI, acute lung injury; ARDS, acute respiratory distress syndrome; PE, pulmonary embolism.

Upper versus lower cSCI: Steroid administration

Steroid use was demonstrated in 30% of patients during their admission (Table 5). When analyzed, 22% were prescribed methylprednisolone/Solu-Medrol, whereas the other 8% were prescribed Dexamethason/Decadron. When comparing the two groups, there was no significant difference in the use of steroids nor the type of steroid prescribed.

Table 5.

Steroid Administration in Participants with Spinal Cord Injuries

			2 groups
		All cervical	C1-C3	C4-C7		FDR-adjusted
		(n = 470)	(n = 158)	(n = 312)	p value	p value
Steroid use		n = 458	n = 156	n = 302
	No steroid, n (%)	320 (70%)	119 (76%)	201 (67%)
	Steroid use, n (%)	138 (30%)	37 (24%)	101 (33%)	0.0316	0.0875
	No steroid, n (%)	320 (70%)	119 (76%)	201 (67%)	0.0316	0.0875
	Methylprednisolone/solumedrol, n (%)	101 (22%)	29 (19%)	72 (24%)	0.1989	0.2984
	Dexamethasone/decadron, n (%)	37 (8%)	8 (5%)	29 (10%)	0.0959	0.1726

FDR, false discovery rate.

Discussion

Our study is the first to attempt to find a natural demarcation of upper versus lower cervical spinal cord injuries based on demographics, mechanism of injury, clinical presentation, and hospital course. By comparing continuous levels from C1-C7, we found a natural demarcation occurs between C3 and C4 as it relates to the mechanism of injury and ISNCSCI motor scores. At the level of C3, the mechanism of injury was more associated with falls; however, at C4, injuries were more associated with MVAs. When participants were analyzed as groups (upper vs. lower cervical injury), differences remained regarding the mechanism of injury, with falls seen more often in the upper cervical group. Additionally, there were differences in clinical presentation. Upper cSCI patients were more likely to present as an AIS C, while lower cSCI patients were more likely to present as an AIS grade A with worse motor and pin prick scores as compared with upper cSCI. The management and hospital course also differed between the two groups. Lower cervical injuries were more likely to have surgery before 24 h and have more complications.

Upper versus lower cervical SCI have different demographics and mechanisms of injury

Although there was no significant difference in age, gender, and race between upper and lower cSCI, there was a notable trend with upper cervical injuries in older patients. Additionally, upper cervical injuries were more associated with falls. This information suggests that upper cervical injuries occur in older patients that are more prone to falls, whereas lower cervical injuries are more likely to be seen in younger patients involved in sports. The difference in injury pattern seen may be related to the velocity of impact in activities other than falls (such as sports or MVAs). For example, a recent study found that motorcycle passengers that suffered spinal cord injuries were more likely to have fractures in the subaxial spine and consistent with our results, more likely to suffer complete SCI.¹⁷

One reason that the lower cervical spinal cord may be more susceptible to injury after high-velocity trauma, such as sports may be the smaller diameter of the cord at these levels. In a study of 140 healthy volunteers, Ulbrich and colleagues found that the canal at C6 was significantly smaller than at C1 and C3.¹⁸ Additionally, the canal at C3 was smaller in diameter than C1. Moreover, there was an age-dependent decrease in these diameters. In a study of 225 cervical spine injuries in 149 consecutive patients 65 years old and above, Lomoschitz and colleagues found that patients over 75 years independent of causative mechanism and patients who fell from standing height independent of age were more likely to have injuries of the upper cervical spine.¹⁹ These authors posit that these differences may be due to the progression of the degenerative changes seen in the cervical spine. One must also consider the biomechanical forces that arise within the cervical spine. Depending on vertebral level, there are varying contributions to flexion and extension. Research has shown that flexion and extension are often initiated in the C4-C7 levels.²⁰ In the younger patients, C4-C7 is the most mobile, and thus this may be the reason the younger patients have injuries at these lower levels. With aging, C1 and C2 become more mobile, and that is why these upper levels may be more susceptible to injury in the older population.²¹

