The Impact of Hypoalbuminemia on Outcomes in Non-Surgically Treated Patients With Central Cord Injury

Abstract

Study design

Database study.

Objectives

Acute traumatic central cord syndrome (atCCS) is the most common incomplete spinal cord injury in the US, characterized by motor weakness of the upper extremities with relative sparing of the lower extremities and varying degrees of bladder dysfunction and sensory changes caudal to the lesion. Hypoalbuminemia (HA) has previously been associated with poorer outcomes following acute spinal cord injury. We hypothesized that patients with atCCS and HA treated non-surgically experienced worse outcomes than those without HA.

Methods

Data was collected using the Pearldiver^TM database and included adult patients who sustained atCCS from 2010-2022Q2. Pearson χ2 test and Welch’s t-test were used to evaluate differences in demographic and clinical data. Univariate analysis and multivariate logistic regression were performed.

Results

HA in conjunction with non-surgical care of atCCS was associated with an increased risk of development of renal failure and pressure ulcers as well as longer length of stay and two-year mortality. There was no significant increase in odds ratios for 90-day hospital readmission, pneumonia, UTI, respiratory failure, or sepsis.

Conclusions

Patients with hypoalbuminemia after atCCS treated non-surgically are at an increased risk of developing complications such as renal failure, pressure ulcers, longer lengths of stay and increased mortality. Surprisingly, respiratory failure, pneumonia, and sepsis were not found to be statistically different between patients with atCCS with and without HA. Inherent to a database study there are predictable limitations, however a large-scale analysis could help further delineate physiologic factors affecting outcomes of atCCS patients.

Introduction

Acute traumatic central cord syndrome (atCCS) is the most common incomplete spinal cord injury in the US which typically results from a hyperextension injury either affecting elderly patients with pre-existing cervical spondylotic stenosis or younger adults following motor vehicle crashes. ¹ Together they represent about 9% of all adult spinal cord injuries and about 6.6% of pediatric spinal cord injuries. ² The characterizing features of atCCS involve a disproportionate weakness of the upper extremities compared to the lower extremities combined with sacral sparing and, depending on the severity and extent of the lesion relative to its injury region, will be associated with some degree of spasticity, bladder dysfunction in addition to varying levels of sensory dysfunction. Within this injury constellation, motor function is typically more affected than sensory function. ³

While atCCS is a clinical diagnosis with strong imaging correlation that holds a relatively favorable recovery potential there remains a significant outcomes variability based on the underlying patient baseline health.^4-6 Regardless of chosen management strategy, atCCS recovery likely depends in part on a number of underlying patient health factors such as the baseline nutritional status represented in the form of key blood-based proteins, such as albumin. Previous studies have confirmed hypoalbuminemia to be associated with complications following major reconstructive spine surgery, such as surgical site infections, poor wound healing, sepsis, pulmonary embolism, and even mortality.^7,8 Hypoalbuminemia as a surrogate for malnutrition or active inflammatory disease state could also be applied as a potential predictor of poor outcomes in the setting of severe systemic challenges such as spinal cord injury.

The potential interrelationship of hypoalbuminemia and adverse clinical outcomes in patients with atCCS has not yet been explored. By applying the benefits of insights gained by analysis gained from a large database, we aimed to investigate outcomes in patients with and without hypoalbuminemia that were non-surgically treated, prompted by the intent to remove the many variables associated with different types of surgical care and explore as purely as possible the impact of the interaction of spinal cord injury and subsequent complications related to nutritional deficiency as expressed in the form of hypoalbuminemia. We hypothesized that there is a direct relationship of poor outcomes in the postinjury clinical course with hypoalbuminemia of patients treated non-surgically for atCCS.

Materials and Methods

Data Source

We performed a formal query of the PearlDiver database (https://www.pearldiverinc.com/, Colorado Springs, CO), a fee-based patient records database. This insurance-based database contains a combination of patient demographics, charge data, and procedural records for patients with disease coding using the International Classification of Diseases, 9th Revision (ICD-9-CM), 10th revision (ICD-10-CM) diagnoses and procedures or Current Procedural Terminology (CPT) codes from multiple different insurers, including both Medicare and Humana (representing a private insurer). The PearlDiver database houses records of over 165 million patients from either the Mariner claims or the Medicare Standard Analytical Files (SAF), which refers to any records that are billed to Medicare. Access to this database, which was stored on a password-protected server, was obtained from PearlDiver Technologies for the purpose of academic research. Patient data was de-identified and therefore exempt from institutional review board approval.

