Article Text
PDF
Abstract
Objectives To describe the natural history and clinical features of sporadic amyotrophic lateral sclerosis (ALS) in Chinese patients, and to report data on the prognostic factors for survival.
Methods All patients referred to our ALS centre between 2003 and 2012 were followed up every 3 months. Survival and tracheotomy were predefined as primary outcome measures. Group differences were analysed using parametric and non-parametric tests as appropriate. Survival was analysed using the Kaplan-Meier method and Cox regression analysis.
Results Of the 1624 patients with ALS, 75.1% had limb-onset, 14.0% had bulbar-onset, 7.8% had flail-arm syndrome (FAS), 2.6% had progressive muscular atrophy and 0.5% had primary lateral sclerosis. The male:female ratio was 1.7:1, and the mean age at onset was 49.8 years. The median diagnostic delay was 14 months, and the median survival time after symptom onset was 71 months. Male gender, older age at symptom onset, lower body mass index, shorter diagnostic delay, bulbar-onset ALS phenotype, higher Airlie House category at presentation, rural place of residence, use of traditional Chinese medicine and a history of contact with pesticides were associated with poorer survival, whereas female gender or an FAS phenotype may have a better prognosis.
Conclusions The clinical characteristics and outcomes of Chinese patients with sporadic ALS were different compared with patients from other countries. Compared with other studies, the age at onset of Chinese patients was earlier, the percentage of bulbar-onset ALS was lower and the prognosis was better. This study substantially advances the understanding of the clinical features and epidemiology of this rare disease.
Statistics from Altmetric.com
Introduction
Amyotrophic lateral sclerosis (ALS) is characterised by the progressive degeneration of upper and lower motor neurons; its typical clinical features include limb paralysis, muscle atrophy, dysphagia, dysarthria, shortness of breath and respiratory failure.1–3 There are several phenotypes of ALS, including limb-onset ALS, bulbar-onset ALS, progressive muscular atrophy (PMA), primary lateral sclerosis (PLS) and several regional limb variants, such as flail-arm syndrome (FAS).4 ,5 Although most studies report a median survival time of 3–5 years for patients with ALS, the disease progresses rapidly in some patients, while the survival time of some patients is very long (over 120 months).1–4 ,6 ,7
The aetiology of ALS remains unknown, and its prognostic factors have not been satisfactorily defined.1 Researchers have found that the survival of patients with ALS is related to several factors, including the clinical phenotype, age at onset, gender, early presence of respiratory failure and treatment with riluzole.1 ,3 A better understanding of the prognostic factors of ALS may guide therapeutic interventions and the design of clinical trials.1 However, the natural history and clinical features of ALS have never been studied in mainland China. Therefore, we investigated the characteristics of patients with ALS from a large cohort, and we discuss some of the factors that influence outcomes of ALS.
Methods
Protocol approvals, registrations and patient consent
The institutional ethics committee of Peking University Third Hospital (PUTH) approved this study (IRB 00006761). Written, informed consent was obtained from each patient.
Subjects and definitions
This study was a clinic-based prospective cohort study that analysed information from the ALS patient database of PUTH, the largest ALS centre in China. Patients were recruited from January 2003 to December 2012, and each patient was given a follow-up evaluation by telephone every 3 months. The patients were from throughout China. For all cases, baseline demographic information and clinical data were collected during the patient's first visit to PUTH and during follow-up evaluations. Survival and tracheotomy were predefined as primary outcome measures.
Patients were diagnosed and classified according to the Airlie House diagnostic criteria.8 Those with pure lower motor syndromes were classified into an additional category of suspected ALS because they could not be classified using these established criteria.5 All patients were interviewed and examined by board-certified neurologists from the study group who had experience with motor neuron diseases. Each patient was independently examined by two neurologists, and if their diagnoses or disease categories differed, a third and more experienced neurologist examined the patient to make a final determination. According to the site of onset and clinical features, patients were categorised with limb-onset ALS, bulbar-onset ALS, FAS, PMA or PLS.4 ,5 Patients presenting with a pure lower motor neuron syndrome not characteristic of the FAS phenotype were classified into the PMA or limb-onset ALS categories, depending on their subsequent clinical evaluations.
