Study characteristics:
The baseline characteristics of participants by quartile levels of serum bicarbonate level in the total population are shown in Table 1. A total of 2,981 Qatari participants were included in the study. The mean age was 36.4 ± 11.1 years. The mean level of serum bicarbonate was within the normal range (26.02 ± 2.12). As shown in Table 1, the univariate analysis demonstrated a significant association between bicarbonate quartiles level and other variables such as age, gender, BMI, smoking, renal function, and uric acid levels (p-value <0.01). Serum uric acid level was gradually increasing with the increase of serum bicarbonate level (p-value <0.001). The age, gender distribution, body mass index (BMI), and the number of smokers also gradually increased with the increase in bicarbonate level. However, the estimated glomerular filtration rate gradually decreased with the increased levels of serum bicarbonate (p-value <0.001).
Table 1. Sample characteristics per quartiles of bicarbonate among participants attending the Qatar Biobank Study
|
Total
|
Q1
|
Q2
|
Q3
|
Q4
|
p-value
|
|
N=2,981
|
N=748
|
N=915
|
N=580
|
N=738
|
|
Bicarbonate (mEq/L)
|
26.02 (2.12)
|
23.32 (1.05)
|
25.51 (0.47)
|
26.86 (0.30)
|
28.72 (1.07)
|
<0.001
|
Age
|
36.4 (11.1)
|
34.1 (10.3)
|
36.3 (11.0)
|
37.0 (11.3)
|
38.3 (11.4)
|
<0.001
|
Sex
|
|
|
|
|
|
<0.001
|
Men
|
1,430 (48.0%)
|
250 (33.4%)
|
406 (44.4%)
|
294 (50.7%)
|
480 (65.0%)
|
|
Women
|
1,551 (52.0%)
|
498 (66.6%)
|
509 (55.6%)
|
286 (49.3%)
|
258 (35.0%)
|
|
Smoking
|
|
|
|
|
|
0.001
|
Non-smoker
|
2,022 (67.8%)
|
529 (70.7%)
|
646 (70.6%)
|
394 (67.9%)
|
453 (61.4%)
|
|
Ex-smoker
|
189 ( 6.3%)
|
48 ( 6.4%)
|
47 ( 5.1%)
|
41 ( 7.1%)
|
53 ( 7.2%)
|
|
Current smoker
|
770 (25.8%)
|
171 (22.9%)
|
222 (24.3%)
|
145 (25.0%)
|
232 (31.4%)
|
|
Uric acid (umol/L)
|
296.53 (81.89)
|
283.29 (85.55)
|
296.08 (82.54)
|
297.58 (81.59)
|
309.69 (75.26)
|
<0.001
|
BMI (kg/m2)
|
28.4 (5.9)
|
28.8 (6.3)
|
28.8 (5.8)
|
28.1 (5.6)
|
27.8 (5.6)
|
0.003
|
Calcium (mmol/L)
|
2.28 (0.08)
|
2.27 (0.08)
|
2.28 (0.07)
|
2.28 (0.08)
|
2.29 (0.07)
|
<0.001
|
Phosphorus (mmol/L)
|
1.15 (0.17)
|
1.15 (0.16)
|
1.15 (0.16)
|
1.16 (0.18)
|
1.16 (0.17)
|
0.28
|
Thyroid Stimulating Hormone (mIU/L)
|
1.61 (1.15)
|
1.63 (1.05)
|
1.57 (0.97)
|
1.56 (0.92)
|
1.68 (1.53)
|
0.14
|
Total Testosterone
|
9.72 (10.67)
|
6.48 (8.58)
|
8.84 (10.45)
|
10.08 (10.34)
|
13.79 (11.73)
|
<0.001
|
Estradiol (pmol/L)
|
231.12 (281.04)
|
290.68 (298.82)
|
247.35 (294.62)
|
215.32 (296.50)
|
163.11 (208.55)
|
<0.001
|
AST (GOT) (U/L)
|
19.69 (10.72)
|
19.09 (11.83)
|
19.05 (7.67)
|
20.05 (10.88)
|
20.81 (12.43)
|
0.002
|
ALT ( GPT ) (U/L)
|
22.57 (18.24)
|
21.34 (19.98)
|
21.82 (14.36)
|
22.54 (14.90)
|
24.75 (22.44)
|
0.001
|
eGFR (mL/min/1.73m2)
|
111.35 (15.07)
|
115.16 (14.84)
|
111.46 (14.84)
|
110.28 (14.82)
|
108.19 (14.96)
|
<0.001
|
eGFR<90 mL/min/1.73m2
|
256 ( 8.6%)
|
42 ( 5.6%)
|
77 ( 8.4%)
|
52 ( 9.0%)
|
85 (11.5%)
|
<0.001
|
BMI: body mass index; AST: Alanine aminotransferase; ALT: Alanine Aminotransferase; eGFR: estimated glomerular filtration rate
Association between Serum bicarbonate and uric acid levels:
Gender-specific correlation pattern was also evident in this study in which there was a negative correlation between serum bicarbonate and uric acid levels in men without adjustments as shown in figure 1. While in the female gender, no association was evident between serum bicarbonate and uric acid levels.
To further analyze the association, a multivariate regression analysis with variably adjusted associations was conducted between the quartiles of bicarbonate and serum uric acid. As shown in Table 2, two models were adjusted in which compared to the first quartile, the second, third, and fourth quartiles of serum bicarbonate were independently associated with lower serum uric acid levels after adjustment for age, BMI, smoking, and kidney function.
Table 2 Regression coefficients (95%CI) for the association of serum uric acid by quartiles of bicarbonate in men and women Qatari adults attending Qatar Biobank
|
Serum Bicarbonate level
|
|
|
Q1
|
Q2
|
Q3
|
Q4
|
P for trend value
|
Men
|
|
|
|
|
|
Model 1
|
0.00
|
-14.05
|
(-24.76 - -3.35)
|
-18.74
|
(-30.19 --7.29)
|
-25.79
|
(-36.17 --15.40)
|
0.000
|
Model 2
|
0.00
|
-12.32
|
(-22.69 - -1.95)
|
-11.50
|
(-22.65 --0.35)
|
-15.59
|
(-25.88 --5.30)
|
0.011
|
Women
|
|
|
|
|
|
|
|
|
Model 1
|
0.00
|
6.68
|
(-0.54 -13.91)
|
-0.01
|
(-8.54 - 8.53)
|
1.34
|
(-7.70 -10.38)
|
0.980
|
Model 2
|
0.00
|
6.48
|
(-0.49 -13.45)
|
3.13
|
(-5.14 -11.41)
|
3.51
|
(-5.19 -12.21)
|
0.492
|
Model 1 adjusted for age; Model 2 further adjusted for BMI, smoking, and eGFR<90 mL/min/1.73m2. The mean serum bicarbonate levels in Men (Q1: 23.39, Q2: 25.56, Q3: 26.90, and Q4: 28.77); in women (23.27, 25.47, 26.8, and 28.6 respectively).
To confirm the dose-response association in men, the association was tested for nonlinearity using the restricted cubic spline regression method following the same adjustments for age, BMI, smoking, and renal function. As shown in Figure 2, there was a significant linear association between serum bicarbonate and the regression coefficients of uric acid confirming the dose-response effect in men. In women, no significant association was found between quartiles of serum bicarbonate and uric acid level following the same adjustments. Interestingly, upon using the same restricted cubic spline method, a significant association with binomial distribution was demonstrated at the serum bicarbonate level of 25 mEq/L. The association between serum bicarbonate and the variation coefficients of uric acid seemed to be positive at lower levels and negative at higher levels as shown in Figure 2.