Nutritional status is a prognostic factor for survival in ALS patients

Methods.

A total of 55 patients with definite or probable ALS according to the criteria of El Escorial7 seen in our referral neurology practice were enrolled in the study between March 15, 1996, and October 16, 1997. No patients were excluded or refused. The patients underwent an interventional assessment that consisted of coupled, detailed evaluations in the neurology and nutritional clinics. The protocol was approved by the regional ethics committee. During the neurologic evaluation, the date of onset of the first disease symptoms (fasciculations; cramps; weakness; speaking, chewing, and swallowing disorders; etc.) was elucidated by neurologic history. The variable of interest was time since entry into the study.8,9 Patients were followed until death or until the closing date of the study (December 1, 1997). No patient was lost to follow-up. Neurologic deficit was quantified by manual muscle testing of the neck and upper and lower extremities as defined by the Medical Research Council10 (maximal value, 150 points), the modified Norris scales for limbs (maximal value, 63 points) and bulbar function11 (maximal value, 39 points), and the ALS Functional Rating Scale12 (ALS FRS; maximal value, 40 points). Slow respiratory vital capacity (VC) was measured using a Hans Rudolph pneumotachograph, integrated in a body plethysmography system (CPF Medical Graphics, St. Paul, MN). Results were expressed in relation to a theoretical calculated index value. All patients were treated with a benzothiazole (riluzole) with the goal of reducing neuronal excitotoxin-related disease progression,13 and the starting date of treatment was noted.

For the nutritional assessment, weight was recorded to the 0.1 kg with a scale (Seca, Vogel & Halke, Hamburg, Germany), and height was measured to 0.5 cm using a SECA height gauge. BMI was calculated according to the formula BMI = weight (kg)/height² (m). BMI ≤ 18.5 kg/m² defined malnutrition.14 A possible subsequent placement of gastrostomy for enteral feeding was noted.

Statistical analyses were conducted using Statview 4.5 F software (Abacus Concept, Inc., Berkeley, CA) and Epitable 2.0 (Centers for Disease Control and Prevention, Atlanta, GA, and Epicentre, Paris, France). Qualitative results were compared using Pearson’s chi-square test or Fisher’s exact test. Quantitative averages were compared using the Mann–Whitney U test. Survival was analyzed using the Kaplan–Meier method,9 and comparisons between malnourished and nonmalnourished patient groups were performed using a logrank test. Multivariate analysis was performed using a Cox regression model.15 The threshold of significance was set at 0.05 for all statistical analyses.

Results.

A total of 55 different patients were seen during the study period. Their principal characteristics, obtained during the interventional (first) study visit, are presented in table 1. Time between initial signs and interventional assessment was 29 ± 25 months (average ± SD). The cumulative 75% survival (25% mortality) was reached by the 26th month from initial signs. The mean follow-up period was 7 ± 4 months. A total of 32% of patients died during the study period (bulbar, 41.7%; spinal, 25.8%).

Neurologic and respiratory evolution results were as follows: survival was worse for bulbar ALS patients (mean bulbar, 30 months; spinal, 37 months; p = 0.05), and this form had age at onset older than the spinal form (63 years versus 58 years; p = 0.05). Higher age at onset showed a trend for reduced survival (age > 60 years, 31 months; age < 60 years, 37 months; p = 0.06). Survival was greater for patients with VC > 60% (8.5 months) than those with VC reduced below this level (3.8 months; p < 0.0001; figure 1). Survival duration was independent of the intensity of motor deficit for all the scoring scales used.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Kaplan–Meier survival plot of patients according to slow respiratory vital capacity (VC). The difference between the patients with VC ≥ 60% (thick line) and the patients with VC < 60% (thin line) is highly significant (p < 0.0001).

Malnutrition was diagnosed in 16.4% of patients. The main characteristics of malnourished or nonmalnourished patients are compared in table 1.

Survival duration was shorter in malnourished than nonmalnourished patients (p < 0.0001), and nutritional status was a fundamental predictive factor for survival (figure 2). Two alternative methods of nutritional assessment were used, and they agreed with the BMI results (unpublished observations). Univariate relative risk of death was increased 7.7-fold (95% CI, 3.0 to 20.2) for patients with malnutrition.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Kaplan–Meier survival plot of patients in the nonmalnourished (NM; thick line) and malnourished (M; thin line) groups according to nutritional status as determined by body mass index. The difference between malnourished and nonmalnourished patients is highly significant (p < 0.0001).

On average, the VC percentage was lower in the malnourished group. The difference between the two groups, however, did not reach statistical significance (see table 1), suggesting that nutritional status and VC were not correlated. Among the malnourished patients, there were no more bulbar than spinal ALS patients (bulbar, 8 of 9; spinal, 8 of 9).

Lastly, a multivariate analysis determined only two significant, independent prognostic factors after adjusting for ALS type, disease duration prior to consultation, duration of riluzole treatment, age at onset, and presence of a gastrostomy. These factors were VC and malnutrition as assessed by BMI. In this model, patients with VC < 60% multiplied the risk of death by 22.0 (95% CI, 5.0 to 97.1; p < 0.0001), and being malnourished multiplied the risk of death by 7.4 (95% CI, 1.7 to 32.1; p < 0.01).

