Introduction

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease characterized by progressive dysarthria, limb weakness, and atrophy. Patients usually die of respiratory failure and nutrition dysfunction several years after onset1. Weight loss and malnutrition are common in patients with ALS, which suggest a poor prognosis and increased mortality2,3. Various factors are considered responsible for this condition. It is well established that dysphagia in ALS is linked to reduced dietary intake and weight loss4, while recent studies also suggested that loss of appetite5 and taste changes6 may also be involved in dietary changes. Additionally, evidence from clinical research revealed the presence of hypermetabolism in ALS patients7. Although controversial, a large number of clinical studies found an association between hypermetabolism and a faster rate of progression and shorter survival8,9.

Given the difficulty in the development of an effective treatment, metabolism exploration and management of nutrition in ALS patients has generated great interest in recent years. Until now, the nutritional state of patients with ALS has mostly focused on weight, body mass index (BMI), waist-hip ratio (WHR) and some serum nutritional biomarkers10. A few studies of body composition in ALS have showed different pattern of changes in body composition among ALS patients, focusing on the fat-free mass (FFM) and fat mass (FM)3,5,11. Also, some researchers suggested that muscle wasting was associated with poor prognosis, while gaining fat might be associated with better survival in ALS patients3,12,13. Accordingly, targeting nutritional status and supplementation of nutrients might be a promising therapeutic approach.

In this study, we aimed to evaluate the nutritional status of Chinese ALS patients by exploring weight and body composition during the course of disease and to explore the correlation between nutritional parameters and clinical parameters. Furthermore, the effect of nutritional status on rate of disease progression was also studied to identify the potential prognostic role of nutritional status in ALS.

Results

A total of 100 newly diagnosed ALS patients (50 men and 50 women) and 50 normal controls were included in this study. The mean age at onset of patients was 51.90 ± 10.53 years, with a median (interquartile range) disease duration of 13 (8, 19) months at diagnosis. Seven patients had a positive family history and the bulbar form of onset was present in 35 (35%) ALS patients. The clinical characteristics of patients with ALS and the comparison of body composition between patients and controls are shown in Table 1. There were no significant differences in gender and age between patients and controls. The BMI was significantly lower in patients with ALS than in controls (p < 0.05), while no significant difference was found in weight, FFM, FM, and other nutritional parameters between the two groups.

Table 1 Clinical features of patients and the comparison of nutritional parameters between patients and controls.
Full size table

Weight and body composition at diagnosis

The mean weight of ALS patients at diagnosis was 65.25 ± 12.03 kg, which was significantly lower than the premorbid weight (68.70 ± 13.39 kg, p < 0.001). Comparing with weight before disease onset, the median weight variation rate at diagnosis was − 4.11 (− 9.32, 0.56) %, and the median BMI variation was − 0.97 (− 2.42, 0.13) kg/m2. Of the 100 ALS patients included in this study, 4% were underweight (BMI < 18.5 kg/m2), 39% were overweight (BMI: 24–27.9 kg/m2) and 7% were obese (BMI ≥ 28 kg/m2). Compared with their premorbid weight, 66% had weight loss at diagnosis; among these patients, 65.2% (43/66) had a weight loss exceeding 5%, and 34.8% (23/66) had a weight loss exceeding 10%. According to their weight loss at diagnosis, the patients were divided into two groups: patients with significant weight loss (≥ 5%, n = 43); patients without significant weight loss (no weight loss or weight loss < 5%, n = 57). A comparison of clinical features and nutritional parameters between patients grouped by weight variation is shown in Table 2.

Table 2 Demographic, anthropometric and clinical features in different groups stratified by weight changes.
Full size table

No significant differences in age, sex, family history, or disease duration were found among patients grouped by weight variation. There were 21(48.8%) bulbar-onset cases in patients with significant weight loss at diagnosis, which was significantly higher than those without significant weight loss (24.6%, p < 0.05). Among these bulbar-onset cases with significant weight loss at diagnosis, 76.2% (16/21) also had spinal involvement at baseline. With regard to body composition, we found that muscle mass (MM), bone mass (BM), WHR, and visceral fat index (VFI) were significantly lower in patients with bulbar onset (Fig. 1A). Interestingly, fat mass in the limbs was relatively higher in these patients than in patients with spinal onset, although the result did not reach significance (Fig. 1B). The predominant involvement of the lower motor neurons occurred in 72.1% of patients with significant weight loss and 56.1% without significant weight loss, while the difference was not statistically significant (p > 0.05). Moreover, no significant difference in body composition was noted between patients with predominant involvement of upper motor neurons and those with predominant involvement of lower motor neurons (p > 0.05).

