Association between serum uric acid and α-klotho protein levels in the middle-aged population

Characteristics of the subjects

The characteristics of the study population based on the klotho level are summarized in Table 1. Table 1 shows the number of subjects (N), geometric mean klotho concentration (pg/mL), 95% CI, and p-value according to each study characteristic. The whole study population had a geometric klotho concentration mean of 806.5 pg/mL (95% CI: 801.7–811.4). The klotho concentration differed significantly by age, gender, ethnicity, education, monthly income, smoking status, alcohol consumption, obesity, kidney function, and history of diseases (hypertension, gout, kidney disease, and malignancy). The geometric mean klotho levels were more likely to be low in subjects who were older (lower than those in other age groups, p-value < 0.0001), male (male vs. female, 788.4 vs. 824.1; p-value < 0.0001), white (lower than those in other races, p-value < 0.0001), less educated (≤12 years vs. >12 years, 796.3 vs. 816.5; p-value < 0.0001), paid less (≤$20,000 vs. >$20,000, 796.2 vs. 809.8; p-value = 0.0218), current smokers (lower than that in former smokers or never smokers, p-value < 0.0001), drinkers (at least 12 alcohol drinks per year vs. less than 12 drinks, 792.3 vs. 842.6; p-value < 0.0001), obese (non-obese vs. obese, 827.0 vs. 800.4; p-value < 0.0001) and who had a history of hypertension (yes vs. no, 796.3 vs. 818.3; p-value < 0.0001), gout (yes vs. no, 743.3 vs. 810.7; p-value < 0.0001), kidney disease (yes vs. no, 734.3 vs. 809.1; p-value < 0.0001) or malignancy (yes vs. no, 780.8 vs. 810.1; p-value < 0.0001). However, the data did not show any significant differences among the subjects in terms of physical activity (p-value = 0.3446) or a history of diabetes (p-value = 0.4966).

Correlation analysis between uric acid and klotho levels

The scatter plot of serum uric acid levels and log klotho concentrations is shown on the left part of Figure 1. The serum uric acid level and log of klotho protein level was treated as continuous variables. The uric acid levels were negatively correlated with the log klotho concentrations, and the correlation coefficient (r) was −0.154 (p-value < 0.0001). A blue regression line with a 95% CI is shown in the graph. On the right part of Figure 1 is the scatter plot of eGFR and log of klotho concentration. eGFR was positively correlated with the log klotho concentrations, whose r was 0.149 (p-value < 0.0001). Herein, eGFR is treated as a continuous variable.

Figure 1. Scatter plot of log of klotho concentration with serum uric acid levels (Left) and estimated glomerular filtration rate (Right).

Analysis of serum klotho levels according to medical history

Figure 2 shows a comparison of the mean (SE) log klotho concentration, along with significant differences according to disease states. When a total of 11,734 subjects were divided into two groups according to their uric acid level (normal vs. hyperuricemia), the mean of log klotho concentration in the normal group was 6.711 pg/mL (0.003) and the mean log of the hyperuricemia group was 6.629 pg/mL (0.006), which was lower than that of the normal group (p-value < 0.0001). When the same participants were divided into two groups according to the presence of gout history (normal vs. gout patients), 11,031 participants without a history of gout had a mean log klotho (SE) of 6.698 pg/mL (0.003), and the other 703 participants with a history of gout had a mean of 6.611 pg/mL (0.012), which was lower than that of the normal group (p-value < 0.0001). Of the 716 participants with a history of gout, the subgroup with a normal uric acid level had a log mean klotho (SE) of 6.637 pg/mL (0.018), while the other subgroup with hyperuricemia had a log mean klotho (SE) of 6.579 pg/mL (0.016), which was lower than that of the normal uric acid group (p-value = 0.0169).

Figure 2. Mean logarithm of klotho levels by the presence of hyperuricemia and history of gout (pg/mL).

