Infections, particularly pneumonia, are the primary cause of mortality in individuals with spinal cord injury (SCI).1 Several factors may contribute to the high rate of infections in the SCI population, including motor paralysis and reduced reflexes resulting in aspiration, invasive procedures, and so-called SCI-induced immune depression syndrome (SCI-IDS).2,3 SCI-IDS is thought to occur when the connection between the CNS and the immune system are disrupted by a lesion in the spinal cord, resulting in a decrease in immune function.3 In addition to increased morbidity and mortality, infections after SCI may affect neurologic recovery.2,3 A recent study found that infections impaired the return of muscle strength up to 1 year postinjury; however, the long-term consequences remain uncertain.2
The aim of this study was to investigate whether infections occurring in an acute care setting after SCI affected long-term functional recovery and survival.4
METHODS
Data source and inclusion criteria.
The study by Kopp et al.4 used data gathered from the multicenter National Spinal Cord Injury Database (NSCID). The NSCID prospectively gathers data from 25 specialized traumatic SCI care centers.2 Inclusion criteria were admission within 24 hours after injury between 1995 and 2005, age between 17 and 75, a cervical SCI, complete baseline and infection data, and an American Spinal Cord Injury Association Impairment Scale (AIS) grade of A, B, or C (see table e-1 at Neurology.org for a full description of the AIS grading system). Participants were excluded if they resided in a hospital or nursing home prior to injury, had serious concomitant injuries affecting consciousness, or were rehospitalized for unspecified infectious or parasitic diseases during the follow-up period.
Exposures and outcomes.
The primary outcomes were the motor items of the Functional Independence Measure (FIM; FIMmotor total score and 4 additional FIMmotor subscores) at 4 timepoints (admission, discharge, 1 year, and 5 years postinjury). The secondary outcome measure was infection-associated mortality at 10 years postinjury. Hospital-acquired infections included pneumonia and postoperative wound infections that occurred during acute care or inpatient rehabilitation.
Statistics.
FIMmotor scores as a dependent variable were analyzed in 3 ways: an explorative analysis of scores at each of the 4 timepoints, a linear mixed model (LMM), and an adjusted LMM. Both models used multiple imputation and a sensitivity analysis using only complete cases. The measure of interest was a time × infection interaction term; that is, if early infection had a time-dependent effect on the recovery of FIMmotor scores. The LMMs included the initial model (adjusted only for baseline FIMmotor scores), an additional model (adjusted for AIS, level of injury, age, ethnic group, and working status), and finally 3 models stratified for AIS grade and adjusted for the aforementioned variables.
Mortality data were analyzed using an exploratory analysis and a Cox regression model (adjusted and unadjusted) over 10 years. Similar to the LMM, the Cox models were adjusted for AIS, stratified for AIS, and further adjusted for neurologic level of injury, age, working status, and educational level.
RESULTS
A total of 1,203 individuals were included in the study, with 47% (564) acquiring an infection during early care, the majority of which were pneumonia (540 acquired pneumonia, 11 postoperative wound infections, and 13 both). The group that acquired infections was significantly different from the group that did not with regards to sex (a higher proportion of males), injury severity (a higher proportion of AIS As), level of injury (a higher proportion of C1–C4 injuries), and median baseline FIMmotor scores (a smaller interquartile range).
The exploratory analysis in the complete cohort and the AIS A group identified significantly higher FIMmotor scores in those without acute infections at each time point. The AIS B and C group analyses only found a significant difference between the 2 groups at discharge.
The FIMmotor score LMMs not stratified by AIS (imputed and complete cases models) all revealed a significant association between acquired infections and impaired recovery of FIMmotor scores at each time point, and overall. Specifically, the imputed model adjusted for all variables revealed a −7.4 point difference (95% confidence interval [CI] −11.5 to −3.3) and the complete cases model revealed a −5.2 point difference (95% CI −8.4 to −2.0). The model was then stratified by AIS and adjusted for level of injury and sociodemographic variables. When the model was stratified by AIS grade, there was an overall significant association between infections and impaired recovery in AIS A (both models) and in AIS B at 5 years for the imputed model only. The AIS C model did not confer an overall significant association.
In-hospital mortality was not significantly different between the groups with and without infections; however, Kaplan-Meier survival curves indicated a significant difference in survival in both the total cohort and AIS subgroups over a 10-year period. Initial Cox regression confirmed a significant association between infection and mortality in the total cohort with and without adjustment for AIS grade, and stratification by AIS grade revealed an association in AIS A and B. Further adjustment for level of injury and sociodemographic variables (age, working status, and educational level) confirmed the significant association in the complete cohort (hazard ratio 1.65, 95% CI 1.26–2.16) and the AIS A stratification.
INTERPRETATION
This work indicates that pneumonia and postoperative infections in acute care after SCI negatively affect both long-term functional recovery and survival. This study was conducted using an observational database. Although this design introduces the possibility of additional factors related to the exposure (infections) influencing the outcomes, the study design is suitable for the nature of the study as it is not possible to randomize individuals to the study exposure.
