Abstract
Context
The evidence regarding bicarbonate status and mortality among diabetes is scarce.
Objective
The purpose of this study was to investigate the associations of bicarbonate concentrations with risk of all-cause, cardiovascular disease (CVD), and cancer mortality among patients with type 2 diabetes (T2D).
Methods
This study included 8163 adult diabetic patients from the National Health and Nutrition Examination Survey (NHANES), 1999 to 2018. Death outcomes were ascertained by linkage to National Death Index records through 31 December 2019. The Cox proportional-risk model was used to estimate hazard ratios (HR) and 95% CIs for mortality from all causes, CVD, and cancer. The mediating effects of 11 metabolic, cardiovascular, and renal biomarkers were evaluated using a logistic regression model within a counterfactual framework.
Results
During 8163 person-years of follow-up, 2310 deaths were documented, including 659 CVD deaths and 399 cancer deaths. After multivariate adjustment, lower serum bicarbonate levels were significantly linearly correlated with higher all-cause, CVD, and cancer mortality: The risk of all-cause death increased by 40%, the risk of CVD death increased by 48%, and the risk of cancer death increased by 84% compared with the normal group (all P < .05). Altered levels of estimated glomerular filtration rate explained 12.10% and 16.94% of the relation between serum bicarbonate with all-cause and CVD mortality, respectively. Total cholesterol mediated 4.70% and 10.51% of the associations of all-cause and CVD mortality, respectively.
Conclusion
Lower serum bicarbonate concentrations were significantly associated with higher all-cause, CVD, and cancer mortality. These findings suggest that maintaining adequate bicarbonate status may lower mortality risk in individuals with T2D.
Keywords: serum bicarbonate, NHANES, diabetes, mortality, cardiovascular diseases
Type 2 diabetes mellitus (T2D) is a complex metabolic disease that has become a major public health problem worldwide (1, 2). More than 463 million people are now suffering from diabetes, and the prevalence rate is expected to increase to 578 million by 2030 (3). Diabetes has been well known to lead to increased cardiovascular disease (CVD) and all-cause mortality (4). Compared with the general population, the risk of death from CVD in patients with T2D is increased by about 2 to 4 fold (2). It is therefore important to fully identify modifiable factors associated with premature mortality in people with diabetes.
Bicarbonate is a major ion that is involved in many essential metabolic processes and is responsible for maintaining the acid-base balance in the body fluid. Recently, it was observed that low bicarbonate concentration significantly decreased the glucose-induced insulin secretion of different mouse pancreatic β cells (5). Serum bicarbonate has been proposed as an independent predictor of all-cause mortality and CVD death in studies particularly in patients with chronic kidney disease (CKD), but also in the general population (6–16). However, among patients with diabetes who often have a heightened proinflammatory and oxidative status, as well as an impaired metabolism of acid-base balance, studies regarding the potential health effect of bicarbonate are limited and inconsistent.
Just 2 of these studies included only patients with diabetes, and the evidence regarding the relationship between serum bicarbonate status and mortality risk among diabetes is limited and somewhat mixed without a specific focus on diabetes. For instance, one prospective study of 1283 community-based diabetic patients from the Fremantle Diabetes Study found that serum bicarbonate is an important independent predictor of cardiovascular events, but does not include the death or heart failure of patients with T2D in the community (11). Another study of 2628 adults with T2D and kidney disease found that after adjusting the estimated glomerular filtration rate (eGFR), the correlation between serum bicarbonate level and the end point of kidney disease was not retained, which was in contrast with the results of early studies in nondiabetic people (12). Moreover, these existing studies are subject to a few limitations, such as small sample sizes, participants only had established CKD, and insufficient adjustment of some important covariates (eg, lifestyle factors including physical activity).
Therefore, the purpose of our study was to investigate the effect of serum bicarbonate levels on all-cause and specific-cause mortality of the national representative sample of US adults with T2D. We validated the hypothesis that diabetes state modulates the effect of serum bicarbonate levels on mortality.
Materials and Methods
Study Population
The National Center for Health Statistics (NCHS) established the National Health and Nutrition Examination Survey (NHANES), a series of cross-sectional surveys using a complex multistage probability design, that obtained representative samples of the noninstitutionalized civilian population in the United States. The NHANES examination includes interview survey questions on health and nutrition and physical examination with laboratory tests. Details about the study have been published elsewhere (17–20). The NHANES research was approved and agreed to by the NCHS Research Ethics Review Committee. We analyzed data from the survey interview and physical examination within continuous NHANES (1999-2018, n = 101 369). For this analysis, 34 735 individuals (age < 18 years) were excluded and the present study enrolled 59 204 adult participants (age ≥ 18 years). In the present study, we defined diabetes using 1 of the following 5 criteria: self-reported doctor diagnosis of diabetes, intake of antidiabetic medications or insulin, glycated hemoglobin A1c (HbA1c) level greater than or equal to 6.5%, fasting glucose level of greater than or equal to 7.0 mmol/L, or a 2-hour glucose level of 11.1 mmol/L after an oral glucose tolerance test. Among those with diabetes, we excluded individuals with possible type 1 diabetes (n = 306), defined by use of insulin and age at diabetes onset younger than 30 years (validated as accurate in 97% of cases); the remaining 9004 individuals were defined as having T2D (21). Ultimately, 8163 diabetic patients measured serum bicarbonate concentration with available mortality follow-up data (Fig. 1).
Figure 1.
This flowchart of the study population describes the composition of the present sample of participants. NHANES, National Health and Nutrition Examination Survey; T1D, type 1 diabetes; T2D, type 2 diabetes.
Measurement of Bicarbonate
In the NHANES 1999 to 2000, serum bicarbonate concentrations were measured by a Hitachi Model 917 multichannel analyzer (Roche Diagnostics). Starting from the 2001 to 2006 cycle, serum bicarbonate concentrations were measured by a Beckman Synchron LX20 (Beckman Coulter Inc). Starting from the 2007 to 2008 cycle, serum bicarbonate concentrations were measured by a Beckman Synchron LX20 and Beckman UniCel DxC800 Synchron (Beckman Coulter Inc). Starting from the 2009 to 2014 cycle, serum bicarbonate concentrations were measured by a Beckman UniCel DxC800 Synchron (Beckman Coulter Inc). Starting from the 2015 to 2016 cycle, serum bicarbonate concentrations were measured by a Beckman UniCel DxC800 Synchron (Beckman Coulter Inc) and Beckman Coulter UniCel DxC 660i Synchron Access chemistry analyzer (Beckman Coulter Inc). Starting from the 2017 to 2018 cycle, serum bicarbonate concentrations were measured by a Roche Cobas 6000 Chemistry Analyzer (Roche Diagnostics Corp). The baseline concentration of bicarbonate in serum was measured at the baseline level of the study and analyzed according to the following clinically relevant stratification as categorical variables: low (< 22 mEq/L), normal (22-26 mEq/L), and high (> 26 mEq/L).
