Survival of Patients with Amyotrophic
Lateral Sclerosis in a Population-Based
Registry
A. Millul E. Beghi G. Logroscino A. Micheli E. Vitelli A. Zardi
for the ‘Registro Lombardo SLA’ (SLALOM)
1
Istituto di Ricerche Farmacologiche ‘Mario Negri’ di Milano, Milano , Italia
od in the whole sample, by level of diagnostic certainty,
and by selected prognostic indicators (age, sex, bulbar
or spinal onset, and disease duration). Multivariate analysis
was done with the Cox proportional hazard function.
Results:
The sample comprised 79 patients (33 female;
46 male) aged 28–85 years (mean age 64.4 years). Onset
of symptoms was bulbar in 30% of cases. Mean symptom
duration at diagnosis was 13.3 months. ALS was
defi nite in 43%, probable in 29%, possible in 6%, and
suspected in 22%. By December 31, 2002, 56 cases (71%)
had died. The cumulative probability of surviving after
diagnosis was 78% at 12 months, 56% at 24 months, and
32% at 48 months. Median survival from onset was 39.2
months and from diagnosis 30.6 months. Multivariate
analysis confi rmed defi nite ALS at diagnosis and older
age as adverse prognostic factors.
Conclusions: Survival
of ALS patients in the present sample was slightly longer
than previously reported. Better palliative care and supportive
treatment may explain the difference. Older age
and the presence of defi nite ALS at diagnosis are poor
prognostic predictors.
Copyright © 2005 S. Karger AG, Basel
Introduction
Amyotrophic lateral sclerosis (ALS) is a progressive
disease causing degeneration of the upper and lower motor
neurons, and average survival is between 2 and 3 years
Key Words
Amyotrophic lateral sclerosis
Mortality
Prognostic predictors
Abstract
Objective:
To evaluate the survival of patients with amyotrophic
lateral sclerosis (ALS) in an Italian population and
to assess the effects of selected prognostic indicators on
survival.
Background: Median survival of ALS patients
has been reported to range between 12 and 23 months
from diagnosis and between 23 and 36 months from onset
of symptoms. Although several negative prognostic
factors have been identifi ed, the overall picture still
needs clarifi cation.
Methods: We included patients enrolled
in an Italian ALS Regional Register (population
4,529,003) during the calendar year 1998. The diagnosis
was confi rmed by an ad hoc committee using the original
El Escorial criteria. Each case was regularly followed
up until death or December 31, 2002, whichever came
fi rst. Survival was assessed with the Kaplan-Meier meth-
Published online: June 13, 2005
Dr. Ettore Beghi
Istituto di Ricerche Farmacologiche ‘Mario Negri’
Via Eritrea 62
IT–20157 Milano (Italy)
Tel. +39 02 39014542, Fax +39 02 33200231, E-Mail beghi@marionegri.it
© 2005 S. Karger AG, Basel
0251–5350/05/0253–0114$22.00/0
Accessible online at:
www.karger.com/ned
1
SLALOM: D. Alimonti, D. Baldini, C. Balzarini, C. Bendotti, G. Bianchi,
G. Bogliun, V. Bonito, A. Brambilla, L. Brunati, R. Causarano, M. Ceroni, A.
Cheldi, A. Chiò, L. Chiveri, A. Citterio, M. Clerici, G. Comi, M. Corbo , M.L.
Delodovici, E. Donati, C. Ferrarese, G. Filippini, P. Gambaro, M. Guidotti, I.
La Spina, P. Liberini, G. Mariani, T. Mennini, G. Meola, S. Moroni, E. Munerati,
M. Perini, P. Perrone, M. Poloni, M. Porta, A. Prelle, M. Rezzonico, R.
Riva, A. Romorini, F. Sasanelli, L. Serlenga, V. Silani, D. Testa, F. Tavernelli,
M.C. Tonini.