In the database, ‘familial ALS’ was defined as patients with at least one more patient with ALS in their family, and ‘juvenile ALS’ was defined as patients with an age at onset less than 18 years. Since the aetiology and prognosis of familial or juvenile ALS can be quite different from sporadic ALS, patients with familial and juvenile ALS were excluded from the analysis. ‘Lost to follow-up’ was defined as a change in the telephone number of a patient or refusal of a patient to participate in our follow-up evaluation more than three times. ‘Diagnostic delay’ was defined as the time from symptom onset to a confirmed diagnosis of ALS made by a board-certified neurologist. ‘Contact with pesticides’ was defined as selling, delivering or spraying pesticides on at least one occasion in daily life. ‘Alcohol abuse’ was defined as consuming an alcoholic drink more than twice a week for more than 1 year. ‘Use of riluzole’ was defined as treatment with riluzole (50 mg) twice a day for longer than 2 weeks. ‘Use of traditional Chinese medicine (TCM)’ was defined as treatment with TCM for more than 2 weeks with different dosages and types of herbs after disease onset. ‘Use of vitamins’ was defined as taking vitamin B1, B12, C or E, or multivitamins for more than 2 weeks after disease onset, and prior use was considered as a relevant factor. A ‘Family history’ of neurodegenerative diseases and cancer was defined as a history of such diseases in second-degree relatives.
Statistical analysis
Paper copies of case report forms were returned from the study group on a weekly basis. Data were independently entered into our database by two investigators and examined by two other investigators to avoid mistakes. Significance was tested at the 5% level, and all analyses were performed using the SPSS V.16.0 software package (SPSS, Chicago, Illinois, USA). Q–Q plots were used to determine whether data were normally distributed. Diagnostic delay data were normalised by a log transformation because these data were non-normally distributed. Continuous clinical and demographic variables that were normally distributed were compared using parametric tests (one-way analysis of variance (ANOVA) or a Student t test) between subgroups, and categorical variables were analysed using non-parametric tests (χ2 test, Fisher's exact test, Kruskal-Wallis one-way ANOVA by ranks, or Mann-Whitney U test). The censoring date for the survival data was 1 May 2013. Patients lost to follow-up were censored at the last known living date. Survival curves were estimated using the Kaplan-Meier analysis, and covariates were compared using the log-rank test and Cox regression analysis. The 5-year and 10-year survival rates were calculated by life-table analysis. Since no patient in the PLS group had died, this phenotype was excluded from the survival analysis when the survival times of the different phenotypes were examined.
Results
Features of the cases in the database
We screened 1900 patients, and a total of 1624 patients met the inclusion criteria for this study (figure 1). A small percentage of the patients (17.8%) was local from Beijing, and the remaining patients (82.2%) were recruited from other parts of China. The demographic and clinical characteristics of the study subjects are shown in table 1. The male:female (M:F) ratio was 1.7:1. The mean age at symptom onset was 49.8 years, and the percentage of patients between 50 and 59 years of age (31.9%) was higher than that of any other age range. The mean body mass index (BMI) was 22.9 kg/m2. The median diagnostic delay was 14.0 months. By May 2013, 606 patients had died.
Demographic and clinical characteristics of the patients with ALS examined in this study
Study flow diagram (ALS, amyotrophic lateral sclerosis).
We found that 28.5% of the patients were smokers, which was a higher rate compared with the entire Chinese population in 2002 (24%);9 19.6% had a history of alcohol abuse compared with 21% of the overall Chinese population in 2002,9 and 16.4% had a history of contact with pesticides. Riluzole was used by 32.3% of the patients, and only 10% were treated with non-invasive positive pressure ventilation (NIPPV); 40.3% of the patients took vitamins, and 11.3% had used TCM. In the survival analysis, the median survival time for all patients was 71 months. The 5-year survival rate was 49%, and the 10-year survival rate was 32% (table 1).