Outcome Ascertainment
The main results of this study are all-cause, CVD, and cancer mortality. The mortality data as of December 31, 2019, were determined by linking with the National Death Index, whose death certificate records meet the following data: Social Security number, name, date of birth, race/nationality, sex, birth status, and residence status. The root cause of death is coded in accordance with the 10th revision of the International Classification of Diseases (ICD). We divided the mortality into all-causes, CVDs (ICD codes: I00-I09, I11, I13, I20-I51), cancer-related diseases (ICD codes: C00-C97), and other causes (including chronic lower respiratory diseases [ICD 10 codes: J40-J47], accidents [ICD 10 codes: V01-X59, Y85-Y86], cerebrovascular disease [ICD 10 codes: I60-I69], Alzheimer disease [ICD 10 code: G30], diabetes mellitus [ICD 10 codes: E10-E14], influenza and pneumonia [ICD 10 codes: J10-J18], nephritis, nephrotic syndrome, and nephrosis [ICD 10 codes: N00-N07, N17-N19, N25-N27], and all other causes [residual]).
Covariant Evaluation
Information about age, race, alcohol, smoking in the past month, education level, family income, physical activity, diabetes, drug use, and CVD status (using Rose angina questionnaire) is self-reported. Participants received medical examination to measure weight, standing height, and waist circumference in a standardized way. Race and ethnicity are divided into 4 categories: non-Hispanic White; non-Hispanic Black; Mexican American, and others, including other Hispanic, Asian, and multiracial participants. Body mass index (BMI) is defined as body weight (kilograms) divided by height (meters) squared. Alcohol consumption was defined as regular drinking in the past 12 months (yes or no). Smoking was coded as nonsmoker, former smoker, and current smoker. Participants who smoked less than 100 cigarettes in their lifetime were classified as never smoking. Former smokers are defined as people who have smoked more than 100 cigarettes in their lifetime but have given up smoking. At present, smokers are defined as those who have smoked more than 100 cigarettes in their lives but still smoke. We divided the education level into less than high school, high school graduation or general educational development certificate, and greater than high school. According to the relationship between self-reported family income and poverty line, family size, and calendar year, the poverty-income ratio (PIR) of each family is calculated. A value of 1 or less is lower than the official poverty threshold, whereas a PIR value higher than 1 indicates that the income is higher than the poverty level (range, 0-5). Physical activity is assessed by the number of moderate to high-intensity activities (such as walking, jogging, running, swimming, cycling, dancing, or yard work) per week, while lack of physical activity is defined as never performing moderate or high-intensity activities. The use of prescription drugs was evaluated by self-report and verified by the interviewer by checking the drug container. Hypertension was defined as self-reported history of hypertension, measured systolic blood pressure greater than or equal to 140 mm Hg, measured diastolic blood pressure greater than or equal to 90 mm Hg, or self-reported blood pressure medication. Participants are considered to have CVD if they were told by doctors or other health experts that they have congestive heart failure, coronary heart disease, angina pectoris, heart attack, or stroke. The eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration Study equation, and CKD was defined as an eGFR less than or equal to 60 ml/min/1.73 m2 (22). The Homeostasis Model Assessment of Insulin Resistance (HOMA-IR) was determined by the formula: HOMA-IR = fasting plasma insulin (μU/L) × fasting blood glucose (mmol/L)/22.5.
Statistical Analysis
Due to the complex sampling design of NHANES, all the analyses include the research visit weight, main sampling units, and hierarchical design of the NHANES survey(17). A P value less than .05 was used as a cutoff for statistical significance. Analyses were conducted using IBM SPSS statistical software (version 24, IBM), R packages (http://www.r-project.org), and Stata statistical software (version 16.0, Stata Corp).
Continuous variables are expressed as mean SD, while classified variables are expressed as numbers and their proportions. We used the chi-square test for classified variables, one-way analysis of variance for normal continuous variables, and the Kruskal-Wallis test for skewed continuous variables.
The generalized linear model was applied to examine the associations of serum bicarbonate levels with cardiometabolic biomarkers at baseline, including plasma glucose, insulin, HOMA-IR, HbA1c, total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), triglycerides, C-reactive protein (CRP), urine albumin/creatinine, and eGFR, adjusted for potential confounders of the serum bicarbonate–mortality associations including age, sex, race/ethnicity, BMI, drinking status, smoking status, education, PIR, physical activity, duration of diabetes, medication use, HbA1c, self-reported diseases, and CKD. Poisson distribution was used to estimate the mortality rates per 1000 person-years and the 95% CI of each major cause of death in every layer of serum bicarbonate.
The Cox proportional-hazard regression model was used to calculate the hazard ratio (HRs) and 95% CI of the relationship between serum bicarbonate level and all-cause mortality and specific-cause mortality (CVD and cancer). We tested our models 1 to 3 by adjusting the risk factors step by step. Model 1 was adjusted only for age (years), sex (male or female), and race/ethnicity (non-Hispanic White, non-Hispanic Black, Mexican American, or other), and model 2 was further adjusted for BMI (continuous), drinking status (< 12 dozen drinks/year; ≥ 12 dozen drinks/year), smoking status (never, former, or current smoker), education level (< high school, high school or equivalent, or > high school), family PIR (continuous), as well as physical activity (never, moderate, or vigorous). Then, model 3 was simultaneously adjusted for the variables in model 2 plus the duration of diabetes (≤ 3, 3-10, or 10 years), and diabetes medication use (none, oral medication, and/or insulin), HbA1c (continuous), self-reported diseases (hypertension, congestive heart failure, coronary heart disease, angina, heart attack, stroke), and CKD.
The associations between levels of bicarbonate and all end points were evaluated on a continuous scale with restricted cubic spline curves (4 knots) based on Cox proportional-hazards models using 25 mEq/L as the reference value.
The mediation analysis was performed with R package “mediation,” version 4.2.1, and 1000 permutations and bootstrapping to estimate the CI. We used the calculated coefficients from the generalized linear model to characterize the outcome variables and the calculated coefficients from the linear regression model to characterize the mediator variables.
Results
Baseline Characteristics
Table 1 shows the selected demographic characteristics and potential confounding factors of 3 types of serum bicarbonate status. In the continuous NHANES (1999-2018) data set, a total of 8163 participants older than 18 years with diabetes had measures of serum bicarbonate. The average age of the final study population was 63 years (interquartile range, 53-72 years); 51.9% were male. Participants who had lower bicarbonate levels are more prone to be younger, non-Hispanic White, obese, current smokers, materially deprived, and physically active. In addition, the prevalence of CKD was higher (all P < .05).
Table 1.