As to the question of the mechanism of injury, there are several factors to consider in this study. First, younger patients have a higher rate of being involved in sports or high-velocity impacts, and thus this could explain the differences seen in the mechanisms of injury. The Centers for Disease Control and Prevention recently released data collected from 1999-2019 that demonstrated males aged 15 to 24 were more likely to be involved in a fatal motor vehicle accident when compared with older adult age groups.²² A similar trend was seen with women, although not as significant. Similarly, as patients age, the propensity to be involved in falls due to age-related physiologic changes (lower body weakness, unsteady gait, confusion, and certain medications) increases. Thus, the incidence of these falls is higher in the older population.²³

Upper cSCI patients receive surgery later

We found that those with upper cSCI were more likely to have surgery after 24 h of admission compared with lower cSCI. This is especially relevant given evidence demonstrating early surgical intervention is beneficial in SCI. This evidence includes both an improved 2 grade AIS²⁴ recovery and less time in the critical care unit post stabilization.²⁵ In 2016, Wilson and colleagues published an article in the Spine Journal with recommendations for decompressive surgery in less than 24 h.²⁶ Fehlings and colleagues had similar recommendations a year later.²⁷ Additionally, although 24 h decompression is considered early, some studies have proposed decompression surgery at even earlier intervals, including within 12, 8, and 4 h from injury.^28
-30 Each of these time-points demonstrated some level of improvement in neurologic outcomes when compared with later surgical interventions.

A possible explanation for these differences in surgery times within our study is the recent guidelines related to the timing of surgery in patients with central cord injury. In addition to research demonstrating the importance of early decompression in central cord syndrome,³¹ in 2017 it was recommended that patients with central cord syndrome could be offered surgery earlier.³² Previous research has demonstrated that central cord syndrome occurs in a bimodal age distribution.³³ When seen in the young, it is typically due to high-energy events such as MVAs and sports. However, when seen in older groups, it is likely due to spinal canal stenosis predisposing patients to injury. It would be interesting for future research to investigate whether the level of cervical injury is associated with specific cord injury patterns. It should also be reiterated that patients with upper cSCI were of more advanced age and had less severe ASIA grades. These factors likely played a role in deciding to perform surgery less urgently.

Lower cervical injuries have worse hospital courses

We found that lower cSCI had a worse hospital course compared with upper cSCI. Not only did they have surgery earlier, but they also reported more complications. The highest reported complication in both groups was pulmonary; however, it occurred significantly more often in the lower cSCI group. The high incidence of pulmonary complications within our entire cohort align well with what is reported by the National Spinal Cord Injury Statistical Center and the current literature.^34,35 Previous research has demonstrated that pulmonary complications are among the highest reported complications in patients with SCI and significantly increase hospital length of stay.³⁶

Interestingly, in our study, there was no statistical difference in LOS between the lower and upper cSCI. This is surprising given the worse ASIA scores on presentation and higher incidence of complications seen in the lower cSCI. Further, research has demonstrated ASIA scores to be the best predictor of pulmonary complications and hospital length of stay.³⁶ Future research should investigate the incidences of specific pulmonary complications that occur within cSCI patients according to their specific level of injury and how this may be correlated with hospital LOS.

Lower cSCI were also more likely to have other complications such as cardiac, GIGU, and skin. When analyzed, the most common of these was GI hemorrhage, ileus, and sacral skin complications. Interestingly, this contradicts another study that found that patients with high vertebral level injury (C1-C5) had greater requirements for cardiovascular intervention.³⁷ Although a direct comparison to this study remains challenging given the different classification systems used. For example, our cutoff for upper injuries was at the C4 level, not C5. It is unclear whether the bradycardia and shock seen in our study were more of an acute versus ongoing process. Especially when considering that hemodynamic instability is often seen in the acute setting of SCI.³⁸ If such complications were acute and resolved after decompressive surgery, this would explain the lack of significant difference in hospital length of stays. Further, sacral ulcers are a common complication of immobility regardless of the underlying illness.³⁹ Future research should investigate the timing of onset and resolution of cardiac complications within the cSCI population as well as the timing and severity of skin complications to understand how they may affect hospital LOS within this group.

