ALS patients from the MND Clinic at the Neurology Department of Hanyang University Hospital in Seoul, Korea were prospectively enrolled from May 2009 to November 2012. All subjects were selected according to the revised El Escorial criteria, and fulfilled the criteria for possible, probable, probable-laboratory-supported or definite ALS.14 None of the patients had a family history of ALS or mutations in the gene encoding superoxide dismutase 1 (SOD1), so their diagnosis was seemingly sporadic ALS without SOD1 mutations. The ALS Functional Rating Scale-Revised (ALSFRS-R) was used to assess the patients' functional status.15 Patients in the early stages of ALS were recruited, as defined by a disease duration of ≤3 years. This study was approved by the Institutional Review Board of Hanyang University Hospital. The controls were recruited from outpatients who visited the Health Promotion Center of Hanyang University Hospital during the same period. The availability of more than 3,000 healthy individuals enabled the identification of well-matched controls for each patient in terms of sex, age, and body mass index (BMI). Thus, 136 healthy matched controls were included in the study.

Subjects who had been diagnosed with any condition associated with changes in serum UA concentrations were excluded, such as 1) stroke, angina, or myocardial infarction; 2) renal dysfunction; 3) history of alcohol abuse; 4) acute inflammatory state or medical disease (e.g., pneumonia or gastroenteritis); 5) administration of UA-lowering or UA-increasing medications (thiazide or allopurinol); 6) history of gout; or 7) history of percutaneous gastrostomy.

Relevant demographic and clinical data were collected at enrollment, including age, sex, age at symptom onset and ALS examination, region of symptom onset, disease duration (interval between symptom onset and examination), forced vital capacity (FVC), medical history, alcohol consumption, and medications and treatment provided.

The progression rate of the disease (ΔFS) from symptom onset to the time of examination was calculated as follows: ΔFS=(48-ALSFRS-R score at time of examination)/[duration between symptom onset and time of examination (months)].16

Baseline serological data were obtained from serum samples obtained at the first visit to our clinic for the referral diagnosis of ALS. Patients who fulfilled the inclusion criteria after at least 6 months had elapsed since they entered the study had a second blood examination to obtain a longitudinal follow-up of serum UA level. Nonfasting blood was collected and centrifuged, and UA was assayed using the kinetic method (SICDIA L UA reagent, Shin Yang Pharm., Seoul, Korea). Within 30 minutes of collection, samples of blood were centrifuged at 3,000 rpm for 10 minutes. The serum levels of blood urea nitrogen, creatinine, and UA were measured using standard methods with the aid of an automatic chemistry analyzer (Hitachi 7600-210, Hitachi, Tokyo, Japan).

Continuous variables (e.g., age, BMI, and serum UA level) are presented as mean±SD values. Comparisons between ALS patients and control subjects regarding demographic and laboratory characteristics were performed using Student's t-test and chi-square tests. A paired t-test was used to compare serum UA levels at the times of the first and second examinations. The correlations between serum UA levels and the variables were calculated using Pearson's correlation. Multiple regression analyses were used to examine the association between serum levels of UA and the other variables (age, sex, BMI, and ALSFRS-R score). The cutoff for statistical significance was set at p<0.05 for all of the data analyses. Statistical analyses were carried out using SPSS (version 18; SPSS Inc., Chicago, IL, USA).

The survival analysis was conducted after stratifying UA levels according to sex-specific tertiles; survival was analyzed using the Kaplan-Meier method with the log-rank test. Survival was defined as the duration from the time of examination to death or tracheostomy.

During the study period, 206 patients visited the MND Clinic of the Neurology Department at Hanyang University Hospital. Of these, 19 were excluded due to refusal to participate in the study, and 16 were excluded due to a disease duration of longer than 36 months. Patients were further excluded due to a family history of motor neuron disease (n=11), presence of a concurrent respiratory tract infection (n=3), history of gout (n=3), cancer (n=10), cardiovascular disease (myocardial ischemia, angina, or stroke; n=6), a renal disease (n=4), and gastrostomy at the time of the first examination (n=1). After these patients were excluded, 136 patients were finally enrolled in the study group. Males constituted 69 (50.7%) of the study participants.