Baseline characteristics of National Health and Nutrition Examination Survey participants with diabetes according to serum bicarbonate
| Characteristics | Serum bicarbonate, mEq/La | ||||
|---|---|---|---|---|---|
| No. of patients | Total (N = 8163) | < 22 (n = 656) | 22-26 (n = 5416) | > 26 (n = 2091) | P |
| Demographics and clinical characteristics | |||||
| Age, y | 63 (53-72) | 60 (46-69) | 63 (52-72) | 66 (57-74) | < .001 |
| Male, % | 4240 (51.942) | 318 (48.476) | 2828 (52.216) | 1094 (52.319) | .179 |
| Race/ethnicity | < .001 | ||||
| Non-Hispanic White | 2958 (36.237) | 254 (38.720) | 1943 (35.875) | 761 (36.394) | |
| Non-Hispanic Black | 1977 (24.219) | 151 (23.018) | 1248 (23.043) | 578 (27.642) | |
| Mexican American | 1692 (20.728) | 170 (25.915) | 1203 (22.212) | 319 (15.256) | |
| Other | 1536 (18.817) | 81 (12.348) | 1022 (18.870) | 433 (20.708) | |
| BMI | 30.84 (26.97-35.88) | 31.56 (27.21-37.63) | 31 (27.14-35.6) | 30.33 (26.58-35.9) | < .001 |
| Drinking status | .200 | ||||
| < 12 dozen drinks/y | 2588 (35.584) | 219 (36.379) | 1688 (34.883) | 681 (37.172) | |
| ≥ 12 dozen drinks/y | 4685 (64.416) | 383 (63.621) | 3151 (65.117) | 1151 (62.828) | |
| Smoking status | < .001 | ||||
| Nonsmoker | 4028 (75.360) | 308 (68.597) | 2658 (74.684) | 1062 (79.432) | |
| Former smoker | 218 (4.079) | 23 (5.122) | 151 (4.243) | 44 (3.291) | |
| Current smoker | 1099 (20.561) | 118 (26.281) | 750 (21.073) | 231 (17.277) | |
| Education level | .733 | ||||
| < High school | 3080 (37.801) | 255 (38.931) | 2052 (37.951) | 773 (37.057) | |
| High school or equivalent | 1865 (22.889) | 157 (23.969) | 1226 (22.674) | 482 (23.106) | |
| > High school | 3203 (39.310) | 243 (37.099) | 2129 (39.375) | 831 (39.837) | |
| Family poverty-income ratio | 1.8 (1.03-3.4) | 1.5 (0.92-3.12) | 1.84 (1.05-3.43) | 1.79 (1.03-3.38) | .001 |
| Leisure time physical activity | .034 | ||||
| Never | 4483 (54.925) | 348(53.049) | 2967 (54.792) | 1168 (55.858) | |
| Moderate | 1758 (21.539) | 151 (23.018) | 1205 (22.253) | 402 (19.225) | |
| Vigorous | 1921 (23.536) | 157 (23.933) | 1243 (22.955) | 521 (24.916) | |
| Duration of diabetes, y | .114 | ||||
| ≤ 3 | 3624 (44.636) | 289 (44.257) | 2421 (44.858) | 914 (44.176) | |
| 3-10 | 2205 (27.159) | 167 (25.574) | 1502 (27.830) | 536 (25.906) | |
| > 10 | 2290 (28.205) | 197 (30.168) | 1474 (27.311) | 619 (29.918) | |
| Medication use | .016 | ||||
| No insulin or pills | 1352 (16.563) | 97 (14.787) | 906 (16.728) | 349 (16.691) | |
| Only diabetes pills | 3670 (44.959) | 264 (40.244) | 2472 (45.643) | 934 (44.668) | |
| Any insulin use | 1372 (16.808) | 133 (20.274) | 871 (16.082) | 368 (17.599) | |
| Other | 1769 (21.671) | 162 (24.695) | 1167 (21.547) | 440 (21.043) | |
| HbA1c | 6.7 (6.1-7.8) | 6.8 (6-8) | 6.8 (6.1-7.9) | 6.7 (6.1-7.6) | .006 |
| Self-reported diseases | |||||
| Hypertension | 5846 (71.633) | 461 (70.274) | 3790 (69.991) | 1595 (76.316) | < .001 |
| Congestive heart failure | 732 (9.062) | 54 (8.372) | 429 (7.999) | 249 (12.029) | < .001 |
| Coronary heart disease | 820 (10.184) | 60 (9.274) | 516 (9.663) | 244 (11.816) | .017 |
| Angina | 583 (7.232) | 39 (6.084) | 384 (7.174) | 160 (7.741) | .353 |
| Heart attack | 853 (10.517) | 65 (10.015) | 541 (10.050) | 247 (11.881) | .063 |
| Stroke | 704 (8.672) | 50 (7.692) | 468 (8.691) | 186 (8.929) | .617 |
| CKD | 1661 (20.353) | 181 (27.634) | 1076 (19.871) | 404 (19.321) | < .001 |
Data are numbers (percentages). All estimates accounted for complex survey designs.
Abbreviations: BMI, body mass index; CKD, chronic kidney disease; HbA1c, glycated hemoglobin A1c.
Serum bicarbonate status was categorized into 3 groups: < 22 mEq/L, 22-26 mEq/L, and > 26 mEq/L.
Least Squares Means of Cardiometabolic Markers
The least squares means of cardiometabolic biomarkers according to serum bicarbonate concentrations are shown in Table 2. Lower levels of serum bicarbonate were significantly correlated with higher levels of glucose, HOMA-IR, HbA1c, total cholesterol, LDL, triglycerides, CRP, urine albumin/creatinine, and with lower levels of HDL at baseline (all P < .05).
Table 2.
Least squares means of metabolic, cardiovascular, and renal markers according to serum bicarbonate concentrations among patients with diabetes in the National Health and Nutrition Examination Survey 1999 to 2018
| Characteristics | Serum bicarbonate, mEq/L | |||
|---|---|---|---|---|
| <22 (n = 656) | 22-26 (n = 5416) | >26 (n = 2091) | P | |
| Glucose (n = 4439), mmol/L | 8.75 ± 0.25 | 8.63 ± 0.09 | 7.99 ± 0.14 | <0.001 |
| Insulin (n = 4414), pmol/L | 25.07 ± 2.07 | 20.65 ± 0.76 | 19.94 ± 1.21 | 0.092 |
| HOMA-IR (n = 4404) | 10.27 ± 0.87 | 8.02 ± 0.32 | 7.28 ± 0.51 | 0.012 |
| HbA1c (n = 8148), % | 7.34 ± 0.09 | 7.32 ± 0.03 | 7.12 ± 0.06 | 0.007 |
| Total cholesterol (n = 8161), mmol/L | 5.09 ± 0.06 | 5.04 ± 0.02 | 4.86 ± 0.04 | <0.001 |
| HDL (n = 8160), mmol/L | 1.19 ± 0.02 | 1.25 ± 0.01 | 1.32 ± 0.01 | <0.001 |
| LDL (n = 4098), mmol/L | 2.87 ± 0.07 | 2.91 ± 0.03 | 2.73 ± 0.04 | 0.001 |
| Triglycerides (n = 4406), mmol/L | 2.69 ± 0.13 | 1.94 ± 0.05 | 1.55 ± 0.08 | <0.001 |
| CRP (n = 6357), mg/L | 8.04 ± 0.65 | 6.50 ± 0.24 | 5.76 ± 0.40 | 0.012 |
| Urine albumin/creatinine (n = 7925), mg/g | 222.07 ± 34.55 | 130.80 ± 12.51 | 125.84 ± 20.51 | 0.037 |
| eGFR (n = 8161), mL/min/1.73 m2 | 83.17 ± 1.30 | 84.93 ± 0.47 | 83.00 ± 0.77 | 0.066 |
The least squares (mean ± SE) was estimated using general linear model with adjustment of age (continuous), sex (male or female) and race/ethnicity (non-Hispanic White, non-Hispanic Black, Mexican American, or other), body mass index (continuous), drinking status (< 12 dozen drinks/y, ≥ 12 dozen drinks/y), smoking status (never smoker, former smoker, or current smoker), education level (< high school, high school or equivalent, or > high school), family poverty-income ratio (continuous), physical activity (never, moderate, or vigorous), duration of diabetes (≤ 3, 3-10, or 10 years), diabetes medication use (none, oral medication, and/or insulin), glycated hemoglobin A1c (continuous), self-reported diseases (hypertension, congestive heart failure, coronary heart disease, angina, heart attack, stroke), and chronic kidney disease.