Comparisons between phenotypes
The characteristics of the different phenotype groups are shown in table 1. Of the 1624 patients included in the analysis, 1220 (75.1%) had the typical limb-onset form of ALS, 227 (14.0%) had the bulbar-onset form, 126 (7.8%) had FAS, 43 (2.6%) had PMA and 8 (0.5%) had PLS. The mean age at symptom onset was different between phenotypes (F=15.339, p<0.001). In the post hoc analysis, the mean ages at onset of the patients with limb-onset ALS (48.7 years) and PMA (47.9 years) were lower than those of the patients with bulbar-onset ALS (53.2 years) (compared with limb-onset ALS, p<0.001; with PMA, p=0.003) and FAS (55 years, both p<0.001). Using a logistic regression and after adjusting for gender as a covariate, the frequency of bulbar-onset ALS and FAS was independently positively correlated with a higher age at symptom onset (both p<0.001), but the limb-onset frequency was negatively correlated (p<0.001). The percentage of patients who had either clinically definite or clinically probable disease at presentation was 53% in the limb-onset ALS group, 59.9% in the bulbar-onset ALS group and 19% in the FAS group. Cases of PMA were classified into an additional category of suspected ALS because they could not be classified according to the established criteria. Some cases of limb-onset ALS, bulbar-onset ALS and FAS that could not be classified using the established criteria at first, but that were confirmed to be ALS during follow-up examinations, were also classified into this additional group. In the post hoc analysis, mean BMI was not significantly different between phenotypes, with the exception of the mean BMI in the bulbar-onset ALS group, which was significantly lower than that in the limb-onset ALS group (p=0.013). The median diagnostic delay was different between phenotypes (p<0.001). In the post hoc analysis, the diagnostic delay was longest in the PLS group (49.0 months; all p<0.001). The shortest diagnostic delay time was found in the bulbar-onset ALS group (12.0 months; compared with the limb-onset ALS group, p=0.048; compared with the PMA group, p=0.056; compared with the FAS group, p<0.001; compared with the PLS group, p<0.001), and the diagnostic delay in the FAS group (19.0 months) was longer than those in the limb-onset ALS (p=0.002) and bulbar-onset ALS (p<0.001) groups. No significant difference in diagnostic delay was observed between the PMA group and the other phenotype groups, with the exception of the PLS group.
The percentage of patients who were smokers differed significantly among the phenotype groups (p=0.013); the highest percentage was observed in the PMA group (46.2%), and the lowest was observed in the PLS group (12.5%). The percentage of patients who had a history of alcohol abuse or a history of contact with pesticides was not different between groups. The percentage of patients who took vitamins was significantly different between phenotype groups (limb-onset ALS 42.8%, bulbar-onset ALS 29.8%, FAS 45.7%, PMA 15.4% and PLS 50%; p=0.001). No significant differences in the percentages of patients receiving treatment with riluzole, NIPPV or TCM were observed between groups.
Since no case reached the end point in the PLS group, this phenotype was not included in the survival analysis. The median survival times of patients were as follows: limb-onset ALS group, 72 months; bulbar-onset ALS group, 48 months; FAS group, 97 months; and PMA group, 110 months. The 5-year and 10-year survival rates were higher in the FAS and PMA groups compared with the limb-onset and bulbar-onset ALS groups. The Kaplan-Meier curves for the four phenotypes were different overall (figure 2). Post hoc tests showed that the median survival time of patients with bulbar-onset ALS was significantly shorter than those of patients in the other phenotype groups (compared with limb-onset ALS, p<0.001; compared with FAS, p=0.001; compared with PMA, p=0.049); no significant difference in the median survival time was observed among the other phenotype groups. After adjusting for gender, post hoc tests for the overall survival time showed that the survival time was shortest in the bulbar-onset ALS group (compared with limb-onset ALS, p<0.001; with FAS, p<0.001; with PMA, p=0.016), while the survival time of patients with FAS was greater than the survival time of those with limb-onset ALS (p=0.039) and bulbar-onset ALS (p<0.001). Since there were fewer cases of PMA, no significant difference in survival time between PMA and limb-onset ALS or FAS was observed (figure 2).