Steroid administration

When investigating the differences in complication rates between upper versus lower cSCI, it is important to note there were no statistical differences in steroid administration between the two groups. This is surprising given the higher rates of shock, pulmonary, GIGU, and skin complications. It is unclear based on study design whether these complication rates and steroid administration are linked. In 2015, a large multi-center study in Latin America published data indicating that of 970 surgeons interviewed, 56.1% reported they routinely treat the acute SCI patient with methylprednisolone.⁴⁰ This is considerably higher compared with the 30% seen in our current study. A similar expanded study was performed in 2018, which collected information regarding steroid administration including Europe, Asia Pacific, North America, and the Middle East.⁴¹ Again, this study reported higher rates of steroid administration, with 52.9% of surgeons prescribing steroids in acute SCI.⁴¹

As mentioned, given the design of the study, it is challenging to understand the true relationship between steroid administration and complication rates in the lower cSCI group. There have been a few recent studies; however, investigating the safety and efficacy of steroid administration in acute SCI patients. In 2020, a meta-analysis of 12 studies demonstrated a relative risk of 2.9 for hyperglycemia in patients prescribed methylprednisolone for acute SCI.⁴² Interestingly, there was an increased risk of pneumonia in the steroid-prescribed population in observational studies, but not in the randomized clinical trials. They also did not find any significant increased risk for gastrointestinal bleeding, decubitus ulcers, surgical site infections, atelectasis, or urinary tract infections. Based on the current literature, the relationship between steroid administration and its risk of complications for acute SCI remains unresolved.

Limitations

As with all publications that use a large registry database, our study has several limitations. Specifically, given that we applied retrospective hypotheses to the prospective NACTN database, we are unable to establish the causality of identified relationships. Additionally, our data were collected over a 15-year time frame during which the age demographic showed considerable change. Future studies should aim to look at more recent time-points.

Conclusion

Our study represents one of the first to highlight how the initial clinical presentations, mechanisms of injury, and hospital course differ in acute cervical spinal cord injury patients based on the cervical level of injury. A demarcation point was found at which C1-C3 could be classified as upper cSCI and C4-C7 could be classified as lower cSCI. Additionally, upper cervical injuries are more associated with falls and observed in older populations. Lower cervical injuries may be more likely in younger populations, are more often due to sports, and have more complications.

Footnotes

Acknowledgments

The views expressed in this manuscript are those of the author(s) and do not necessarily reflect the official policy of the Department of Defense of the U.S. Government.

Authors' Contributions

All authors have provided critical feedback and helped shape the research, analysis, and manuscript.

Funding Information

This material is based upon work supported by the U.S. Army Medical Research Acquisition Activity. under Grant Nos.W81XWH-07-1-0361, W81XWH-10-2-0042, W81XWH-13-2-0040 and Contract No. W81XWH-16-C-0031, and the Christopher & Dana Reeve Foundation.

Author Disclosure Statement

No competing financial interests exist.

References

Ahuja

, Wilson

, Nori

, et al. Traumatic spinal cord injury. Nat Rev Dis Primers, 2017; 3:1–21; doi: 10.1038/nrdp.2017.18

Ropper

, Neal

, Theodore

. Acute management of traumatic cervical spinal cord injury. Pract Neurol, 2015; 15(4):266–272; doi: 10.1136/practneurol-2015-001094

Mneimneh

, Moussalem

, Ghaddar

, et al. Influence of cervical spinal cord injury on thermoregulatory and cardiovascular responses in the human body: literature review. J Clin Neurosci, 2019; 69:7–14; doi: 10.1016/j.jocn.2019.08.022

Sekhon

, Fehlings

. Epidemiology, demographics, and pathophysiology of acute spinal cord injury. Spine, 2001; 26:S2–S12; doi: 10.1097/00007632-200112151-00002

Downey

, Downey

, Newhouse

, et al. Fatal hyperthermia in a quadriplegic man: possible evidence for a peripheral action of haloperidol in neuroleptic malignant syndrome. Chest, 1992; 101:1728–1730; doi: 10.1378/chest.101.6.1728

Khan

, Plummer

, Martinez-Arizala

, et al. Hypothermia in patients with chronic spinal cord injury. J Spinal Cord Med, 2007; 30:27; doi: 10.1080/10790268.2007.11753910