At the time of the interventional consultation, 3 patients already had a feeding gastrostomy in place, and 16 patients had a feeding gastrostomy placed during the study period, with placement determined by a loss of 5 to 10% of former body weight. Gastrostomy placement occurred an average of 71 ± 117 days after the interventional examination date. None of these results was modified if one excludes the patients who had enteral nutritional support during the study period.

Discussion.

Comparison between patient populations in the current study with other published reports is problematic, because the only five prior studies analyzing the nutritional status of patients with ALS have small study populations with heterogeneous patient characteristics2,6,16-18 (table 2). The mean BMI of patients in our study is normal (23 ± 5 kg/m²), similar to the values reported elsewhere (18 to 30 kg/m²), although this index measure is only given in two of five studies.2,16 The percent of bulbar and spinal ALS types are provided in only one study,16 in which bulbar type initial presentation is only 14% of studied patients, and there is no explanation for this low number. The comparable percent we report—43.6%—is slightly higher or similar to the ratio with bulbar onset reported elsewhere in the published literature.4,19-21 The time between initial signs and interventional assessment in our patients (29 ± 25 months) is similar to that reported by different authors.2,16,17 Just one study16 addresses the percent of study patients having a gastrostomy (45%), which is slightly higher than in our patient population (34%). Additionally, our patients have similar age characteristics compared with those reported in two earlier studies conducted in our hospital, and compared with those reported classically for ALS populations.4,21-23

Using our criteria of malnutrition with BMI, 16.4% of our patients are malnourished. BMI is often taken into consideration2,16 with different threshold levels. We have chosen those proposed by Torun and Chew14 according to James et al.24 Concerning the occurrence of malnutrition in ALS patients, the compared criteria are quite disparate in the literature. It has already been reported that alimentary intake of patients is balanced with energy expenditure at the onset of their medical care, and subsequently that intake drops below their requirements if no nutritional intervention is made.17 By contrast, Mazzini et al.16 report that at the onset of medical attendance, 53% of their patients had a BMI < 20 kg/m², and 55% had a weight loss > 15% of their usual weight, indicating a high occurrence of malnutrition. A study published in 1983 found weight loss of 10% or greater in 24% of patients, insufficient alimentation (or food intake) in 70% of patients, and, based on anthropometric criteria, a high occurrence of malnutrition; however, these criteria have since been rejected for having too low a threshold.6 According to Kasarskis et al.,2,25 inadequate oral intake relative to energy consumption is observed in acccordance with disease progression, and in parallel with a drop in BMI. Therefore, although malnutrition is reported as being frequent in studied ALS populations, agreement regarding its occurrence is lacking. Yet it seems reasonable to accept an occurrence that ranges from 15 to 55% if one combines patients at all stages of the disease.

For the population studied, the existence of a state of malnutrition is a clearly negative prognostic factor for survival. It has already been suggested in the study by Mazzini et al.,16 in which an improvement in patients’ nutritional status via enteral feeding by gastrostomy, quantified by an improvement in BMI, yielded improved survival by 6 months follow-up. They propose attentive surveillance of patients’ nutritional status, particularly if bulbar signs appear. They advocate prompt consideration of gastrostomy placement for enteral feeding according to the recommendations of Leigh and Ray–Chaudhuri,20 and set the tolerable limit for total weight loss at 10%. Kasarskis et al. (personal communication cited in Kasarskis et al.2) cite the strong correlation between reduced BMI and shortened survival, and insist on the importance of providing nutritional supplementation to all patients with ALS as their alimentary intake falls below their requirements well before the terminal phase of the disease.2,25 They recommend enteral feeding when the weight loss relative to usual weight reaches 5% or when clear bulbar dysfunction develops.25 Nevertheless, neither of these studies document how the threshold of 5% or 10% weight loss is chosen.

Factors influencing survival significantly in our study are compared with literature data: ALS with bulbar type onset is a negative prognostic factor,1,21 as is reduction in VC.26 Higher age at disease onset is at threshold significance.

In our study, the three methods for evaluating neurologic deficit have no prognostic value. This may be due primarily to the short follow-up period. Moreover, muscle strength testing is a semiquantitative, insensitive method, and yields scores with nonuniform distribution.27 The Norris scales provide a good global disease assessment, but rely on subjective criteria,27 and although the ALSFRS addresses all functions that may be affected by ALS, our study does not confirm the hypothesized prognostic value of this method,28 perhaps because of its subjective nature.

Malnutrition causes rapid muscle wasting,29 yet the magnitude of neurologic deficits in our study are independent of nutritional status. This may be related to the fact that assessment of disease severity is determined globally, and thus takes into account indicators with equal or greater importance than nutritional state, but with a faster progression.

Although bulbar onset of ALS can be regarded as a much greater risk of malnutrition than spinal-onset disease, in our study occurrence of malnutrition does not differ between the two disease types. If inadequate caloric intake and weight loss correlate with severity of dysphagia in most ALS patients,1,3,4,25 it is possible that a higher energy expenditure due to spasticity and fasciculations occurs in some spinal-onset forms of ALS.