Survival of Patients with ALS
Neuroepidemiology 2005;25:114–119 115
[1–3] . However, the clinical spectrum of the disease is
wide and tends to refl ect the heterogeneity of the symptoms
at onset and the variability of its course. Information
on the prognosis of ALS has mostly been obtained
from population-based studies or large clinical series [1,
4–16] . In these studies, survival of ALS patients varied
signifi cantly depending on the methods used for case ascertainment,
the different diagnostic criteria, and the
small numbers of cases surviving beyond the fi rst few
years. In three prospective population-based studies [9,
14, 16] , several negative prognostic factors were identifi
ed, including older age, female sex, lack of marital partner,
bulbar features at onset, level of diagnostic certainty,
progressive bulbar palsy, short interval between symptom
onset and diagnosis, progression rate of respiratory, bulbar
and lower limb symptoms, and overall disease severity.
However, the fi ndings are only partly consistent. We
therefore arranged a population-based study of newly diagnosed
ALS patients with the following aims: (1) to defi
ne survival of ALS patients after a long observation period,
and (2) to verify the role of selected putative clinical
predictors of long-term survival.
Materials and Methods
Eligible patients included all ALS cases in selected geographic
areas of Lombardy, a largely industrialized region in Northern Italy
( fi g. 1 ). For each designated area, the choice was mostly based
on the accuracy of case ascertainment, the case of source identifi cation,
and the presence of reliable health care settings. Patients were
identifi ed through a prospective regional registry of ALS. The study
population amounted to 4,529,003 (1991 national census). Participants
in the registry were neurologists working in the designated
areas who screened all patients admitted to local hospitals and discharged
with a diagnosis of ALS. The study cohort included patients
ascertained during the period from January 1 to December
31, 1998. To assure complete case ascertainment, all the hospital
discharge diagnoses were examined and the medical records of patients
discharged with an ICD code 335.2 (motor neuron disease)
were reviewed. Neurophysiology units were inspected searching for
any electrophysiological records which suggest motor neuron involvement.
Discharge diagnoses and mortality statistics for the
years 1998 through 2002 were also examined for cases with ICD
code 335.2.
For each eligible case, the neurologists were asked to collect the
main demographic, clinical, electrophysiological and (where needed)
neuroimaging fi ndings on ad hoc semistructured forms. An effort
was made to obtain details from a thorough neurological examination,
focused on the main regions (bulbar, cervical, thoracic,
and lumbosacral) and the involvement of the upper and lower motor
neurons. The diagnosis of ALS was based on the original El
Escorial criteria [17] and reviewed by an expert panel. Age, sex, site
of fi rst symptoms (bulbar or spinal), disease duration at diagnosis,
Italy
Varese
Como
Lecco
Sondrio
Bergamo
Brescia
Lodi
Pavia
Milano
Cremona
Mantova
Fig. 1.
Map of Lombardy (study areas in black).
Millul/Beghi/Logroscino/Micheli/Vitelli/
Zardi
116
Neuroepidemiology 2005;25:114–119
and diagnostic category at study entry were indicated as the prognostic
predictors. Each patient with ALS was then followed until
death or December 31, 2002, whichever came fi rst. Suspected or
possible ALS cases were revised during follow-up and patients who
shifted to different diagnostic categories were identifi ed. Death certifi
cates were examined for diagnostic confi rmation of all possible
and suspected ALS cases who died in these stages.
The signifi cance of variables was tested using the
2 test, Student’s
t test and univariate analysis of variance as appropriate.
Survival of the entire cohort from diagnosis and from onset of
symptoms was estimated using the Kaplan-Meier method. The logrank
test was used to measure differences in survival from diagnosis
for each prognostic predictor. The Cox proportional hazard
model was used to calculate the adjusted hazard ratio, with different
prognostic factors independently (age, sex, disease duration at
diagnosis, site of onset, and diagnostic category). A backward stepwise
(likelihood ratio) method was used. The proportional hazard
model was fi tted using the method of maximum likelihood and the
log-cumulative hazard plot. The El Escorial diagnostic categories
were included as dummy variables. Interactions between age and
site of onset, sex and disease duration, and site of onset and disease
duration were also included in the model. Signifi cance was tested
at the 5% level. No missing values were present for each variable
included in the model. Calculations were done using SPSS (version
11.0 SPSS Inc., Chicago, Ill., USA).
Results
The study population included 79 patients (33 women
and 46 men aged 28–85 years). Mean age at diagnosis was
64.4 years (women 67.1; men 62.3; p = 0.14). Bulbar-onset
ALS was diagnosed in 24 cases (30%) and spinal-onset
ALS in 55 (70%). Mean age at onset was similar in bulbar-
onset and spinal-onset ALS patients (67.5 vs. 63.0;
p = 0.30). The two forms were evenly distributed between
the two sexes. However, bulbar-onset ALS was present in
18.8% of patients aged less than 55 and in 33.3% of patients
55 years and older (p = 0.21). Mean disease duration
at diagnosis was 13.2 months (range 1–86 months).