Kaplan-Meier survival curves. (A) Survival analysis stratified by phenotype; (B) survival analysis stratified by gender; (C) survival analysis stratified by phenotype in male patients and (D) survival analysis stratified by phenotype in female patients (ALS, amyotrophic lateral sclerosis; FAS, flail arm syndrome; PMA, progressive muscular atrophy).
Comparisons between males and females
The number of male patients was greater than that of female patients in all phenotype groups. The highest M:F ratio was found in the FAS group (8.0:1), while the lowest M:F ratio was found in the bulbar-onset ALS group (1.3:1). The mean age at onset in females (49 years) was younger than that in males (50.2 years; p=0.048). Neither the difference in the mean BMI (p=0.526) nor the difference in the diagnostic delay (p=0.645) between males and females was significant. The percentage of male patients who had reached the end point by 1 May 2013 was 39.5%, which was significantly higher than that of female patients (33.5%; p=0.017).
A higher percentage of male patients smoked (44.7%, which was lower than the percentage in Chinese male adults (50.2%)) compared with female patients (1.7%, which was also lower than in Chinese female adults (2.8%)) (p<0.001).9 The percentage of patients who had a history of alcohol abuse was also higher in males (males, 30.5% vs females, 1.4%, p<0.001). The percentages of males (15.5%) and females (17.8%) who had a history of contact with pesticides were not significantly different (p=0.356). There were no significant differences in the percentages of males and females using riluzole, NIPPV, vitamins or TCM.
In the survival analysis, the survival time of females (87 months) was significantly longer than that of males (63 months; p=0.008). In the analysis of genders stratified by phenotype, the survival time of females was significantly longer than that of males in the limb-onset ALS (p=0.003) and bulbar-onset ALS (p=0.018) groups, while in the FAS and PMA groups the survival time did not statistically significantly differ between males and females. Among the male patients, the survival time for bulbar-onset ALS was the shortest among all the phenotypes, and the survival time for FAS was the second longest (compared with limb-onset ALS, p=0.012; compared with bulbar-onset ALS, p<0.001), next to that of the PLS group, which was the longest. However, among the female patients, there was no significant difference in survival time between the phenotype groups (figure 2).
Analysis of the prognostic factors of survival
We analysed the following factors to determine which of them were related to disease outcome: gender; age at onset; diagnostic delay; disease phenotype; Airlie House category at presentation; BMI; rural versus urban place of residence; treatments, including riluzole, NIPPV, vitamins and TCM; history of smoking and alcohol abuse; history of contact with pesticides; family history of other neurodegenerative diseases; and family history of cancer.
In the univariate regression analysis (table 2), several factors were found to be predictive of a shorter survival time, including the following patient characteristics: male gender (p=0.009), older age at onset (p<0.001), bulbar-onset ALS phenotype (p<0.001) and living in a rural area (p=0.039). Since a history of contact with pesticides could be biased by urban versus rural residence, pesticide contact was only used in the multivariate analysis. In contrast, a longer diagnostic delay (p<0.001), a higher BMI (p=0.016) and an initial diagnosis of probable ALS supported by laboratory findings (p=0.008) were associated with a better prognosis. After we adjusted for age at onset and gender, factors indicative of a poorer prognosis included a bulbar-onset ALS phenotype (p=0.017) and living in a rural area (p<0.001), while a longer diagnostic delay (p<0.001), a higher BMI (p=0.021), a phenotype of FAS (p=0.001) and an initial diagnosis of probable ALS supported by laboratory findings (p=0.001) or possible ALS (p=0.003) were predictors of longer survival time.