Study Population

The study population included adult patients who were managed nonoperatively for acute traumatic central cord syndrome from 2010 to 2022-Q2. Patients were categorized to having HA if they received the diagnosis at any point prior to admission for atCCS. Patients were then stratified into two groups based on the diagnosis of hypoalbuminemia (ICD-9-D-2738, ICD-10-D-E8809) and matched using a 1 to 10 ratio of HA vs control patients. Baseline status included a detailed assessment of comorbidities that collected sufficient data to enable calculation of the Charlson Comorbidity Index (CCI). We also specifically looked at typical comorbidities such as obesity, smoking, diabetes mellitus, chronic kidney disease, depression, and others. For our study we excluded patients who had any form of surgery as treatment for their spinal cord injury.

Outcome Measures

We assessed 90-day complications, including medical events such as arrhythmia, myocardial infarction, cardiac arrest, respiratory failure, pleural effusion, pulmonary embolism, deep vein thrombosis, cardiovascular accident, electrolyte abnormality, renal failure, urinary retention, pneumonia, sepsis, urinary tract infection, and pressure ulcer. Both 30-day and 90-day readmissions, length of stay, and mortality rates by the conclusion of 2 years were also documented. Patient discharge status was obtained from the Standard Analytic Files dataset.

Statistical Analysis

Pearson χ2 test and Welch’s t-test were used to assess for differences in demographic and clinical data. Univariate analysis was conducted first to determine significant postoperative outcomes. Multivariable logistic regression was performed using the significant variables derived from the univariate analysis and clinical expert opinion in order to determine the effects of hypoalbuminemia on various outcomes. All statistical analyses were performed using the research query interface provided by PearlDiver Bellwether. The common statistical metrics, including frequencies, mean values, and odds ratios (ORs), were collected by the Bellwether system using the R statistical package. Statistical significance was set at P < .05. Using the matplotlib library in Python, forest plots were created with the adjusted ORs plotted alongside their 95% confidence interval (CIs) for each outcome. For outcomes where the CIs crossed the null value of 1, the lines were bolded.

Results

A total of 2695 patients with at CCS that were managed non-operatively from 2010-2022Q2 met inclusion criteria. Of these 245 patients were placed into our study group by meeting our clinical parameters of hypoalbuminemia, with the remaining 2450 who did not have hypoalbuminemia assigned to serve as our control group. A summary of demographic variables including age, sex, and comorbidities including BMI, smoking history, chronic obstructive pulmonary disease, peripheral vascular disease, diabetes mellitus, hyperlipidemia, hypertension, hypothyroidism, congestive heart failure, coronary artery disease, renal disease, depression, and dementia, as well as a Charleston Comorbidity Index (CCI) are represented in Table 1. We found that on average, patients with hypoalbuminemia had a higher mean CCI compared to the control group as well as statistically higher rates of all other comorbidities measured in the analysis. Our subsequent analyses controlled for these potentially confounding variables.

Table 1.

Summary of Demographic Variables for 2695 Patients Managed Nonoperatively for Central Cord Syndrome.

	Hypoalbuminemia		Control		P-value
	n = 245		n = 2450
Age	65.1 ± 12.3		65.0 ± 12.2		.839
Sex (Female)	80	32.7%	800	32.7%	1.000
Comorbidities
CCI	5.3 ± 3.9		2.5 ± 2.7		<.001*
BMI 30-40	37	15.1%	207	8.4%	<.001*
BMI 40+	20	8.2%	8	.3%	.031*
Smoke	113	46.1%	581	23.7%	<.001*
Chronic obstructive pulmonary disease	120	49.0%	581	23.7%	<.001*
Peripheral vascular disease	88	35.9%	437	17.8%	<.001*
Diabetes mellitus	138	56.3%	951	38.8%	<.001*
Hyperlipidemia	143	58.4%	1212	49.5%	.009*
Hypertension	212	86.5%	1623	66.2%	<.001*
Hypothyroidism	57	23.3%	399	16.3%	.007*
Congestive heart failure	56	22.9%	217	8.9%	<.001*
Coronary artery disease	98	40.0%	608	24.8%	<.001*
Renal disease	91	37.1%	387	15.8%	<.001*
Depression	124	50.6%	642	26.2%	<.001*
Dementia	17	6.9%	86	3.5%	.012*

CCI, Charlson comorbidity index.

Boldface type indicates statistical significance with P < .05.

The outcomes after 90 days are summarized in Table 2. Patients with hypoalbuminemia were more likely to experience care complications following their nonsurgical management of their atCCS. We also observed a higher rate of cardiopulmonary complications such as cardiac arrhythmia, respiratory failure, and pneumonia in the hypoalbuminemia cohort as well as metabolic and renal derailments such as electrolyte abnormality, renal failure, and urinary retention. In addition, infections and related complications such as urinary tract infections, pneumonia, and sepsis were found to be more frequent in patients with hypoalbuminemia, as were pressure ulcers. Patients with hypoalbuminemia were also found to have an increased length of stay compared to our control cohort (P < .001), similar to a higher 90-day hospital readmission rate for the HA study group (P < .001), whereas no difference was found for the 30-day hospital readmission rate.