Figure 1
figure 1

Weight and body composition in ALS patients grouped by site of onset. (A) Patients with bulbar onset had a significantly lower weight and MM than those with spinal onset. (B) FM in trunk was relatively lower in patients with bulbar onset while FM in limbs was relatively higher in these patients than in those with spinal onset. MM muscle mass, FM fat mass. *p < 0.05; **p < 0.01.

Full size image

At diagnosis, patients with significant weight loss had lower ALSFRS-R scores, lower respiratory scores of ALSFRS-R, and a higher rate of disease progression (p < 0.05). Seven patients had a positive family history of ALS, six of whom did genetic screening (4 were SOD1 mutation, 1 was FUS mutation, 1 was DCTN1 mutation; Supplementary Table 1). No significant difference in nutritional parameters, including weight and body composition, was found between patients with and without a family history, but the negative results were probably due to the small size of patients with a family history. According to the DPR at diagnosis, patients were divided into two groups: slow disease progression (DPR < 1) and fast disease progression (DPR ≥ 1). However, the nutritional parameters also did not differ between patients with slow disease progression and those with rapid disease progression.

Additionally, parameters of body composition varied in the groups with different weight variations (Table 2). The levels of BMI, FM, and FM in limbs and trunk were significantly lower in patients with weight loss exceeding 5% (p < 0.05). Patients with significant weight loss at diagnosis tend to have a relatively higher proportion of muscle mass and a relatively lower body fat percentage (p = 0.092, p = 0.089). Moreover, patients with weight loss ≥ 5% had a significantly lower WHR, VFI, and FM% than those without weight loss (p < 0.05).

The correlation analysis showed that the total ALSFRS-R score was positively correlated with the weight at diagnosis (r = 0.323, p = 0.001), muscle mass (r = 0.267, p = 0.008), fat mass (r = 0.207, p = 0.043), fat in the limbs (r = 0.21, p = 0.04), fat in the trunk (r = 0.204, p = 0.046), and the bone mass (r = 0.267, p = 0.008). Moreover, the correlation analysis revealed a significant correlation between the bulbar score of the ALSFRS-R and the parameters of body fat, including total fat mass (r = 0.237, p = 0.018), fat mass in the limbs (r = 0.238, p = 0.017), VFI (r = 0.237, p = 0.018), and WHR (r = 0.311, p = 0.002). However, no significant correlation was found between the respiratory score of ALSFRS-R and nutritional parameters including weight, BMI, and body composition.

Short-term change of nutritional parameters

There were 88 patients followed up at 6 months after first visit. During the follow-up period, endpoints occurred in five patients (4 patients died and 1 patient was tracheotomized), and PEG was used in 3 patients. The weight of 6 patients at follow-up was unknown due to their mobility problems and difficulty in standing on their own. Of 77 living patients without tracheotomy, 10.4% (8/77) were underweight at the follow-ups. Weight loss occurred in 52 (67.5%) patients at follow-up, of which 51.9% (27/52) had weight loss exceeding 5% and 26.9% (14/52) had weight loss exceeding 10% comparing with weight at diagnosis. Among patients with weight loss at diagnosis, 36 (54.5%) continued to lose weight as noted at subsequent follow-ups. Onset age, ALSFRS-R’ at follow-ups, and DPR’ at follow-ups were significantly different among patients with weight loss ≥ 5% and those without weight loss ≥ 5% during follow-ups (Supplementary Table 2).

Twenty-one patients reassessed their body composition during outpatient visits, and 14 (66.7%) patients had a weight loss at follow-up. The mean age of onset was 51.19 ± 11.88 years, and 7 (33.3%) patients were bulbar onset. The ALSFRS-R scores were significantly decreased during follow-ups compared with those in diagnosis (Table 3). Of all 21 patients with twice the evaluation of body composition, FFM, and MM values were significantly decreased during follow-ups, while no significant difference was found in FM, WHR, and VFI (Table 3). Among those patients with weight loss during follow-ups, we found that the FFM, MM and FM were significantly decreased during follow-ups (p < 0.01, Supplementary Table 3). Further analysis of changes in fat showed that fat in limbs was significantly decreased while no significant difference was found in WHR and VFI (p < 0.01, Supplementary Table 3). We also investigated the changes in nutritional parameters in patients with weight gain, but no significant difference was found in weight and body composition.