Hostettler

, Leuthold

, Brechbühl

, et al. Maximal cardiac output during arm exercise in the sitting position after cervical spinal cord injury. J Rehabil Med, 2012; 44:131; doi: 10.2340/16501977-0913

Masson de Almeida Prado

, Masson de Almeida Prado

, Ueta

RHS

, et al. Subaxial spine trauma: radiological approach and practical implications. Clin Radiol, 2021; 76(12):941.e1–941.e10; doi: 10.1016/j.crad.2021.09.006

Joaquim

, Ghizoni

, Tedeschi

, et al. Upper cervical injuries—a rational approach to guide surgical management. J Spinal Cord Med, 2014; 37(2):139–151; doi: 10.1179/2045772313Y.0000000158

10.

Vaccaro

, Koerner

, Radcliff

, et al. AOSpine subaxial cervical spine injury classification system. Eur Spine J, 2016; 25(7):2173–2184; doi: 10.1007/s00586-015-3831-3

11.

Marcon

, Cristante

, Teixeira

, et al. Fractures of the cervical spine. Clinics, 2013; 68(11):1455–1461; doi: 10.6061/clinics/2013(11)12

12.

, Zheng

, Ren

, et al. Current epidemiological profile and characteristics of traumatic cervical spinal cord injury in Nanchang, China. J Spinal Cord Med, 2022; 45(4):556–563; doi: 10.1080/10790268.2021.1949188

13.

Devivo

. Epidemiology of traumatic spinal cord injury: trends and future implications. Spinal Cord, 2012; 50(5):365–372; doi: 10.1038/sc.2011.178

14.

, Li

, Ning

, et al. Epidemiology of traumatic cervical spinal cord injury in Tianjin, China. Spinal Cord, 2012; 50(10):740–744; doi: 10.1038/sc.2012.42

15.

Christopher and Dana Reeve Foundation. NACTN SCI Registry. Available from: https://www.christopherreeve.org/research/nactn/the-nactn-data-registry [Last accessed: July 31, 2022 ].

16.

Knaus

, Draper

, Wagner

, Zimmerman

. APACHE II: a severity of disease classification system. Crit Care Med, 1985; 13(10):818–829.

17.

Wang

, Liu

, Zhao

, et al. Incidence and pattern of traumatic spinal fractures and associated spinal cord injury resulting from motor vehicle collisions in China over 11 years: An observational study. Medicine, 2016; 95(43):e5220; doi: 10.1097/MD.0000000000005220

18.

Ulbrich

, Schraner

, Boesch

, et al. Normative MR cervical spinal canal dimensions. Radiology, 2014; 271(1):172–182; doi: 10.1148/radiol.13120370

19.

Lomoschitz

, Blackmore

, Mirza

, et al. Cervical spine injury in patients 65 years old and older: epidemiologic analysis regarding the effects of age and injury mechanism on distribution, type, and stability of injuries. AJR Am J Roentgenol, 2002; 178:573–577; doi: 10.2214/ajr.178.3.1780573

20.

Swartz

, Floyd

, Cendoma

. Cervical spine functional anatomy and the biomechanics of injury due to compressive loading. J Athl Train, 2005; 40(3):155–161.

21.

White

, Panjabi

. Clinical Biomechanics of the Spine, 2nd ed. Lippincott: Philadelphia, PA;, 1990; 97–119.

22.

Spencer

, Hedegaard

, Garnett

. Motor vehicle traffic death rates, by sex, age group, and road user type: United States, 1999-2019. NCHS Data Brief, 2021;(400):1–8.

23.

Rubenstein

. Falls in older people: epidemiology, risk factors and strategies for prevention. Age Ageing, 2006; 35 Suppl 2:ii37–ii41; doi: 10.1093/ageing/afl084

24.

Fehlings

, Vaccaro

, Wilson

, et al. Early versus delayed decompression for traumatic cervical spinal cord injury: results of the surgical timing in acute spinal cord injury study (STASCIS). PloS one, 2012; 7(2); doi: 10.1371/journal.pone.0032037

25.

Balas

, Guttman

, Badhiwala

, et al. Earlier surgery reduces complications in acute traumatic thoracolumbar spinal cord injury: analysis of a multi-center cohort of 4108 Patients. J Neurotrauma, 2022; 39(3-4):277–284; doi: 10.1089/neu.2020.7525

26.