The authors provide a strong statistical analysis investigating the association between early infections after SCI and long-term functional recovery and mortality utilizing multiple imputation to account for attrition bias. Attrition, or loss to follow-up, can occur for numerous reasons, and is of particular concern in longitudinal studies.5 These missing data can introduce bias when the lack of data is related to the outcome measure, and is not missing at random.5 One strategy to avoid this bias is multiple imputation, in which missing values are predicted through other observed values in the dataset, creating multiple imputed datasets, and averaging these imputed estimates.5 This is a common method of accounting for missing data, with varying outcomes.5,6 Presented in conjunction with the complete case sensitivity analysis, the imputed data allow for a transparent interpretation of the results. In addition, the authors also adjust for variables known to be related to functional recovery after SCI (e.g., level of injury, AIS). The results consistently confirm that early infections are associated with impaired functional outcomes after SCI in complete SCI. The authors note that these effects are likely due to pneumonia, as this accounted for the majority of recorded infections.
Based on stratified sensitivity analyses, the authors reported that attrition bias existed in AIS groups B and C, but not A, due to the fact that the imputed and complete case only models produced differing results. This could indeed be due to attrition bias, or inaccuracy in the imputations. It would be of interest to examine whether this was due to a nonrandom dropout in the groups AIS B and C, and not A.
Given that the cohort included patients enrolled between 1995 and 2005, it would be of interest to know how the management of care after SCI (e.g., infections and treatments for infections) changed throughout this time frame,7,8 or if year of admission differed significantly between the groups. Perhaps related to care, length of acute care stay has also been shown to affect functional outcomes after SCI, and could provide further insight.9,10
Regarding inclusion/exclusion criteria, the authors provided valid reasoning for the inclusion of specific injury characteristics (limiting heterogeneity and ceiling effects in recovery), though no reason was given for the exclusion of individuals in prior care. Second, patients who were rehospitalized for unspecified infectious or parasitic diseases during follow-up were also excluded, and a justification was also not provided. Both prior care and rehospitalization could potentially confound the relationship between infection and function/survival, which may have been the reason for their exclusion. However, it would be interesting to quantify how these factors affected the study outcomes in a sensitivity analysis. Third, of all the individuals who were enrolled in NSCID and met the initial selection criteria (n = 1,427), 1,203 (84%) remained after the application of the exclusion criteria. Exclusion was largely due to age and missing baseline assessments. Although the proportion of patients recruited was high, it would be of interest to know if those excluded due to missing data differed significantly with regards to injury characteristics or outcomes.
Overall, this article provides a strong statistical analysis of the available data and presents a comprehensive report of the outcomes. Future analyses on novel cohorts and potential confounding and mediating variables could increase our understanding of how acute infections after SCI are related to long-term functional recovery.
Supplementary Material
Footnotes
Supplemental data at Neurology.org
AUTHOR CONTRIBUTIONS
F.M. Warner: drafting/revising the manuscript, analysis or interpretation of data. B. Tong: revising the manuscript, analysis or interpretation of data. Dr. Jutzeler: revising the manuscript, analysis or interpretation of data. Dr. Cragg: revising the manuscript, analysis or interpretation of data. P.S. Scheuren: revising the manuscript. Dr. Kramer: drafting/revising the manuscript, analysis or interpretation of data.
STUDY FUNDING
No targeted funding reported.
DISCLOSURE
The authors declare no competing financial interests. Go to Neurology.org for full disclosures.
REFERENCES
- 1.DeVivo MJ, Kartus PL, Stover SL, Rutt RD, Fine PR. Cause of death for patients with spinal cord injuries. Arch Intern Med 1989;149:1761–1766. [PubMed] [Google Scholar]
- 2.Failli V, Kopp MA, Gericke C, et al. Functional neurological recovery after spinal cord injury is impaired in patients with infections. Brain 2012;135:3238–3250. [DOI] [PubMed] [Google Scholar]
- 3.Brommer B, Engel O, Kopp MA, et al. Spinal cord injury-induced immune deficiency syndrome enhances infection susceptibility dependent on lesion level. Brain 2016;139:692–707. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kopp MA, Watzlawick R, Martus P, et al. Long-term functional outcome in patients with acquired infections after acute spinal cord injury. Neurology 2017;88:892–900. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Biering K, Hjollund NH, Frydenberg M. Using multiple imputation to deal with missing data and attrition in longitudinal studies with repeated measures of patient-reported outcomes. Clin Epidemiol 2015;7:91–106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Twisk J, De Boer M, De Vente W, Heymans M. Multiple imputation of missing values was not necessary before performing a longitudinal mixed-model analysis. J Clin Epidemiol 2013;66:1022–1028. [DOI] [PubMed] [Google Scholar]
- 7.Gerndt SJ, Rodriguez JL, Pawlik JW, et al. Consequences of high-dose steroid therapy for acute spinal cord injury. J Trauma 1997;42:279–284. [DOI] [PubMed] [Google Scholar]
- 8.Berger A, Edelsberg J, Oster G, Huang X, Weber DJ. Patterns of initial antibiotic therapy for community-acquired pneumonia in U.S. hospitals, 2000 to 2009. Am J Med Sci 2014;347:347–356. [DOI] [PubMed] [Google Scholar]
- 9.Richard-Denis A, Feldman D, Thompson C, Mac-Thiong JM. Prediction of functional recovery six months following traumatic spinal cord injury during acute care hospitalization. J Spinal Cord Med 2017:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ahn H, Singh J, Nathens A, et al. Pre-hospital care management of a potential spinal cord injured patient: a systematic review of the literature and evidence-based guidelines. J Neurotrauma 2011;28:1341–1361. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.