Table 3 Comparison of nutritional parameters at diagnosis and follow-up in patients (n = 21).
Full size table

Using the Pearson correlation analysis, we found a significant correlation between weight’ at follow-ups and total ALSFRS-R’ score at follow-ups (r = 0.343, p = 0.002). There was no correlation between total ALSFRS-R’ score at the follow-up and nutritional parameters at diagnosis (Supplementary Table 4). The mean rate of disease progression at follow-ups was 1.0, by which patients were divided into two groups: slow progression (DPR’ < 1) and rapid progression (DPR’ ≥ 1). Clinical features and nutritional parameters were compared between groups (Supplementary Table 5). No significant difference was found in onset age, site of onset, and disease duration between patients with different progression rates. The proportion of patients with weight loss during the follow-up period did not significantly differ between patients with slow or fast disease progression, while patients with weight loss ≥ 5% at follow-ups tend to progress rapidly (p < 0.001). Patients with rapid disease progression during follow-ups had lower VFI, a lower proportion of FM, and a higher proportion of MM at baseline (p < 0.05). In multifactor linear regression, we found that FFM at diagnosis and weight’ during follow-ups were independently correlated with DPR’ at follow-ups after adjusting for other contributing factors (p < 0.05, Table 4).

Table 4 Multivariate linear regression analysis of disease progression rate during follow-ups (n = 83).
Full size table

Discussion

Our study found that a large number of Chinese ALS patients (66%) had weight loss at diagnosis, and 43.0% had a weight loss greater than 5%. Weight loss occurred most commonly in patients with the bulbar form of onset and worse neurological function, correlating with faster disease progression. These results are consistent with a large-scale population-based study in Netherlands including 2420 ALS cases, among which 67.5% of patients reported weight loss at diagnosis with a mean loss of weight of 6.2 (9.7)%14. As shown in our study, weight loss is a long-term ongoing feature in the course of ALS. Previous studies also reported a loss of weight in ALS patients during different periods even before disease onset, which is associated with a poor prognosis14,15,16,17. Moglia et al. found that the median survival in patients with a mean monthly weight loss exceeding 1% at diagnosis was less than half that in those who gained weight4. Marin et al.3 also identified a nearly 30% and 34% increased risk of death with each 5% weight loss at diagnosis and follow-up, respectively.

The exact mechanism of weight loss remains largely unclear. Investigation of body composition may offer insights to address this issue. Previous studies with small sample sizes18,19 found that ALS patients have a lower BMI and FFM than healthy controls, but no significant difference in body composition was found among patients and controls in this study. The inconsistent results might be explained by the different disease duration of patients and the demographic differences of controls. Although muscle and fat were both lower in ALS patients than in controls in this study, the difference in parameters of body composition did not reach significance at the early stage of the disease. Ngo et al.5 also did not find significant differences in body composition between 62 patients with ALS and 45 healthy controls, while the study didn’t match age among patients and controls. Moreover, the difference in techniques for measuring body composition may also lead to different results. In our study, the difference in fat mass is relatively higher between patients and controls among parameters of body composition, which may reach statistical significance as the disease progresses or by expanding the sample size. In longitudinal observation, Nau et al.11 reported a loss of lean mass while gaining FM over 6 months, with weight loss and energy storage. Marin et al.3 also suggested a significant decrease in weight, BMI, and lean mass with increased FM and triceps skinfold thickness during follow-up. We also found that FFM and FM decreased in patients with weight loss as the disease progressed, which is consistent with previous studies, while the FM also decrease in our patients with weight loss and FM did not significantly increase in our patients with weight gain. The contradictory findings might be explained by the difference in energy expenditure and calorie intake. Nau et al.11 observed a loss of lean mass and increased fat mass, resulting in weight loss, while the energy store was increased in ALS patients. When comparing intake with consumption, no difference was found at the beginning, while calorie intake was higher than expenditure during follow-ups. Based on that study, we speculated that when patients consumed more calories than their expenditure, residual energy would be stored in the form of fat, maintaining body weight. In contrast, insufficient intake resulted in fat burning for energy, similar to the patients in our study. This hypothesis could also be indirectly supported by Ngo’s study5, which found that FM decreased significantly in ALS patients with loss of appetite during the 18-month follow-up but increased in patients with intact appetite.