Wilson

, Tetreault

, Aarabi

, et al. Guidelines for the management of patients with spinal cord injury: the optimal timing of decompression. Spine J, 2016; 63:172; doi: 10.1016/j.spinee.2016.07.123

27.

Fehlings

, Tetreault

, Wilson

, et al. A clinical practice guideline for the management of acute spinal cord injury: Introduction, rationale, and scope. Global Spine J, 2017; 7(S3):84S–94S; doi: 10.1177/2192568217703387

28.

Burke

, Yue

, Ngwenya

, et al. Ultra-early (<12 hours) surgery correlates with higher rate of American Spinal Injury Association impairment scale conversion after cervical spinal cord injury. Neurosurgery, 2019; 85(2):199–203; doi: 10.1093/neuros/nyy537

29.

Jug

, Kejžar

, Vesel

, et al. Neurological recovery after traumatic cervical spinal cord injury is superior if surgical decompression and instrumented fusion are performed within 8 hours versus 8 to 24 hours after injury: a single center experience. J Neurotrauma, 2015; 32(18):1385–1392; doi: 10.1089/neu.2014.3767

30.

Jug

, Kejžar

, Cimerman

, Bajrović

. Window of opportunity for surgical decompression in patients with acute traumatic cervical spinal cord injury. J Neurosurg Spine, 2019; 1–9; doi: 10.3171/2019.10.SPINE19888

31.

Qin

, Liu

, Wang

, et al. Impact of timing of surgery on traumatic central cord syndrome without fracture or dislocation. Interdiscip Neurosurg, 2022; 28:101507; https://doi.org/10.1016/j.inat.2022.101507

32.

Fehlings

, Tetreault

, Wilson

, et al. A clinical practice guideline for the management of patients with acute spinal cord injury and central cord syndrome: recommendations on the timing (≤24 hours versus >24 hours) of decompressive surgery. Global Spine J, 2017; 7(3 Suppl):195S–202S; doi: 10.1177/2192568217706367

33.

Brooks

. Central Cord Syndrome. Neurosurg Clin N Am, 2017; 28(1):41–47; doi: 10.1016/j.nec.2016.08.002

34.

Aito

. Complications during the acute phase of traumatic spinal cord lesions. Spinal Cord, 2003; 41(11):629–635; doi: 10.1038/sj.sc.3101513

35.

Tator

, Duncan

, Edmonds

, et al. Complications and costs of management of acute spinal cord injury. Paraplegia, 1993; 31:700–714.

36.

Aarabi

, Harrop

, Tator

, et al. Predictors of pulmonary complications in blunt traumatic spinal cord injury. J Neurosurg Spine, 2012; 17(1 Suppl):38–45; doi: 10.3171/2012.4.AOSPINE1295

37.

Bilello

, Davis

, Cunningham

, et al. Cervical spinal cord injury and the need for cardiovascular intervention. Arch Surg, 2003; 138(10):1127–1129; doi: 10.10001/archsurg.138.10.1127

38.

Eldahan

, Rabchevsky

. Autonomic dysreflexia after spinal cord injury: Systemic pathophysiology and methods of management. Auton Neurosci, 2018; 209:59–70; doi: 10.1016/j.autneu.2017.05.002

39.

Young

, Dobrzanski

. Pressure sores. Epidemiology and current management concepts. Drugs Aging, 1992; 2(1):42–57; doi: 10.2165/00002512-199202010-00006

40.

Teles

, Cabrera

, Riew

, Falavigna

. Steroid use for acute spinal cord injury in latin merica: a potentially dangerous practice guided by fear of a lawsuit. World Neurosurg, 2016; 88:342–349; doi: 10.1016/j.wneu.2015.12.045

41.

Falavigna

, Quadros

, Teles

, et al. Worldwide steroid prescription for acute spinal cord injury. Global Spine J, 2018; 8(3):303–310; doi: 10.1177/2192568217735804

42.

Sultan

, Lamba

, Liew

, et al. The safety and efficacy of steroid treatment for acute spinal cord injury: a systematic review and meta-analysis. Heliyon, 2020; 6(2):e03414; doi: 10.1016/j.heliyon.2020.e03414