Abbreviations: CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; HOMA-IR, homeostasis model assessment for insulin resistance; LDL, low-density lipoprotein.
Causes of Death Mortality Rates
Over a median follow-up of 7.0 years (interquartile range, 3.1-11.3 years), 2310 patients died (no weight). The main causes of death were CVD (n = 659) and malignant tumors (n = 399) (no weight). The highest mortality rate per 1000 person-years was observed for the lower levels of serum bicarbonate. The all-cause mortality rate was 3.52 per 1000 person-years (95% CI, 3.07-3.98), the CVD mortality rate was 0.90 per 1000 person-years (95% CI, 0.67-1.13), and the cancer-related mortality was 0.69 per 1000 person-years (95% CI, 0.49-0.89). Table 3 describes the overall and specific-cause mortality rates per 1000 person-years in the serum bicarbonate layer.
Table 3.
Causes of death and mortality rates per 1000 person-years across serum bicarbonate strata
| Serum bicarbonate, mEq/L | |||
|---|---|---|---|
| < 22 (n = 656) | 22-26 (n = 5416) | > 26 (n = 2091) | |
| Cause of death | Mortality rate per 1000 person-y (95% CI) | ||
| All-cause deaths | 3.52 (3.07-3.98) | 2.80 (2.66-2.94) | 2.69 (2.47-2.91) |
| Cardiovascular diseases | 0.90 (0.67-1.13) | 0.81 (0.73-0.89) | 0.77 (0.65-0.89) |
| Malignancy | 0.69 (0.49-0.89) | 0.49 (0.43-0.55) | 0.43 (0.34-0.52) |
| Chronic lower respiratory diseases | 0.11 (0.03-0.19) | 0.10 (0.07-0.12) | 0.20 (0.14-0.26) |
| Accidents (unintentional injuries) | 0.15 (0.06-0.25) | 0.04 (0.03-0.06) | 0.03 (0.01-0.05) |
| Cerebrovascular disease | 0.17 (0.07-0.27) | 0.18 (0.15-0.22) | 0.14 (0.09-0.19) |
| Alzheimer disease | 0.02 (-0.01-0.05) | 0.08 (0.05-0.10) | 0.09 (0.05-0.13) |
| Diabetes mellitus | 0.40 (0.24-0.55) | 0.25 (0.21-0.29) | 0.28 (0.21-0.35) |
| Influenza and pneumonia | 0.06 (0.00-0.12) | 0.06 (0.04-0.08) | 0.06 (0.02-0.09) |
| Nephritis, nephrotic syndrome, and nephrosis | 0.20 (0.09-0.31) | 0.09 (0.07-0.12) | 0.09 (0.05-0.13) |
Associations Between Serum Bicarbonate Level and All-Cause Mortality
The multivariate-adjusted associations of serum bicarbonate levels with mortality are shown in Fig. 2. As shown in Table 4, after adjusting for demographics, lower level of serum bicarbonate was positively associated with all-cause mortality (HR 1.51; 95% CI, 1.32-1.74; P < .001). In the multivariate-adjusted Cox regression model (model 3), the risk of all-cause mortality of individuals in the lower group was 40% higher (HR, 1.40; 95% CI, 1.15-1.71; P = .001) than that of participants with normal serum bicarbonate concentration (see Table 4). The detailed analysis of Cox proportional risk confirms these observations (Fig. 3A).
Figure 2.
Full multivariable model–adjusted survival curve for the target results by baseline serum bicarbonate stratification. A, All-cause mortality. B, Cardiovascular disease (CVD) mortality. C, Cancer mortality. Cox regression model with complex investigation design was used for analysis. Hazard ratios (HRs) were adjusted according to age (continuous), sex (male or female), and race/ethnicity (non-Hispanic White, non-Hispanic Black, Mexican American, or other), body mass index (continuous), drinking status (< 12 dozen drinks/y, ≥ 12 dozen drinks/y), smoking status (never smoker, former smoker, or current smoker), education level (< high school, high school or equivalent, or > high school), family poverty-income ratio (continuous), physical activity (never, moderate, or vigorous), duration of diabetes (≤ 3, 3-10, or 10 years), diabetes medication use (none, oral medication, and/or insulin), glycated hemoglobin A1c (continuous), self-reported diseases (hypertension, congestive heart failure, coronary heart disease, angina, heart attack, stroke), and chronic kidney disease.
Table 4.
Cox regression models for the association between serum bicarbonate groups and clinical outcomes accounting for complex survey design
| Serum bicarbonate, mEq/L | |||
|---|---|---|---|
| < 22 (n = 656) | 22-26 (n = 5416) | > 26 (n = 2091) | |
| All-cause mortality | |||
| No. of deaths | 231 | 1517 | 562 |
| Model 1 | 1.51 (1.32-1.74); < 0.001 | Ref | 0.96 (0.87-1.06); 0.407 |
| Model 2 | 1.52 (1.24-1.85); < 0.001 | Ref | 0.90 (0.78-1.05); 0.190 |
| Model 3 | 1.40 (1.15-1.71); 0.001 | Ref | 0.93 (0.80-1.09); 0.366 |
| CVD mortality | |||
| No. of deaths | 59 | 439 | 161 |
| Model 1 | 1.36 (1.03-1.79); 0.028 | Ref | 0.95 (0.79-1.14); 0.573 |
| Model 2 | 1.65 (1.15-2.36); 0.006 | Ref | 1.02 (0.78-1.33); 0.884 |
| Model 3 | 1.48 (1.03-2.13); 0.035 | Ref | 1.05 (0.80-1.37); 0.743 |
| Cancer mortality | |||
| No. of deaths | 45 | 264 | 90 |
| Model 1 | 1.68 (1.22-2.30); 0.001 | Ref | 0.87 (0.68-1.11); 0.264 |
| Model 2 | 1.91 (1.22-2.97); 0.004 | Ref | 0.90 (0.62-1.29); 0.560 |
| Model 3 | 1.84 (1.17-2.89); 0.008 | Ref | 0.92 (0.64-1.34); 0.677 |
Model 1 is adjusted for age (continuous), sex (male, or female) and race/ethnicity (non-Hispanic White, non-Hispanic Black, Mexican American, or other).