Regression analysis of serum klotho levels according to several health conditions

Table 2 shows the results of regression analysis of klotho concentration according to various health conditions. The estimated beta coefficients, SE, and significance differences (p-value) were calculated using regression models depending on the variables mentioned above. In an unadjusted linear regression model without any variable adjustment, there was a significant negative association between each one-unit increase in uric acid levels and change in log klotho level (beta = −0.034, SE = 0.003; p-value < 0.0001). In the logistic regression model, we also found out that the group with hyperuricemia had a significantly lower log klotho level (beta = −0.085, SE = 0.009; p-value < 0.0001) compared to the reference group without hyperuricemia. The gout group also demonstrated a significantly a lower log klotho level (beta = −0.074, SE = 0.020; p-value = 0.0006) compared to the reference group who never had gout. Within the group with a history of gout, the subgroup with hyperuricemia at the moment of investigation showed a lower log klotho level (beta = −0.072, SE = 0.026; p-value = 0.0059) compared to the reference subgroup who did not have hyperuricemia. Additionally, we analyzed the sample using four adjusted models. All adjustment variables used in these regression models, including kidney function, were treated as categorical variables. Adjusted Model 1 was initially adjusted for age, gender, and ethnicity. Model 2 was further adjusted for related disease history (history of hypertension, diabetes, kidney disease, and malignancy). Model 3 was adjusted for demographic factors, and Model 4 for kidney function (eGFR). For the fully adjusted linear regression model, the model was finally adjusted for age, gender, ethnicity, history of hypertension, diabetes, kidney disease, malignancy, education, family income, exercise, alcohol, smoking, obesity, and kidney function. In the fully adjusted Model 4, there were significant negative associations between each one-unit increase in uric acid levels and change in log klotho (beta = −0.028, SE = 0.004; p-value < 0.0001). There were still significant negative relationships between hyperuricemia state and log klotho level (beta = −0.062, SE = 0.009; p-value < 0.0001) and between having a history of gout and log klotho level (beta = −0.039, SE = 0.019; p-value = 0.0443) compared to those in each reference group. However, within the group with a history of gout, there was no significant difference between the hyperuricemia subgroup and the normal uric acid subgroup when fully adjusted for all covariates (beta = −0.043, SE = 0.030; p-value = 0.1462).

Study population

The NHANES is conducted by the National Center for Health Statistics (NCHS) to assess the health and nutritional status of adults and children in the US. The survey included a representative sample of the civilian non-institutionalized US population and conducted interviews on socio-demographics, diet intake, health behaviors, medical history, and physiological and laboratory examinations [38]. The study protocols and NHANES testing procedures were reviewed and approved by the NCHS Institutional Review Board. Both oral and written consents were obtained from all the participants.

For the current study, we initially selected 13,758 participants aged 40–80 years old from the NHANES 2007–2016, wherein serum klotho data were available. Among them, 1,971 participants with missing data on demographics (n = 1,658) or medical history (n = 313) were excluded. We further excluded 53 participants whose eGFR was <15 mL/min/1.73 m² because this eGFR was interpreted as kidney failure. The final study population comprised 11,734 adults.

Uric acid

Blood samples were collected and frozen until they were ready for transfer to a microtube. The entire procedure can be found in the procedure manual on the NHANES website [39].

Briefly, uric acid is oxidized by uricase to produce allantoin and hydrogen peroxide. Hydrogen peroxide reacts with 4-aminoantipyrine and 3,5-dichloro-2-hydroxybenzene sulfonate in a reaction catalyzed by peroxidase to produce a colored product. The system monitors the change in absorbance at 520 nm over a fixed time interval. The change in absorbance is directly proportional to the concentration of uric acid in the sample, which is measured using the Beckman Unicel DxC 800 Synchron Clinical System. The lower limit of detection of uric acid is 0.5 mg/dL.

Hyperuricemia is a condition characterized by excess uric acid, and there is no universally accepted definition. We defined hyperuricemia as serum uric acid concentrations of ≥7.0 mg/dL for males and ≥6.0 mg/dL for females [40, 41].