Table 2.

Univariate Analysis of 90-Day Complications.

	Hypoalbuminemia		Control		Adjusted OR (95% CI)	P-value
	N = 245		N = 2450
Arrhythmia	44	18.0%	301	12.3%	1.56 (1.10-2.21)	.015*
Myocardial infarction	11	4.5%	65	2.7%	1.72 (.90-3.31)	.146
Cardiac arrest	6	2.4%	34	1.4%	1.78 (.74-4.29)	.302
Respiratory Failure	36	14.7%	183	7.5%	2.13 (1.45-3.13)	<.001*
Pleural effusion	18	7.3%	111	4.5%	1.67 (1.00-2.80)	.070
PE	2	.8%	15	.6%	1.34 (.30-5.88)	1.000
DVT	8	3.3%	47	1.9%	1.73 (.81-3.70)	.236
CVA	11	4.5%	72	2.9%	1.55 (.81-2.97)	.252
Electrolyte Abnormality	35	14.3%	163	6.7%	4.05 (2.84-5.76	<.001*
Renal Failure	50	20.4%	146	6.0%	4.05 (2.84-5.76)	<.001*
Urinary Retention	37	15.1%	228	9.3%	1.73 (1.19-2.52)	.005*
Pneumonia	41	16.7%	201	8.2%	2.25 (1.56-3.24)	<.001*
Sepsis	17	6.9%	81	3.3%	2.18 (1.27-3.74)	.007*
UTI	57	23.3%	384	15.7%	1.63 (1.19-2.24)	.003*
Pressure Ulcer	52	21.2%	168	6.9%	3.66 (2.59-5.16)	<.001*
Hospital readmission (30 d)	54	22.0%	415	16.9%	1.39 (1.01-1.91)	.055
Hospital readmission (90 d)	71	29.0%	504	20.6%	1.58 (1.18-2.11)	.003*
LOS	19.1		12.0		–	<.001*

PE, Pulmonary embolism; DVT, Deep vein thrombosis; CVA, Cerebrovascular accident; UTI, Urinary tract infection; LOS, Length of stay.

Boldface indicates statistical significance with P < .05.

The multivariate analysis evaluating the effect of hypoalbuminemia on various outcomes is represented in Table 3 and is further detailed as a Forest Plot in Figure 1. Hypoalbuminemia was independently associated with an increased risk of development of renal failure and pressure ulcers (P < .001), a longer length of stay and increased likelihood of 2-year mortality (P = .001). Two-year mortality was higher in patients with hypoalbuminemia, with an odds ratio of 1.63. The highest odds (OR 2.30) were observed for pressure ulcers. Although we observed increased odds for 90-day hospital readmission, pneumonia, UTI, respiratory failure, and sepsis, those were not statistically significant.

Table 3.

Multivariate Analysis of 90-Day Complications.

	Hypoalbuminemia		Control		Adjusted OR (95% CI)	P-value
	N = 245		N = 2450
Arrhythmia	44	18.0%	301	12.3%	.99 (.65-1.47)	.952
Respiratory Failure	36	14.7%	183	7.5%	1.47 (.89-2.37)	.118
Electrolyte Abnormality	35	14.3%	163	6.7%	1.44 (.90-2.26)	.119
Renal Failure	50	20.4%	146	6.0%	1.69 (1.10-2.55)	.014*
Urinary Retention	37	15.1%	228	9.3%	1.06 (.67-1.62)	.802
Pneumonia	41	16.7%	201	8.2%	1.16 (.71-1.85)	.539
Sepsis	17	6.9%	81	3.3%	.81 (.39-1.59)	.554
UTI	57	23.3%	384	15.7%	.93 (.63-1.36)	.727
Pressure Ulcer	52	21.2%	168	6.9%	2.30 (1.53-3.41)	<.001*
Hospital readmission (90 d)	71	29.0%	504	20.6%	.96 (.68-1.35)	.828
Mortality (2 y)	135	55.1%	898	36.7%	1.63 (1.21-2.19)	.001*
LOS	19.1		12.0		–	<.001*

UTI, Urinary tract infection; LOS, Length of stay.

Boldface type indicates statistical significance with P < .05.

Figure 1.

Forest plot of adjusted OR with 95% CI.