Body fat might be the major factor involved in energy metabolism resulting in weight variation in ALS patients. A study by Barone et al.20 found that FM was significantly lower in underweight patients and it increased in patients with a higher BMI, while no significant difference was found in FFM. This study also observed lower BMI and FM in patients with significant weight loss. Impaired cellular energy homeostasis and mitochondrial dysfunction have been considered one of the most important mechanisms in ALS21. As one of the major nutrients, fat plays an important role in the energy supplementation of ALS patients. Evidence from mutant SOD1 mice suggests a preferential lipid-based energy metabolism in muscle fibers at a presymptomatic stage, which is independent of motor neuron degeneration22,23. The underlying mechanisms of metabolic changes might be related to altered protein function in metabolic pathways, including glycolysis and β-oxidation22, which could probably be restored by a high-fat diet24,25. Our study found patients with a lower proportion of FM tend to progress rapidly during follow-up. Lee et al.26 also suggested that loss of fat is correlated with faster disease progression in ALS patients (n = 20), indicating the positive effects of fat in ALS. Consistently, Park et al.27 found longer survival in ALS patients (n = 53) with an increased body fat rate. Future studies with larger sample sizes are needed to explore the effect of body fat on the prognosis of ALS patients.

Fat metabolism might be distinctive in different parts of the body in ALS patients. Lindauer et al.13 identified remarkably increased visceral fat and unremarkably decreased subcutaneous fat in ALS patients compared with controls, indicating the tendency of visceral adipose tissues depot in ALS patients. Our study found fat in limbs significantly decreased in patients with weight loss during follow-ups, while no significant difference was found in VFI or WHR, suggesting the predominant wasting of subcutaneous adipose tissues (SAT) in ALS patients. This specific pattern of lipid metabolism and storage may be explained by the abnormalities in the energy metabolism of skeletal muscle28 and probable neural denervation of SAT29 in the pathological process of ALS. The other alternative interpretation is the physiological differences between SAT and VAT30, including the stronger insulin resistance of VAT31,32, different sensitivity to lipolysis, and different expression of adipokines33. Moreover, survival analysis from Lindauer et al.13 indicated that increased subcutaneous fat rather than visceral fat was associated with a better prognosis. Other studies also indicated an association between increased triceps skinfold thickness and a better outcome3. Our study also found that patients with lower fat mass in limbs or trunk tend to have a relatively higher rate of progression during follow-ups; however, the result was only statistically significant in the association between VFI and disease progression rate at follow-ups. Moreover, the effect of VFI on disease progression rate did not reach statistical significance after adjusting for other covariates. It is widely accepted that visceral fat is a predictor of poor prognosis in a variety of diseases34,35, and there has been little studies on the effect of visceral fat and subcutaneous fat on the prognosis of ALS patients. This study observed a trend for increased rate of disease progression in patients with lower visceral fat, but this finding needs to be interpreted with caution due to the relatively small sample size, and more large-sample studies are needed to confirm and explore the findings in the future.

The current study has a number of limitations. First, patients with restricted mobility or difficulty standing on their own were excluded due to the limitations of measurements with the body composition analyser, probable leading to exclusion of patients with a progressed course of ALS. Second, caloric intake and energy expenditure might affect the changes in weight and body composition. Considering the possible effect of cognitive function on diet and nutrition36, the lack of cognitive assessment in this study would also affect the results. In order to correct the effect of diet on the rate of disease progression during the follow-up period, we included patients' weight during the follow-up period in the multivariate regression analysis. There is a wide variety of foods in China and the type of food eaten by Chinese people also varies from region to region. Until now, there is no unified scale for the dietary assessment of ALS patients in China, and we are further exploring and discussing this part. We hope to further explore the influence of diet on morbidity, weight maintenance, and prognosis of ALS patients in the future. Third, this study did not include data from the pulmonary function test (FVC%) to study the effect of respiratory function on nutritional status. Instead of FVC% in the pulmonary function test, we used the respiratory scores of ALSFRS-R to analyse the correlation between nutritional state and respiratory status37,38, which may affect our results. Moreover, we observed the change of nutritional state in a short period (6 months) and study the effect of nutrition on patients' functional progression. We intend to follow these patients for a long-term period to observe the impact of these nutritional parameters on survival in the future.

Conclusion

In conclusion, we found that weight loss was common in the course of ALS with the wasting of muscle and fat, correlating with rapid disease progression. Body composition has potential prognostic value for ALS, and fat may have a potential protective role in ALS patients. Future large-scale research could focus on fat metabolism to provide detailed individual dietary guidelines to delay the progression of the disease.

