Table 3

Relative risks of death associated with nutritional variables measured over the entire follow-up in univariate and multivariate analysis

During follow-up nutritional variables*Univariate analysisMultivariate analysis
Crude hazard ratio95% CIp ValueAdjusted hazard ratio95% CIp Value
Weight variation from usual weight (for 5% decrease)1.401.23 to 1.59<0.00011.341.18 to 1.51<0.0001
BMI variation from usual BMI (for one unit decrease)1.301.18 to 1.42<0.00011.241.13 to 1.36<0.0001
Rate of weight loss from usual weight (%/month)1.111.02 to to 1.170.11
Rate of BMI loss from usual BMI (unit/month)1.441.08 to 1.930.011.280.94 to 1.740.11
Malnutrition versus other BMI categories2.321.25 to 4.290.0072.561.33 to 4.940.005
BMI categories0.030.01
 Malnutrition2.151.14 to to 4.25
 Normality (reference)1.001.00
 Overweight0.980.56 to 1.720.710.40 to 1.28
 Obesity0.430.13 to 1.400.360.11 to 1.19
Mid-arm muscle circumference, cm (for one unit increase)0.890.82 to 0.950.0010.930.86 to 1.0050.07
Triceps skinfold thickness (cm, for one unit increase)0.990.96 to 1.030.660.940.90 to 0.980.003
Phase angle (°, for one unit decrease)1.681.27 to 2.230.00031.270.92 to 1.750.15
Extracellular fluid/intracellular fluid volume (for 0.2 unit increase)1.761.26 to 2.460.00091.671.07 to 2.610.02
Lean mass (kg, or 2.5 unit increase)0.940.89 to 1.0020.061.000.94 to 1.070.97
Fat mass (kg, for 2.5 unit increase)0.980.91 to 1.040.460.900.83 to 0.960.003
  • Significant p values appear in bold.

  • BMI categorisation: (i) malnutrition: BMI <18.5 if age <70 years and BMI<21 if age ≥70 years; (ii) normal status 18.5≤BMI<25 if age <70 years and 21≤BMI<27 if age ≥70 years; (iii) overweight: 25≤BMI<30 if age <70 years and 27≤BMI<30 if age ≥70 years; (iv) obesity BMI≥30.

  • * Survival analyses were performed separately for each nutritional variable.

  • Adjustment on age, gender, bulbar form at onset, amyotrophic lateral sclerosis Functional Rating Scale, manual muscular testing, forced vital capacity, Airlie House criteria during follow-up and diagnostic delay.