Model 2 is adjusted for variables in model 1 + body mass index (continuous), drinking status (< 12 dozen drinks/y; ≥ 12 dozen drinks/y), smoking status (never smoker, former smoker, or current smoker), education level (< high school, high school or equivalent, or > high school), family poverty-income ratio (continuous), and physical activity (never, moderate or vigorous).
Model 3 is adjusted for variables in model 2 + duration of diabetes (≤ 3, 3-10, or 10 years), and diabetes medication use (none, oral medication, and/or insulin), glycated hemoglobin A1c (continuous), self-reported diseases (hypertension, congestive heart failure, coronary heart disease, angina, heart attack, stroke), and chronic kidney disease.
Abbreviations: CVD, cardiovascular disease; Ref, reference.
Figure 3.
Multivariable adjusted smoothing spline plots of serum bicarbonate levels with A, all-cause, and B, cardiovascular disease (CVD), and C, cancer mortality among type 2 diabetes in the National Health and Nutrition Examination Survey 1999 to 2018. Hazard ratios were adjusted according to age (continuous), sex (male or female) and race/ethnicity (non-Hispanic White, non-Hispanic Black, Mexican American, or other), body mass index (continuous), drinking status (< 12 dozen drinks/y, ≥ 12 dozen drinks/y), smoking status (never smoker, former smoker, or current smoker), education level (< high school, high school or equivalent, or > high school), family poverty-income ratio (continuous), physical activity (never, moderate or vigorous), duration of diabetes (≤ 3, 3-10, or 10 years), diabetes medication use (none, oral medication, and/or insulin), glycated hemoglobin A1c (continuous), self-reported diseases (hypertension, congestive heart failure, coronary heart disease, angina, heart attack, stroke), and chronic kidney disease. Both P linearity less than .001. The black curve represents the best fit line, and the gray curve represents 95% CI.
Fig. 4 describes the subgroup analysis by forest plot. In subgroup analyses, people who were female, non-White, and more likely to have increased all-cause mortality had lower levels of serum bicarbonate concentration compared with the normal group (all P < .05) (Fig. 4A). In the subgroup with CKD at baseline, participants with serum bicarbonate concentrations below 22 mEq/L were at increased risk for all-cause mortality compared to the normal bicarbonate group (HR, 1.49; 95% CI, 1.10-2.03; P = .011).
Figure 4.
Serum bicarbonate levels with A, all-cause and B, cardiovascular disease (CVD), and C, cancer mortality in subgroups. Hazard ratios (HRs) were adjusted according to age (continuous), sex (male or female) and race/ethnicity (non-Hispanic White, non-Hispanic Black, Mexican American, or other), body mass index (continuous), drinking status (< 12 dozen drinks/y, ≥ 12 dozen drinks/y), smoking status (never smoker, former smoker, or current smoker), education level (< high school, high school or equivalent, or > high school), family poverty-income ratio (continuous), physical activity (never, moderate or vigorous), duration of diabetes (≤ 3, 3-10, or 10 years), diabetes medication use (none, oral medication, and/or insulin), glycated hemoglobin A1c (continuous), self-reported diseases (hypertension, congestive heart failure, coronary heart disease, angina, heart attack, stroke), and chronic kidney disease (CKD). P values are for interaction between subgroup and bicarbonate levels.
Associations Between Serum Bicarbonate Level and Cardiovascular Mortality
As shown in Fig. 2B, participants with serum bicarbonate concentrations below 22 mEq/L had a statistically significantly higher rate of CVD death. Analysis of serum bicarbonate concentrations using Cox regression estimates as classification predictors showed that patients in the group with the lowest serum bicarbonate concentration had a significantly higher risk of CVD death compared with the reference category after adjusting the variables in model 3 (HR, 1.48; 95% CI, 1.03-2.13; P = .035) (see Table 4). Fig. 3B shows a dose-response relationship between serum bicarbonate concentration (range, 20-28 mEq/L) CVD mortality.
In subgroup analyses, people who were White were more likely to have increased CVD mortality with lower serum bicarbonate concentration compared with the normal group, but there was not statistical significance in these analyses (Fig. 4B). In the subgroup without CKD at baseline, participants with serum bicarbonate concentrations below 22 mEq/L were at increased risk for CVD mortality compared to the normal bicarbonate group (HR, 1.71; 95% CI, 1.06-2.78; P = .029).
Associations Between Serum Bicarbonate Level and Cancer Mortality
After multivariate adjustment including lifestyle factors, BMI, diabetes duration, diabetes medication use, and presence of chronic diseases, compared to the normal bicarbonate group, participants with serum bicarbonate concentrations below 22 mEq/L had a significantly increased the risk of cancer mortality (Fig. 2C). Compared with the normal group, participants with serum bicarbonate concentrations below 22 mEq/L had a cancer mortality rate of 84% higher (HR, 1.84; 95% CI, 1.17-2.89; P = .008) (see Table 4). The finding of multivariate-adjusted smooth spline graph is usually congruent with the categorical results, which indicates that patients with low levels of serum bicarbonate are at higher risk of dying from cancer (Fig. 3C).
In subgroup analyses, people who were non-White were more likely to have increased cancer mortality with lower serum bicarbonate concentration compared with the normal group (HR, 1.98; 95% CI, 1.10-3.57; P = .022) (Fig. 4C). In the subgroup without CKD at baseline, participants with serum bicarbonate concentrations below 22 mEq/L were at increased risk for cancer mortality compared to the normal bicarbonate group (HR, 1.91; 95% CI, 1.05-3.48; P = .035).
Mediation Analysis
When analyzing biological mechanisms potentially linking serum bicarbonate to all-cause and CVD mortality, we found that part of such relations were most pronounced in altered eGFR accounting for 12.10% and 16.94%, respectively, of the excess of all-cause and CVD mortality associated with serum bicarbonate (Table 5). Total cholesterol mediated 4.70% and 10.51% of the associations of all-cause and CVD mortality, respectively (see Table 5). Glucose explained 4.57% of the relationship between serum bicarbonate and all-cause mortality (see Table 5). HbA1c explained 6.37% of the relationship between serum bicarbonate and CVD mortality (see Table 5). We did not find evidence for metabolic, cardiovascular, or renal marker mediation between serum bicarbonate and cancer mortality (see Table 5).
Table 5.