At study entry, defi nite ALS was diagnosed in 34 cases
(43%), probable ALS in 23 (29%), possible ALS in 5 (6%),
and suspected ALS in 17 (22%). The distribution of the
upper and lower motor neuron signs in the four regions
is given in table 1 , where the fi ndings are reported separately
for the four diagnostic categories. As expected, the
number of regions with abnormal fi ndings tended to increase
with the level of diagnostic certainty. Patients with
suspected ALS had fasciculations (n = 15) and hypo-arefl
exia (n = 12). The diagnosis was revised in all patients
Table 1.
Distribution of upper (UMN) and lower motor neuron (LMN) signs in the four regions by the El Escorial diagnostic category
Region ALS suspected (n = 17) ALS possible (n = 5) ALS probable (n = 23) ALS defi nite (n = 34) ALS total (n = 79)
UMN LMN UMN LMN UMN LMN UMN LMN UMN LMN
Bulbar – 8 (47.0) – 3 (60.0) 8 (34.8) 13 (56.5) 19 (55.9) 25 (73.5) 29 (36.7) 49 (62.0)
Cervical – 10 (58.8) 1 (20.0) 3 (60.0) 16 (69.5) 22 (95.6) 27 (79.4) 30 (88.2) 50 (63.3) 65 (82.3)
Thoracic – 2 (11.8) – 2 (40.0) 2 (8.7) 5 (21.7) 8 (23.5) 11 (32.5) 10 (12.6) 20 (25.3)
Lumbosacral – 12 (70.5) 2 (40.0) 3 (60.0) 18 (78.3) 19 (82.6) 25 (73.5) 28 (82.4) 54 (68.4) 62 (78.5)
Figures in parentheses indicate percentages.
ALS suspected
(n =17)
ALS possible
(n = 5)
ALS definite
(n = 4)
(dead 3)
ALS possible
(n = 1)
(dead)
ALS suspected
(n = 1)
(alive)
ALS possible
(n = 3)
(dead 1)
ALS probable
(n = 3)
(dead 2)
ALS definite
(n = 5)
(dead 4)
Lost to follow up
(n = 5)
(dead 3)
Fig. 2.
Diagnostic shift during follow-up
(with status) in patients with suspected and
possible ALS.
Survival of Patients with ALS
Neuroepidemiology 2005;25:114–119 117
with possible ALS and in 12 with suspected ALS ( fi g. 2 ).
Of these, all but 2 (1 with possible ALS and 1 with suspected
ALS) moved to an upper diagnostic category. Eleven
patients with suspected ALS developed upper motor
neuron signs. At the last follow-up, there were 6 patients
with suspected ALS (5 lost to follow-up) and 4 patients
with possible ALS. Five of them (3 with suspected ALS
and 2 with possible ALS) were still alive at the last followup.
In the 3 patients with suspected ALS, disease duration
was 62.4, 61.4 and 99.4 months from symptom onset
(mean 74.5 months) and 49.2, 57.4 and 60.3 months from
diagnosis (mean 55.6 months). Both patients with possible
ALS at the last follow-up had upper and lower motor
neuron signs in one region.
The entire cohort was followed for 2,480 personmonths
(mean 31.4; range 2–60). By the date of the last
follow-up, a total of 56 patients had died (71%). ALS was
reported as the cause of death in all the death certifi cates.
Survival status at the last follow-up by prognostic predictors
is depicted in table 2 . Patients who died were mostly
older, with bulbar-onset ALS, and with shorter disease
duration at diagnosis. The cumulative time-dependent
survival rate at symptom onset and at diagnosis is depicted
in fi gure 3 . Survival from onset of symptoms was
92% at 12 months, 73% at 24 months, and 43% at 48
months. The corresponding values for survival from diagnosis
were 78, 56, and 32%. Median survival from
symptom onset was 39.2 months [95% confi dence interval
(CI) 28.2–49.8] and median survival from diagnosis
was 30.6 months (95% CI 19.5–41.6).