Unadjusted and adjusted univariate survival analysis results
Multivariate Cox regression analysis (table 3) showed that a poorer prognosis was observed with male gender (p=0.004), an older age at disease onset (p<0.001), a bulbar-onset ALS phenotype (p=0.014) and a history of contact with pesticides (p<0.001). Interestingly, the use of TCM seemed to be associated with the trend of poor survival (p=0.080). Additionally, the Cox regression model showed that patients with a longer diagnostic delay (p<0.001), a higher BMI (p=0.032) or a diagnostic category of probable ALS (p=0.034), laboratory-supported probable (p<0.001) or possible ALS (p=0.001) lived longer. Patients who were treated with riluzole did not exhibit a trend of better survival (p=0.143).
Multivariate Cox regression analysis
Discussion
On the basis of a population size of more than 1.3 billion and the incidence rates in other countries,4 over 10 000 cases of ALS in China could exist, although the exact prevalence of ALS in China is currently unknown. Since clinical and epidemiological information about ALS in China is lacking, neurologists use data from other populations to estimate the prognosis of Chinese patients and to guide treatment. Since the genetic background and lifestyle of Chinese patients are different, the direct use of results from other countries may lead to incorrect assessments.
This 10-year prospective, clinic-based cohort study describes, for the first time, the natural history and several important clinical features of ALS cases in mainland China. The mean age at symptom onset was greater than that reported in India (46.2 years)10 and in Hispanics (47.2),11 and lower than that reported in Japan (64.8),12 Italy (64.8)13 and Europe (64.4).14 The overall M:F ratio was comparable with that reported in a population-based study in Taiwan (1.7:1),15 higher than that in Europe (1.2:1)14 and lower than that in Japan (1.96:1).12 The proportion of patients with classic limb-onset ALS was comparable with or higher than that reported in some other studies, but the proportion of patients with bulbar-onset ALS was significantly lower.4 ,5 ,7 ,14 ,16 The median survival time of our patients was similar to that reported in Taiwan (66.6 months).15 The median survival time has been reported to be 114.8 months in India,10 28 months in Italy17 and 68.6 months in Hispanics.11 The percentages of patients who used riluzole and NIPPV were much lower in our study than in other studies,5 ,15 which may be explained by the fact that medical insurance in China does not cover ALS treatment and that most patients have a poor understanding of ALS.
The age at onset of Chinese patients with ALS was about a decade younger compared with European and Japanese patients, and several potential explanations could account for this difference. First, the average life expectancy in China is 74.83 years, which is approximately 5 years shorter than that in developed countries,18 and this difference might partially explain the younger age at onset in our patients. Second, the differences in genetic background and lifestyle might lead to a different disease susceptibility. Third, the frequency of bulbar-onset ALS was much lower in Chinese patients with ALS, while the age at onset of this phenotype was much higher compared with the classic limb-onset ALS.19 Fourth, the factors of ‘social development’, such as pollution and nutrition, may play an important role in the younger age at onset.19 Surely, on the basis of family and personal anxiety, younger patients would seek help more commonly than older patients, and older ones could have more difficulty visiting the clinic, which may create bias. However, a multicentre study in mainland China which recruited 455 patients from 10 centres showed that the mean age at onset for the entire cohort was 52.4±12.1 years, which was similar to that of our present study.20 The finding of this multicentre study made our result convincing and reliable.