Mediation analysis for metabolic, cardiovascular, and renal markers as possible mediators of the association of serum bicarbonate with clinical outcomes among individualsa
| Total effect | Direct effect | Indirect effect | % mediatedb | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OR | Lower 95% CI |
Upper 95% CI |
P | OR | Lower 95% CI |
Upper 95% CI |
P | OR | Lower 95% CI |
Upper 95% CI |
P | ||
| All-cause mortality | |||||||||||||
| Glucose, mmol/L | 0.9848 | 0.9823 | 0.9907 | < .001 | 0.9855 | 0.9828 | 0.9914 | < .001 | 0.9993 | 0.9985 | 0.9999 | .006 | 4.57 |
| Insulin, pmol/L | 0.9848 | 0.9823 | 0.9906 | < .002 | 0.9849 | 0.9823 | 0.9907 | < .002 | 0.9999 | 0.9996 | 1.0001 | .432 | 0.65 |
| HOMA-IR | 0.9850 | 0.9823 | 0.9912 | < .003 | 0.9852 | 0.9824 | 0.9915 | < .003 | 0.9998 | 0.9993 | 1.0001 | .230 | 1.32 |
| HbA1c, % | 0.9850 | 0.9826 | 0.9896 | < .004 | 0.9853 | 0.9827 | 0.99 | < .004 | 0.9997 | 0.9993 | 1.0000 | .056 | 1.99 |
| Total cholesterol, mmol/L | 0.9852 | 0.9826 | 0.9898 | < .005 | 0.9858 | 0.9831 | 0.9906 | < .005 | 0.9993 | 0.9987 | 0.9998 | .004 | 4.70 |
| HDL, mmol/L | 0.9851 | 0.9826 | 0.9899 | < .006 | 0.9853 | 0.9826 | 0.99 | < .006 | 0.9998 | 0.9994 | 1.0003 | .502 | 1.33 |
| LDL, mmol/L | 0.9853 | 0.9820 | 0.9927 | < .007 | 0.9857 | 0.9822 | 0.9931 | < .007 | 0.9996 | 0.9989 | 1.0001 | .106 | 2.70 |
| Triglycerides, mmol/L | 0.9845 | 0.9823 | 0.9906 | < .008 | 0.9860 | 0.9829 | 0.9922 | < .008 | 0.9985 | 0.9967 | 1.0001 | .070 | 9.61 |
| CRP, mg/L | 0.9853 | 0.9834 | 0.9902 | < .009 | 0.9856 | 0.9836 | 0.9905 | < .009 | 0.9998 | 0.9994 | 1.0001 | .136 | 1.35 |
| Urine albumin/creatinine, mg/g | 0.9851 | 0.9824 | 0.9900 | < .010 | 0.9852 | 0.9825 | 0.9901 | < .010 | 0.9999 | 0.9993 | 1.0004 | .570 | 0.67 |
| eGFR, mL/min/1.73 m2 | 0.9844 | 0.9827 | 0.9881 | < .011 | 0.9862 | 0.9841 | 0.9903 | < .011 | 0.9981 | 0.9971 | 0.9989 | < .001 | 12.10 |
| CVD mortality | |||||||||||||
| Glucose, mmol/L | 0.9978 | 0.9900 | 1.0013 | .602 | 0.9979 | 0.9901 | 1.0014 | .638 | 0.9999 | 0.9993 | 1.0002 | .442 | 4.54 |
| Insulin, pmol/L | 0.9976 | 0.9883 | 1.0013 | .606 | 0.9974 | 0.9880 | 1.0012 | .568 | 1.0001 | 0.9999 | 1.0006 | .242 | −4.16 |
| HOMA-IR | 0.9974 | 0.9881 | 1.0013 | .588 | 0.9972 | 0.9880 | 1.0013 | .546 | 1.0002 | 0.9999 | 1.0007 | .240 | −7.68 |
| HbA1c, % | 0.9953 | 0.9867 | 1.0006 | .128 | 0.9956 | 0.9871 | 1.0007 | .150 | 0.9997 | 0.9993 | 1.0000 | .040 | 6.37 |
| Total cholesterol, mmol/L | 0.9962 | 0.9875 | 1.0009 | .208 | 0.9966 | 0.9882 | 1.0011 | .260 | 0.9996 | 0.9991 | 0.9999 | .008 | 10.51 |
| HDL, mmol/L | 0.9959 | 0.9874 | 1.0008 | .170 | 0.9959 | 0.9874 | 1.0008 | .162 | 1.0000 | 0.9995 | 1.0005 | .948 | 0.00 |
| LDL, mmol/L | 0.9980 | 0.9891 | 1.0012 | .678 | 0.9981 | 0.9892 | 1.0013 | .720 | 0.9999 | 0.9994 | 1.0001 | .206 | 5.00 |
| Triglycerides, mmol/L | 0.9974 | 0.9879 | 1.0013 | .564 | 0.9979 | 0.9883 | 1.0015 | .650 | 0.9996 | 0.9987 | 1.0001 | .152 | 15.37 |
| CRP, mg/L | 0.9938 | 0.9824 | 1.0007 | .114 | 0.9938 | 0.9826 | 1.0007 | .116 | 1.0000 | 0.9998 | 1.0001 | .536 | 0.00 |
| Urine albumin/creatinine, mg/g | 0.9952 | 0.9859 | 1.0007 | .138 | 0.9952 | 0.9859 | 1.0007 | .136 | 1.0000 | 0.9998 | 1.0001 | .690 | 0.00 |
| eGFR, mL/min/1.73 m2 | 0.9947 | 0.9860 | 1.0001 | .064 | 0.9957 | 0.9871 | 1.0007 | .148 | 0.9991 | 0.9982 | 0.9997 | < .001 | 16.94 |
| Cancer mortality | |||||||||||||
| Glucose, mmol/L | 0.9883 | 0.9716 | 1.0004 | .086 | 0.9885 | 0.9715 | 1.0005 | .090 | 0.9998 | 0.9988 | 1.0006 | .614 | 1.70 |
| Insulin, pmol/L | 0.9886 | 0.9714 | 1.0004 | .082 | 0.9887 | 0.9715 | 1.0004 | .082 | 0.9999 | 0.9995 | 1.0002 | .586 | 0.87 |
| HOMA-IR | 0.9886 | 0.9710 | 1.0005 | .090 | 0.9888 | 0.9715 | 1.0005 | .088 | 0.9998 | 0.9990 | 1.0003 | .436 | 1.74 |
| HbA1c, % | 0.9946 | 0.9819 | 1.0005 | .146 | 0.9944 | 0.9818 | 1.0005 | .126 | 1.0002 | 1.0000 | 1.0007 | .092 | −3.69 |
| Total cholesterol, mmol/L | 0.9939 | 0.9803 | 1.0006 | .136 | 0.9942 | 0.9807 | 1.0006 | .158 | 0.9998 | 0.9993 | 1.0000 | .058 | 3.27 |
| HDL, mmol/L | 0.9940 | 0.9803 | 1.0005 | .140 | 0.9938 | 0.9803 | 1.0004 | .120 | 1.0002 | 0.9996 | 1.0008 | .492 | −3.32 |
| LDL, mmol/L | 0.9909 | 0.9726 | 1.0008 | .194 | 0.9912 | 0.9733 | 1.0008 | .214 | 0.9997 | 0.9989 | 1.0001 | .182 | 3.28 |
| Triglyceride, mmol/L | 0.9874 | 0.9712 | 1.0000 | .048 | 0.9876 | 0.9710 | 1.0001 | .058 | 0.9998 | 0.9988 | 1.0009 | .682 | 1.58 |
| CRP, mg/L | 0.9964 | 0.9844 | 1.0008 | .412 | 0.9964 | 0.9846 | 1.0009 | .428 | 0.9999 | 0.9996 | 1.0000 | .162 | 2.77 |
| Urine albumin/creatinine, mg/g | 0.9928 | 0.9793 | 1.0002 | .080 | 0.9928 | 0.9792 | 1.0002 | .066 | 1.0001 | 0.9997 | 1.0004 | .684 | −1.38 |
| eGFR, mL/min/1.73 m2 | 0.9934 | 0.9805 | 1.0003 | .092 | 0.9932 | 0.9799 | 1.0003 | .076 | 1.0002 | 0.9994 | 1.0012 | .542 | −3.02 |
Abbreviations: CRP, C-reactive protein; CVD, cardiovascular disease; eGFR, estimated glomerular filtration rate; HbA1c, glycated hemoglobin A1c; HOMA-IR, Homeostasis Model Assessment of Insulin Resistance; HDL, high-density lipoprotein; LDL, low-density lipoprotein; OR, odds ratio.