Survival from diagnosis by different prognostic indicators
is depicted in fi gure 4 . Median survival from onset
was 52.1 months in patients aged less than 55 years, 48.5
months in those aged 55–74 years, and 16.4 months in
those aged 75 years or older (p
! 0.0005). The corresponding
values for survival after diagnosis were 35.1, 38.8, and
8.1 months (p
! 0.0002). Mean (SD) disease duration at
diagnosis was 15.1 (21.9) months in patients aged less
than 55 years, 14.1 (15.9) months in those aged 55–74
years, and 8.8 (7.2) months in those aged 75 years and
older (p = 0.47). In bulbar-onset ALS, median survival
from onset was 26.6 months (spinal-onset ALS 45.9
months; p
! 0.05). Median survival from diagnosis of
bulbar- and spinal-onset ALS patients was 18.7 and 35.1
months, respectively (p
! 0.05). Sex and disease duration
did not affect survival to any signifi cant extent. Median
survival after diagnosis of defi nite ALS patients was 16.3
months. The corresponding values were for probable
ALS, 31.0 months; possible ALS, 40.3 months, and suspected
ALS, 48.8 months (p = 0.16). Multivariate analy-
Table 2.
Survival status of ALS by prognostic predictors
Variable Dead Alive
n % n %
Total 56 71 23 29
Age groups*
<55 years 9 56 7 44
55–74 years 33 70 14 30
>74 years 14 87 2 13
Sex
Men 32 70 14 30
Women 24 73 9 27
Onset*
Bulbar 21 87 3 13
Spinal 35 64 20 36
Duration (months) of symptoms at diagnosis**
<6 27 82 6 18
7–12 15 75 5 25
13–24 8 57 6 43
>24 6 50 6 50
El Escorial category at diagnosis
Defi nite ALS 27 79 7 21
Probable ALS 15 65 8 35
Possible ALS 4 80 1 20
Suspected ALS 10 59 7 41
* p < 0.05; ** p < 0.01.
100
80
60
40
20
0
0 1 2
Duration (years)
3
From diagnosis
Survival (%)
From symptom onset
4 5
Fig. 3.
Cumulative survival of ALS patients from symptom onset
and from diagnosis.
Millul/Beghi/Logroscino/Micheli/Vitelli/
Zardi
118
Neuroepidemiology 2005;25:114–119
sis showed that increasing age and defi nite ALS (hazard
ratio 2.5; 95% CI 1.4–4.2) were the only independent predictors
of mortality.
Discussion
In this population-based prospective study of newly
diagnosed ALS, survival from diagnosis was 78, 56, and
32% at 12, 24, and 48 months. Median survival from diagnosis
was 30.6 months and median survival from onset
was 39.2 months. Our median survival is slightly longer
than that of other reports where it ranged from 12 to 23
months from diagnosis and from 23 to 36 months from
onset [1, 4–11, 13–16]. Although, based on the small sample
size, the possibility of a chance fi nding cannot be excluded
(see confi dence intervals of our rates), the longer
survival of our patients might be explained by earliest
diagnosis and/or by better palliative care and supportive
treatment (e.g. enteral nutrition and BiPap mechanical
100
80
60
Survival (%)
40
20
0
0 1 2 3
Duration (years)
<55 years
Male
Female
>74 years
Spinal
Possible
Suspected
Definite
Probable
Bulbar
55–74 years
4 5
100
80
60
Survival (%)
40
20
0
0 1 2 3
Duration (years)
4 5
100
80
60
Survival (%)
40
20
0
0 1 2 3
Duration (years)
4 5
100
80
60
Survival (%)
40
20
0
0 1 2 3
Duration (years)
4 5
a b
c d
Fig. 4.
Cumulative survival of ALS patients from diagnosis by different prognostic indicators. a Age. b Sex.
c
Type of onset. d El Escorial diagnostic category.
Survival of Patients with ALS Neuroepidemiology 2005;25:114–119 119
No comments:
Post a Comment