Klotho

The blood sample for klotho measurement was collected during the NHANES 2007–2016 and was agreed upon by the study participants at that time that it could be used for future research. Until samples were predefined as a batch, all samples were stored at −80°C and then provided to the technicians. Soluble α-klotho levels were analyzed during 2019–2020, with a commercially available enzyme-linked immunosorbent assay (ELISA) kit manufactured by IBL International, Japan.

To validate the IBL ELISA method in human samples, sample analysis was performed in duplicate according to the manufacturer’s protocol, and the average of the values was calculated as the final value. Two quality control samples with low and high concentrations of klotho were analyzed in duplicate in each ELISA plate. The assay sensitivity was 4.33 pg/mL. The assay measurement indicated outstanding linearity in the plots, with the expected with the expected (r² = 0.998) and obtained values (r² = 0.997).

Other variables of interest

We considered demographic, health behavior, and medical history as the other variables of interest. Demographic variables included age (40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, and 75–79 years old), gender (male or female), ethnicity/race (white, black, Hispanic, or others), monthly family income (<$20,000 or ≥$20,000), and education level (below high school graduate, or over college degree). Health behavior-related variables were moderate physical activity (yes or no), self-reported smoking status (current smoker, former smoker, or never smoked), and self-reported current alcohol consumption (drinker or non-drinker). Obesity was divided by whether body mass index (BMI) was higher than 25 kg/m² (yes or no). Kidney function was classified into 4 stages according to the calculated eGFR (15–29, 30–59, 60–89, and ≥90 mL/min/1.73 m²), which means severe loss of kidney function, moderate loss of kidney function, mild loss of kidney function, and normal respectively. Medical histories were based on the disease diagnosis by a doctor or health professional and asked whether they had been diagnosed with hypertension, diabetes, gout, kidney disease, or malignancy. The histories of the diseases were included as confounding variables because those were suggested as being linked to uric acid and klotho level. All variables except serum uric acid and klotho level were treated as categorical when used as adjustment variables in regression analysis.

Statistical analysis

Weighted estimates of the population parameters were applied based on the NHANES Analytic and Reporting Guidelines. All statistical analyses were performed using the PROC procedures in SAS 9.4 (SAS Institute, Cary, NC, USA) to account for the complex sampling scheme. All the tests were 2-sided, and the level of statistical significance was set to α = 0.05.

Klotho concentration was skewed to the right and log-transformed to ensure a normal distribution. The geometric mean klotho level and 95% confidence interval (95% CI) were computed for each characteristic of the study participants. Statistically significant differences were shown using p-values based on t-test and analysis of variance. Two scatter plots of uric acid, namely, eGFR and log-transformed klotho concentrations, were displayed, together with the correlation coefficient (r) and p-value. Herein, eGFR and klotho concentrations were treated as continuous variables. To explore the association between klotho level and uric acid, participants were classified as follows: 1) those with and without hyperuricemia, 2) those with and without a history of gout, and 3) those with hyperuricemia and without hyperuricemia in subjects who had a history of gout.

We performed unadjusted and adjusted linear regression analyses to estimate changes in klotho levels with a one-unit increase in serum uric acid levels. We further conducted unadjusted and adjusted logistic regression analysis to determine changes in klotho levels in terms of a binary outcome variable (yes or no). Categorical binary outcomes were based on the presence of hyperuricemia, history of gout, and hyperuricemia in the group with a history of gout. The regression model provided an estimated regression coefficient (beta) and standard error (SE). Adjusted Model 1 was basically adjusted for age, gender, and ethnicity. The other models were Model 2, further adjusted for disease history; Model 3, adjusted for demographic factors; and Model 4, adjusted for kidney function. The fully adjusted model was adjusted for age, gender, ethnicity, history of hypertension, diabetes, kidney disease, malignancy, education, family income, exercise, alcohol, smoking, obesity, and kidney function. All variables except uric acid and klotho levels were treated as categorical.