Hazard ratios were adjusted for age (continuous), sex (male or female) and race/ethnicity (non-Hispanic White, non-Hispanic Black, Mexican American, or other), body mass index; (continuous), drinking status (< 12 dozen drinks/y, ≥ 12 dozen drinks/y), smoking status (never smoker, former smoker, or current smoker), education level (< high school, high school or equivalent, or > high school), family poverty-income ratio (continuous), physical activity (never, moderate, or vigorous), duration of diabetes (≤ 3, 3-10, or 10 years), and diabetes medication use (none, oral medication, and/or insulin), HbA1c (continuous), self-reported diseases (hypertension, congestive heart failure, coronary heart disease, angina, heart attack, stroke), and chronic kidney disease.
The percentage mediated was calculated by log (indirect effect)/log (total effect).
Discussion
In a large prospective cohort study of US adults with diabetes, we observed that during a mean follow-up of 7.0 years, patients with serum bicarbonate less than 22 mEq/L had an increased risk of all-cause, CVD, and cancer death compared to normal tertiles in analyses of multivariate-adjusted Cox regression model analyses. These associations all have nothing to do with traditional risk factors, including lifestyle factors, BMI, duration of diabetes, use of diabetes drugs, and self-reported diseases (hypertension, congestive heart failure, coronary heart disease, angina, heart attack, stroke), or CKD. In addition, lower levels of serum bicarbonate were significantly associated with higher levels of glucose, HOMA-IR, HbA1c, total cholesterol, LDL, triglycerides, CRP, urine albumin/creatinine, and with lower levels of HDL at baseline (all P < .05). The largest mediated effect contributing to the indirect effect was eGFR (proportion mediated effect 16.94%). To the best of our knowledge, this is the first large-scale NHANES study to investigate the specific-cause mortality related to serum bicarbonate level in patients with T2D.
We have shown that lower serum bicarbonate (< 22 mEq/L) has a higher risk of death from all causes than the reference group in the model 3 adjustment analysis. Several other studies have also explored the relationship between serum bicarbonate levels and mortality. They all included mixed populations, most of whom being patients without diabetes. The only other 2 studies on the relationship between bicarbonate level and diabetic all-cause mortality are not completely consistent with our observation, that is, they show no independent relationship between adjusted serum bicarbonate and death. In a study of 1283 participants in the Fremantle Diabetes Study, patients with the lower quintile of serum bicarbonate had a higher risk of death than patients with the highest quintile without adjustment analysis. In the fully adjusted model, patients with the lower serum bicarbonate level did not have an increased risk (11). Comprehensive data of patients with CKD are involved in the study of Irbesartan Diabetic Nephropathy Trial (IDNT) or Angiotensin II Antagonist Losartan (RENAAL) trial to reduce the end point of non–insulin-dependent diabetes mellitus. These correlations have nothing to do with age, sex, or cardiovascular risk factors, but disappear after adjusting baseline to eGFR (all P > .05) (12). Results from their study are not entirely consistent with our study because of variable sample sizes and differing cohort composition. Our study included one of the largest cohorts of patients with T2D, and our cohort consisted of non-Hispanic White, non-Hispanic Black, Mexican American, and other populations.
Few studies have assessed the relationship between serum bicarbonate level and CVD, and the results are contradictory. The data obtained from the NHANES (1999-2010) from 31 195 participants representing the general population in the United States also showed that analysis of serum bicarbonate concentrations using Cox regression estimates as classification predictors showed no significant correlation between serum bicarbonate concentrations and CVD mortality. Our study found that there was a significant and independent inverse correlation between lower bicarbonate level and CVD mortality. The association between lower serum bicarbonate and development of diabetes may be explained by metabolic acidosis promoting insulin resistance. Alternatively, it could be due to low concentrations of extracellular bicarbonate significantly impairing the functioning of pancreatic β cells by reducing coupling Ca2+ influx. Therefore, the selective T2D population participating in our study might lead to survival results that differ from outcomes in the general population.
Cancer-related mortality in the low bicarbonate group was higher than that in normal bicarbonate group. In the constantly adjusted model, this association remained unchanged, including lifestyle factors, BMI, diabetes duration, and diabetes drug use. Diabetes has been proved to increase the risk of cancer (23). However, regardless of the history of diabetes, the serum bicarbonate level of the NHANES may be directly related to cancer mortality from 1999 to 2010. The underlying mechanism leading to this association is still speculative and cannot be confirmed by this epidemiological cohort, but there are several crucial points. The findings reported by animal studies suggest that acidosis is a critical factor to induce and propagate carcinogenesis (24). In addition, the acid produced by the tumor microenvironment may promote the local invasion of tumor cells by destroying the extracellular matrix (25).
Advantages of the present research include its prospective research design, relatively large sample size, and the use of national representative samples of American adult diabetic patients, which facilitates the generalization of our findings. In addition, given the comprehensive data acquired in the NHANES, we could adjust for a multitude of potential confounding factors including socioeconomic status, race/ethnicity and lifestyle factors, comorbidities, diabetes duration, and diabetes medication use. Several limitations should be considered as well. First, causality cannot be determined because of the observational study design. Second, the present study measured serum bicarbonate concentrations only once, which would likely underestimate the association of interest (26). Third, the present study did not have detailed information on the severity of diabetes, although the results are still significant when further adjusting for diabetes duration, diabetes medication use, HbA1c levels, eGFR, and the number of self-reported comorbidities. Fourth, mortality outcomes were ascertained by linkage to the National Death Index with a probabilistic match, which might result in misclassification. However, a prior validation study showed the high accuracy of the method (27). Finally, in the present research, we cannot rule out the effect of confounding caused by psychosocial stress or genetic susceptibility, residual or unknown confounding, or chance.
Conclusion
In a nationally representative sample of US adults with diabetes, we found that lower serum bicarbonate concentrations were prominently associated with higher all-cause, CVD, and cancer mortality. These findings support full and powerful future clinical trials of bicarbonate alternatives for high-risk patients with T2D and patients with low bicarbonate levels.
Acknowledgments
The study was approved by the National Center for Health Statistics Research Ethics Review Board, and NHANES obtained the written informed consent from all participants.
The authors thank the data collection team and the management of NHANES for sharing the NHANES data, which enabled us to create this paper.