The ALS patients were 55.0±11.4 years old at the time of examination, had a disease duration of 13.6±7.4 months and a BMI of 22.6±3.4 kg/m². The region of symptom onset was bulbar in 37 patients, limb in 95, and axial muscles in the remaining 4. Thirty-seven patients (27.2%) were being treated with riluzole at the time of the first examination (Table 1). There were no significant differences in age at symptom onset, age at examination, BMI, ALSFRS-R score, disease duration, or FVC between male and female patients.

The age (54.7±10.9 years) and BMI (22.9±2.8 kg/m²) of the matched healthy controls did not differ significantly from those of the ALS patients (p=0.848 and 0.331, respectively) (Table 1).

The serum UA level was significantly lower in the ALS patients (4.50±1.17 mg/dL) than in the healthy controls (5.51±1.22 mg/dL, p<0.001) for both males and females (Table 1). UA levels were lower among female patients and controls at the time of the first examination. The serum UA level was positively correlated with BMI, ALSFRS-R score, and creatinine, and was negatively correlated with ΔFS (r=0.280, p=0.001; r=0.218, p=0.011; r=0.381, p<0.001; and r=-0.317, p<0.001, respectively) (Fig. 1). Age at examination, age at symptom onset, disease duration, and FVC were not correlated with the serum UA level (r=-0.060, p=0.485; r=-0.066, p=0.448; r=0.090, p=0.295; and r=0.029, p=0.744, respectively).

Simple and multiple regression analyses were used to control for confounders. Possible correlations between the serum UA level and the demographic and clinical parameters of the ALS patients were evaluated, such as age at examination, sex, region of onset, administration of riluzole, BMI, ALSFRS-R score, disease duration, ΔFS, FVC, and creatinine. Multiple linear backward regression revealed that sex, BMI, bulbar onset, and ΔFS contributed significantly to serum UA levels (R²=0.331, p<0.001 for the model) (Table 2). This indicates that lower serum UA level is associated with faster disease progression (as expressed by the decrease in ALSFRS-R score per month).

Serum UA levels were higher among males than females, and the tertile ranges of serum UA levels differed with sex. The tertile ranges were classified into bottom, middle, and top tertiles separately for males and females, and the survival rate was analyzed in the ALS patients using the sex-specific tertiles (for males: <4.7, 4.7-5.4, and >5.4 mg/dL; for females: <3.7, 3.7-4.4, and >4.4 mg/dL).

According to our survival analysis of males and females combined in a single model, there was a dose-dependent survival advantage for top-tertile serum UA levels compared to bottom-tertile serum UA levels (log-rank test: p=0.035) (Fig. 2). The mean (SE) survival times in the bottom, middle, and top tertiles were 17.9 (1.6), 21.2 (1.8), and 31.0 (3.4) months, respectively. After separation into male and female groups, the mean survival time did not significantly differ between the top and bottom tertiles among the males [31.4 (5.2) vs. 17.0 (2.3) months, p=0.057] or the females [26.7 (3.2) vs. 17.2 (1.9) months, p=0.218].

The temporal change in serum UA levels and the relationship between disease severity and serum UA level were investigated through follow-up examinations in 58 patients. Similar serum UA results were found in the 58 patients who had a second blood test at least 6 months after the first one (11.1±5.5 months, range 6.0-31.1 months): the serum UA level of the 58 patients who were followed up was 4.55±1.02 mg/dL (Table 3). The ΔFS from symptom onset to the time of the first or second examination (ΔFS_1st or ΔFS_2nd, respectively) was calculated as follows: ΔFS=(48-ALSFRS-R score at time of examination)/[duration between symptom onset and time of first or second examination (months)]. The sequential changes in the correlations between the serum UA level and the other variables, including age at examination, sex, bulbar onset, ALSFRS-R score, BMI, disease duration and the progression rate of the disease, were analyzed using multiple linear regression analysis. Multiple linear backward regression analysis for the time of the first examination revealed that sex and ΔFS_1st were correlated with the serum UA level (R²=0.431, p<0.001 for the model), whereas age at the time of the first examination, bulbar onset, ALSFRS-R score, disease duration, and BMI were not significantly correlated. At the time of the second examination, only sex was correlated with the serum UA level (R²=0.381, p<0.001 for the model), while age at the time of the second examination, bulbar onset, ALSFRS-R score, disease duration, BMI, and ΔFS_2nd were not significantly correlated.