Methods

Participants

We performed a study of 100 consecutive patients presented at Peking Union Medical College Hospital from October 2020 to April 2021. All patients were newly diagnosed with clinically definite, probable, or laboratory-supported probable ALS according to the revised El Escorial criteria39. Exclusion criteria included acute infections, malignant tumours, untreated or uncontrolled endocrine diseases, other malnutrition disorders, and inability to accomplish the anthropometric measurements. Additionally, we included 50 healthy controls for comparison of body composition. This study was approved by the Ethics Committee of Clinical Research of Peking Union Medical College Hospital (Beijing, China) (approval number: JS-2624). All methods were performed in accordance with the Declaration of Helsinki and STROBE guidelines. All participants provided signed informed consent.

Clinical features and nutritional parameters at baseline and follow-ups

Demographic and clinical data were collected at the first meeting, including sex, age of onset, site of onset, disease duration (from onset to diagnosis), predominant involvement of the upper motor neurons or lower motor neurons40,41, and revised ALS functional rating scale (ALSFRS-R). Muscle strength was assessed by the UK Medical Research Council (MRC), and the total strength was calculated by summing the MRC scores of the muscle groups. The rate of disease progression (DPR) at diagnosis was calculated using the formula: DPR = (48-ALSFRS-R)/disease duration (months). We defined rapid disease progression as DPR ≥ 1 according to previous studies42,43.

Anthropometric characteristics were measured at the time of recruitment, including height, weight, WHR, and detailed body composition. Height was accessed when patients stood in an upright position. Weight before disease onset was reported from patients or family members. Weight, WHR, and body composition at diagnosis were evaluated by a body composition analyser (TongFang Health Technology, Beijing, China) using the direct segmental multifrequency bioelectrical impedance analysis method (DSM-BIA). The models for body composition were developed based on the results of body composition measured by isotope dilution, magnetic resonance imaging (MRI), and dual-energy X-ray absorptiometry (DEXA)44:

$$Y = 1.072Xe^{{ - \left( {\frac{{k - 1.786}}{{1.77}}} \right)^{2} }} + 0.3194K^{2} - 2.073K + 3.713$$
(1)

X represents the impedance index; when K = 1, 2, and 3, Y represents total body water (TBW), fat mass (FM), and bone mass, respectively.

$${\text{Protien mass}} = 1.01 \times \left( {{\text{BW}} - {\text{FM}} - {\text{BM}} - {\text{TBW}}} \right) + 1.29$$
(2)
$${\text{FFM}} = {\text{BW}} + {\text{TBW}} + {\text{Protienmass}}$$
$${\text{Muscle Mass}} = {\text{TBW}} + {\text{Protienmass}}$$

The parameters of body composition included FFM, MM, FM, visceral fat index (VFI), and percentage of body mass (FM%, MM%). BMI was calculated using the formula: BMI = weight/height2 (kg/m2).

All patients were followed 6 months after the first visit through a face-to-face evaluation or telephone contact. By telephone call, we collected patients’ weights at follow-ups (Weight’) by self-report from patients who followed our directions with the same scale. Moreover, ALSFRS-R scores were re-evaluated at follow-ups by outpatient visits or telephone follow-ups. The DPR during follow-ups was calculated using the formula: DPR’ = (ALSFRS-R at diagnosis—ALSFRS-R during follow-ups)/6 (months). During clinic follow-ups, weight and body composition of patients were reassessed using the body composition analyzer. Change of each nutritional parameter at different time points was observed and compared. The percentage of weight loss was calculated by comparing the weight from different points in time (before disease onset, at baseline, and during follow-ups). A percentage exceeding 5% was defined as significant weight loss3,14,45.

Statistical analysis

The Statistical Package for the Social Sciences software (SPSS, Version 22) was used for the data analysis. Categorical variables were expressed as frequencies (proportions) and analysed by the chi-square test or Fisher’s exact test. Normally distributed continuous variables such as age, weight, and BMI were expressed as the mean ± SD and compared by independent Student’s t-tests. Non-normally distributed continuous variables, including most parameters of body composition, are presented as the median (interquartile range, IQR) and were analysed using the Mann–Whitney test between groups. The associations between clinical parameters and nutritional parameters were analysed using Spearman correlations. Multivariate logistic regression was performed to identify factors associated with weight loss at diagnosis. The short-term outcome was evaluated by ALSFRS-R at follow-up. A multiple linear regression model was constructed to investigate the correlation of clinical features and nutritional state at diagnosis with ALSFRS-R score at follow-up. The differences in weight and body composition between baseline and follow-ups were analyzed by paired t-test. The results with a p value < 0.05 were considered to be significant.