It is interesting that although the survival time of female patients did not differ between phenotypes, the overall survival time of females was longer than that of males. Some studies have also reported poorer survival in men,12 ,15 ,21 but other studies have not, which is in contrast with the findings of this study.1 ,10 We also noticed a younger age at onset in females compared with males, but the average life expectancy for females was longer (according to the 2010 national census: 72.38 years for males vs 77.37 years for females).18 Previous studies have shown that smoking is a risk factor for ALS, and that patients who smoke may have shorter survival times as well.22–24 Additionally, some scientists believe that the higher ALS incidence rate in males may be explained by their higher smoking rate, and by the increased chances of exposure to metals and chemicals in the workplace, but the percentage of smokers in male and female patients in our database was lower compared with all Chinese adults in 2002.9 ,25 Moreover, some studies have suggested that the reason why the prognosis of female patients is better than that of male patients is that metabolism and hormone status differ between males and females.25 Moreover, a younger age at onset in females might also contribute to a better prognosis. However, these hypotheses have not been confirmed, and the differences between the two genders need further assessment to determine reasonable explanations for these differences.
The Cox regression model used in this study showed that the predictors of survival were gender, age at symptom onset, BMI, diagnostic delay, disease phenotype, Airlie House category at presentation, rural versus urban place of residence, treatment with TCM and a history of contact with pesticides. Many other predictors, such as treatment with NIPPV, that may influence the survival time and improve the cognitive function of patients with ALS have been found previously,1 ,15 ,26 ,27 but no significant effect of treatment with NIPPV was observed in our analysis. Our study demonstrates that the Airlie House status at diagnosis is directly related to the rate of disease progression, in agreement with the findings of other studies.28 ,29 Since a diagnosis of definite ALS is indicative of more widespread clinical involvement, it is reasonable that a diagnosis of definite ALS at presentation carries a significantly poorer prognosis compared with all other diagnostic categories.1
We analysed several factors that were found to be associated with survival but that have seldom been taken into consideration, such as the rural versus urban nature of patients’ places of residence, the use of TCM and whether patients had a history of contact with pesticides. Since there is still a large gap in economic development levels between urban and rural areas, and ALS treatment is not covered by medical insurance in China, patients from rural areas have less opportunity to visit a neurologist to obtain proper treatment. That a history of contact with pesticides may increase the risk of ALS has been verified by several meta-analyses,30–32 and our study suggests that the use of pesticides may not only contribute to the incidence of ALS but also lead to shorter survival times. In this study, the effects of TCM on survival times were examined for the first time; however TCM treatment was a notable predictor of poorer prognosis. Since the components of TCM vary and the quality of herbs is difficult to standardise, more studies need to be performed before a conclusion can be drawn about the effects of TCM in ALS.
Our study has several limitations. First, some important data, such as gastrostomy data and respiratory status, were absent from our database. Second, owing to the difficulty in sample collection and analysis, important information on genetic factors was not available. Third, this was a clinic-based cohort study, and some studies have shown that the results of clinical cohort-based and population-based studies may not always agree.1 ,17 The validity of our observations based on this clinical database is strengthened by the findings from a population-based study in Taiwan.15 Similar data regarding the natural history suggest that our study and the Taiwanese study included patients with a similar genetic background, despite the biases inherent in any clinical cohort. Owing to the limitations of a single-centre cohort study, a nationwide properly constructed multicentre study is absolutely necessary in future.
In summary, our study is the first to investigate the natural history and clinical features of patients with ALS in China. These findings can increase our understanding of the characteristics of Chinese patients with ALS and aid in patient counselling and in the design of clinical trials.
References
Footnotes
LC and BZ contributed equally.
Contributors DF conceived this study and provided financial support. DF, LC and BZ designed the study. LT, RL, YiY, YaY, XL and SY took part in the design of the study and in sample collection. DF, LC, BZ and LT conducted data management. RL, YiY, YaY, XL and SY conducted data follow-up. LC and BZ undertook data checking. LC, BZ, RC and SZ undertook statistical analysis. DF was responsible for project management. LC, BZ and DF were responsible for preparing and revising the manuscript. DF, LC and BZ had key roles in the study.
Funding This study was supported by the National Natural Science Foundation of China (81030019).
Competing interests None declared.
Patient consent Obtained.
Ethics approval The Institutional Ethics Committee of Peking University Third Hospital.
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- Editorial commentary