- Top of page
- Disclosure of Conflicts of Interests
- Supporting Information
Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disorder, characterized by the loss of upper and lower motor neurons. This leads to impairment of motor function at spinal and bulbar level and ultimately causes death within 2–5 years due to respiratory failure.
Patients with ALS present with different phenotypes and histories, so that distinct etiologies have been proposed. ALS is increasingly recognized to be not only a disorder of the motor system but to involve also non-motor areas of the brain and even to affect whole-body physiology .
For several comorbidities, for example, cardiovascular, metabolic, and neuropsychiatric diseases, both differences in prevalence between patients with ALS and the general population and a specific impact of these disorders on ALS progression have been described [2-8]. It is an often reported observation that patients with ALS prior to disease onset are slim, physically fit, and without cardiovascular risk factors [4, 5]. It has been controversially debated whether metabolic alterations, in particular regarding lipid metabolism, are associated with ALS and whether hypercholesterolemia (HC)/hyperlipidemia slows disease progression [6, 7, 9]. A wide range of behavioral abnormalities, ranging from subtle frontal executive dysfunction to manifest frontotemporal dementia (FTD), can occur in patients with ALS, and recent genetic and neuropathologic evidence (mutations in the genes coding for TDP-43 and FUS/TLS as well as the most recently identified hexanucleotide repeat expansion in C9orf72; detection of neuronal cytoplasmic TDP-43-positive aggregates in post-mortem tissue in both ALS and FTD) confirms the link between ALS and FTD [2, 10, 11]. The rate of depression in patients with ALS is, most commonly, estimated to lie between 24% and 28% [12, 13].
Our study assesses the frequency of comorbidities such as cardiovascular diseases and risk factors as well as neuropsychiatric disorders and their impact on survival and disease progression in patients with ALS.
- Top of page
- Disclosure of Conflicts of Interests
- Supporting Information
We retrospectively analyzed a cohort of 514 patients with ALS from our ALS outpatient database (covering a period of 15 years). Patient consent forms have been received from all patients, and no further approval was needed according to the local regulations.
All patients had met the revised El Escorial criteria for probable or definite ALS within the observation period. The initial evaluation in our ALS clinic routinely includes the medical history, current medication, and the time of disease onset as well as the body region of the first symptoms. Amyotrophic Lateral Sclerosis Functional Rating Scale revised (ALSFRS_R) score, progression of symptoms as well as newly developed comorbidities or additional medications are recorded at each visit. The ALSFRS_R score ratio, which means the deterioration of the score per month, is used for the estimation of disease progression .
The prevalence of comorbidities in our ALS cohort was compared with the data of the German general population in different age groups (sources: Robert Koch institute, German society of Alzheimer's disease, German society of epileptology, competence network Parkinson's disease, and publications cited in accordant text passages). These sources provided the prevalence data for the distinct comorbidities/risk factors of the general population, subdivided in slightly different age groups. Therefore, patients of our cohort were divided according to age groups for each comorbidity, respectively. Significant differences of prevalence of the several comorbidities between the general population and age-matched patients with ALS were identified by non-parametric two-sided 95% Clopper Pearson confidence intervals which were compared to the respective prevalence in the general population as a fixed value.
Survival analysis was performed using Cox regression, and the differences in survival were measured by hazard ratios (HR) with corresponding 95% confidence intervals. First, univariate Cox regressions were conducted with only one independent factor in each regression. To adjust for known factors that influence the survival in ALS, we also performed multivariate regressions: a basic model with the prognostic factors ‘age at diagnosis’, ‘region of onset’, and ‘time from first symptom to diagnosis’ was used, in which each comorbidity was included separately (Table 1A). Kaplan–Meier survival plots were used to illustrate parts of the results.