Discussion

First described by Schneider et al, in 1954, ³ central cord syndrome is the most common incomplete spinal cord injury in the United States.^9,10 Understanding of the neuropathophysiology of atCCS has changed over time, with the more recent explanation revolving around a more diffuse shear pattern of injury rather than representing trauma to a specific columnar somatotopic organization of the corticospinal tract in the spinal cord.^10-12 Hypoalbuminemia has previously been associated with poor outcomes following acute spinal cord injury in general.^13-15 From a serologic standpoint, albumin plays a critical role in maintaining intravascular osmotic pressure and serving as a transport protein for hormones, drugs, and other molecules while also participating in immune and inflammatory responses. ¹⁶ Historically, serum albumin levels have been used as a simple yet non-specific surrogate marker for poor nutritional status without actually by itself being diagnostic of the multifactorial pathoentity of malnutrition. ¹⁷ More recently, albumin has been viewed as a negative acute phase reactant, complicating the assumption that hypoalbuminemia is directly related to nutritional status.^18,19 To date, the association between serum albumin and atCCS has not yet been explored.

In our study, hypoalbuminemia (HA) was associated with an increased likelihood of renal failure in patients with atCCS compared to the control group, with an odds ratio of 1.69. Hypoalbuminemia in patients with pre-existing renal failure is attributed to both a reduced synthesis and increased degradation. ²⁰ However, previous studies have reported low albumin to be a risk factor for the development of renal complications ²¹ with a 2010 meta-analysis reporting a causal relationship between hypoalbuminemia and the development of acute kidney injury or acute renal failure. ²² With regard to potential prevention and treatment, previously, administration of human albumin has been found to be nephroprotective in some cases,^20,23,24 but mechanisms are not understood, and this has not been studied in the context of spinal cord injury. Future studies may explore the use of serum albumin as a prevention of complications in patients with hypoalbuminemia and atCCS.

Patients in the analysis with hypoalbuminemia were also more likely to develop pressure ulcers compared to control. In the past, HA has been found to be directly correlated with the development of pressure ulcers in patients in the intensive care unit. ²⁵ Furthermore, albumin administration has been found to both prevent pressure ulcers from developing and progressing to advanced stages. ²⁶ However, a more recent study found that HA alone was not a risk factor for pressure ulcers in surgical patients, but rather the associated comorbidities of HA were causing the ulcers. ²⁷

Hypoalbuminemia has been found to increase mortality with other spinal pathologies ²⁸ This is supported by our results as overall mortality differed significantly between the two study groups, with an odds ratio of 1.63 for the HA compared to the control group. In practice, serum albumin is screened on routine bloodwork ¹³ and can thus be a relatively simple way to help aid in the prognostic evaluation of patients presenting with atCCS.

Of note, HA was not found to be associated with an increased risk of developing respiratory failure, though respiratory complications are common in patients with spinal cord injuries.^29-32 The cause of respiratory failure in SCI is attributed to weakness of respiratory muscles resulting in an inability to cough and clear secretions from the respiratory system; higher level and completeness of injury are associated with a greater risk of respiratory complications.^30-32 One study found that 74% of patients with complete spinal cord injury required mechanical ventilation, with an increased risk with Asia A complete cervical spinal cord injuries. ¹⁴ Extent of severity of deficits in atCCS and association with respiratory complications may be explored further. Furthermore, hypoalbuminemia is itself a risk factor for the development of pulmonary complications such as ARDS, and it is thus overall surprising that our results did not support the existing evidence. ³³ Additionally, patients with HA were not found to have an increased risk of developing pneumonia or sepsis, despite hypoalbuminemia being a known risk factor for the development of both complications. ³⁴ However, we cannot elucidate these findings, which are contradictory to the literature.

Database studies have inherent limitations. Pearldiver, as a private analytics database, includes patients primarily 65 and older with a predilection for the Southern United States ³⁵ and includes patients that are not randomly sampled. As with any database study, there is also the potential for coding error, either clerical or a result of annual updates to ICD-10 code definitions. ³⁶ Furthermore, severity of deficits and neurologic level of injury following central cord injury was not explored and may be related to the development of complications following injury. A potential interrelation of baseline HA and atCSS on outcomes vs acquired HA during hospitalization could not be explored with the data as available. We decided to minimize variables introduced by spine surgery care in spinal cord injury by reducing our investigation to those patients who received nonsurgical treatment only. Patients placed into nonsurgical care as treatment choice may have been afflicted by poor medical baseline conditions, thus creating a selection bias. Lastly, there may be other variables that were not measured or included in the analysis that influence outcomes. Future case-controlled, prospective studies might be more suitable to parse out potential relationships between HA and atCCS complications.

Conclusion

Patients presenting with hypoalbuminemia after acute traumatic central cord injury treated non-surgically are at an increased risk of developing certain complications such as renal failure and decubital pressure ulcers and also have longer lengths of stay, higher longer term readmissions risk after 90 days as well as an increase in overall mortality. Surprisingly, respiratory failure, pneumonia, and sepsis were not found to be statistically different between patients with atCCS with and without HA. This database study has a significant size and claims to have high data quality standards does have limitations, but future case-controlled, prospective studies may further explore complications in atCCS patients and hypoalbuminemia.