Table 1. Effect of basic prognostic factors and comorbidities on survival (A) and progression (B) in univariate and multivariate analyses
| ||Univariate analysis||Multivariate analysis (adjusted for age, bulbar onset and time to diagnosis = basic model)|
| P ||HR [95%-CI]|| P ||HR [95%-CI]|
|Gender (f/m)||0.16||1.16 [0.94, 1.42]|| || |
|Age at diagnosis (years)|| <0.001 ||1.03 [1.02, 1.04]|| || |
|Bulbar onset (bulbar/spinal)|| <0.001 ||1.49 [1.19, 1.86]|| || |
|Time from first symptom to diagnosis (month)|| <0.001 ||0.989 [0.980, 0.998]|| || |
|Hypertension|| 0.001 ||1.41 [1.14, 1.75]||0.31||1.13 [0.90, 1.41]|
|Cardiac arrhythmia||0.70||0.94 [0.67, 1.31]||0.46||0.88 [0.63, 1.24]|
|Coronary heart disease (CHD)|| 0.035 ||1.44 [1.03, 2.03]||0.16||1.28 [0.91, 1.82]|
|Myocardial infarction (MI)|| 0.032 ||1.53 [1.04, 2.24]||0.24||1.27 [0.85, 1.87]|
|Diabetes mellitus (DM)||0.39||1.18 [0.81, 1.70]||0.60||1.11 [0.76, 1.60]|
|Hypercholesterolemia (HC)||0.39||1.11 [0.87, 1.43]||0.69||1.05 [0.82, 1.36]|
|Depressive symptoms||0.87||0.98 [0.78, 1.24]||0.64||0.95 [0.75, 1.19]|
|Dementia||0.53||1.14 [0.76, 1.71]||0.21||1.31 [0.87, 1.97]|
|Parkinson||0.17||0.59 [0.28, 1.26]||0.17||0.58 [0.27, 1.25]|
|Epilepsy|| 0.047 ||0.25 [0.06, 0.98]|| 0.045 ||0.24 [0.06, 0.97]|
| || P ||Mean ratio [95%-CI]|| P ||Mean ratio [95%-CI]|
|Gender (f/m)||0.37||0.90 [0.74, 1.12]|| || |
|Age at diagnosis (years)|| 0.015 ||1.01 [1.00, 1.02]|| || |
|Bulbar onset (bulbar/spinal)|| 0.012 ||1.34 [1.07, 1.68]|| || |
|Time from first symptom to diagnosis (month)|| <0.001 ||0.99 [0.98, 0.99]|| || |
|Hypertension||0.71||1.04 [0.84, 1.31]||0.73||0.96 [0.76, 1.21]|
|Cardiac arrhythmia||0.49||0.89 [0.64, 1.23]||0.39||0.87 [0.63, 1.20]|
|CHD||0.58||1.12 [0.76, 1.62]||0.56||1.12 [0.76, 1.65]|
|MI||0.23||1.28 [0.85, 1.92]||0.35||1.21 [0.81, 1.80]|
|DM||0.50||1.14 [0.77, 1.68]||0.71||1.07 [0.74, 1.57]|
|HC||0.20||1.19 [0.91, 1.54]||0.33||1.14 [0.88, 1.46]|
|Depressive symptoms||0.15||0.84 [0.66, 1.07]||0.36||0.90 [0.70, 1.14]|
|Dementia||0.58||0.89 [0.58, 1.36]||0.75||1.07 [0.70, 1.63]|
|Parkinson|| 0.03 ||0.41 [0.19, 0.90]|| 0.04 ||0.46 [0.22, 0.98]|
|Epilepsy||0.22||0.61 [0.28, 1.34]||0.29||0.66 [0.31, 1.40]|
The primary population for survival analysis (Kaplan–Meier survival plots, multivariate regression) consisted of all patients with at least two visits and an observation time over at least 6 months (n = 488). Survival times were calculated as the time from diagnosis to death (n = 360), tracheostomy (n = 6), or censoring (n = 122). Patients were censored at the last recorded time point of contact (visit or phone call) when they were lost to follow up (n = 58), had died of another cause than ALS (n = 12), or were still alive at the time of analysis (n = 52). For sensitivity analyses, calculations were also made with survival time defined as the time from symptom onset to death, tracheostomy, or censoring.
To estimate the influence of a specific comorbidity on the course of disease, linear regression analysis was carried out, using the logarithmic ALSFRS_R score ratio as the dependent variable and the respective comorbidity as the predictor variable. Similar to the survival analysis, all comorbidities were assessed as additional independent covariates in the aforementioned basic model to adjust for known prognostic factors. Differences in ALSFRS_R score ratio are measured as mean ratio with corresponding 95% confidence intervals (Table 1B).
Statistical analyses were performed using SPSS 19.0 (SPSS, Chicago, IL, USA). R 2.14.1 was used for prevalence bar plots with Clopper Pearson intervals. Because multiple tests were performed without correction of the type one error, P-values are exploratory and P-values < 0.05 are considered to be statistically significant.