Abbreviations
- BMI
body mass index
- CRP
C-reactive protein
- CVD
cardiovascular disease
- eGFR
estimated glomerular filtration rate
- HbA1c
glycated hemoglobin A1c
- HOMA-IR
Homeostasis Model Assessment of Insulin Resistance
- HDL
high-density lipoprotein
- HR
hazard ratio
- ICD
International Classification of Diseases
- LDL
low-density lipoprotein
- NCHS
National Center for Health Statistics
- NHANES
National Health and Nutrition Examination Survey
- PIR
poverty-income ratio
- T2D
type 2 diabetes
Contributor Information
Yilan Li, Department of Cardiology, the Second Affiliated Hospital of Harbin Medical University, Harbin 150001, Nangang, China; Key Laboratory of Myocardial Ischemia, Ministry of Education, Harbin Medical University, Harbin 150001, Nangang, China.
Rong Gao, Department of Cardiology, the Second Affiliated Hospital of Harbin Medical University, Harbin 150001, Nangang, China.
Bing Zhao, Department of Cardiology, the Second Affiliated Hospital of Harbin Medical University, Harbin 150001, Nangang, China; Key Laboratory of Myocardial Ischemia, Ministry of Education, Harbin Medical University, Harbin 150001, Nangang, China.
Yao Zhang, Department of Cardiology, the Second Affiliated Hospital of Harbin Medical University, Harbin 150001, Nangang, China; Key Laboratory of Myocardial Ischemia, Ministry of Education, Harbin Medical University, Harbin 150001, Nangang, China.
Financial Support
This work was supported by the National Natural Science Foundation of China (grant Nos. 81770255 and 82000381), Heilongjiang Province Postdoctoral Science Foundation (grant No. LBH-Z19188), the Open Project of Key Laboratory of Myocardial Ischemia, Ministry of Education (grant No. KF202103), and the Open Project Program of Key Laboratory of Preservation of Human Genetic Resources and Disease Control in China (Harbin Medical University), Ministry of Education (grant No. LPHGRD2022-001).
Author Contributions
Y.L. and R.G. designed the analysis and wrote the first draft of the article. R.G. and B.Z. conducted the model parameterization and the statistical analyses. Y.Z. revised the manuscript. All authors read and approved the final manuscript.
Disclosures
The authors have nothing to disclose.
Data Availability
Original data generated and analyzed during this study are included in this published article or in the data repositories listed in “References.”
References
- 1. Zimmet P, Alberti KG, Shaw J. Global and societal implications of the diabetes epidemic. Nature. 2001;414(6865):782–787. [DOI] [PubMed] [Google Scholar]
- 2. Harding JL, Pavkov ME, Magliano DJ, Shaw JE, Gregg EW. Global trends in diabetes complications: a review of current evidence. Diabetologia. 2019;62(1):3–16. [DOI] [PubMed] [Google Scholar]
- 3. Saeedi P, Petersohn I, Salpea P, et al. ; IDF Diabetes Atlas Committee . Global and regional diabetes prevalence estimates for 2019 and projections for 2030 and 2045: results from the International Diabetes Federation Diabetes Atlas, 9th edition. Diabetes Res Clin Pract. 2019;157:107843. [DOI] [PubMed] [Google Scholar]
- 4. Raghavan S, Vassy JL, Ho YL, et al. Diabetes mellitus-related all-cause and cardiovascular mortality in a national cohort of adults. J Am Heart Assoc. 2019;8(4):e011295. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Zhang YC, Xiong FR, Wang YY, et al. High bicarbonate concentration increases glucose-induced insulin secretion in pancreatic β-cells. Biochem Biophys Res Commun. 2022;589:165–172. [DOI] [PubMed] [Google Scholar]
- 6. Kovesdy CP, Anderson JE, Kalantar-Zadeh K. Association of serum bicarbonate levels with mortality in patients with non-dialysis-dependent CKD. Nephrol Dial Transplant. 2008;24(4):1232–1237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Navaneethan SD, Schold JD, Arrigain S, et al. Serum bicarbonate and mortality in stage 3 and stage 4 chronic kidney disease. Clin J Am Soc Nephrol. 2011;6(10):2395–2402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Raphael KL, Zhang Y, Wei G, Greene T, Cheung AK, Beddhu S. Serum bicarbonate and mortality in adults in NHANES III. Nephrol Dial Transplant. 2013;28(5):1207–1213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Raikou VD, Kyriaki D. Mortality and low serum bicarbonate level in patients on hemodiafiltration versus peritoneal dialysis. Indian J Nephrol. 2018;28(2):105–112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Bommer J, Locatelli F, Satayathum S, et al. Association of predialysis serum bicarbonate levels with risk of mortality and hospitalization in the Dialysis Outcomes and Practice Patterns Study (DOPPS). Am J Kidney Dis. 2004;44(4):661–671. [PubMed] [Google Scholar]
- 11. Paul Chubb SA, Davis WA, Peters KE, Davis TME. Serum bicarbonate concentration and the risk of cardiovascular disease and death in type 2 diabetes: the Fremantle Diabetes Study. Cardiovasc Diabetol. 2016;15(1):143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Schutte E, Lambers Heerspink HJ, Lutgers HL, et al. Serum bicarbonate and kidney disease progression and cardiovascular outcome in patients with diabetic nephropathy: a post hoc analysis of the RENAAL (Reduction of End Points in Non-Insulin-Dependent Diabetes With the Angiotensin II Antagonist Losartan) Study and IDNT (Irbesartan Diabetic Nephropathy Trial). Am J Kidney Dis. 2015;66(3):450–458. [DOI] [PubMed] [Google Scholar]
- 13. Dobre M, Yang W, Pan Q, et al. ; CRIC Study Investigators . Persistent high serum bicarbonate and the risk of heart failure in patients with chronic kidney disease (CKD): a report from the Chronic Renal Insufficiency Cohort (CRIC) study. J Am Heart Assoc. 2015;4(4):e001599. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Raphael KL, Murphy RA, Shlipak MG, et al. ; Health ABC Study . Bicarbonate concentration, acid-base status, and mortality in the Health, Aging, and Body Composition Study. Clin J Am Soc Nephrol. 2016;11(2):308–316. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Raphael KL, Wei G, Baird BC, Greene T, Beddhu S. Higher serum bicarbonate levels within the normal range are associated with better survival and renal outcomes in African Americans. Kidney Int. 2011;79(3):356–362. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Dobre M, Rahman M, Hostetter TH. Current status of bicarbonate in CKD. J Am Soc Nephrol. 2015;26(3):515–523. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Johnson CL, Paulose-Ram R, Ogden CL, et al. National Health and Nutrition Examination Survey: analytic guidelines, 1999-2010. Vital Health Stat 2. 2013;(161):1–24. [PubMed] [Google Scholar]
- 18. Chen TC, Clark J, Riddles MK, Mohadjer LK, Fakhouri THI. National Health and Nutrition Examination Survey, 2015-2018: sample design and estimation procedures. Vital Health Stat 2. 2020;(184):1–35. [PubMed] [Google Scholar]
- 19. Curtin LR, Mohadjer LK, Dohrmann SM, et al. National Health and Nutrition Examination Survey: sample design, 2007-2010. Vital Health Stat 2. 2013;(160):1–23. [PubMed] [Google Scholar]
- 20. Johnson CL, Dohrmann SM, Burt VL, Mohadjer LK. National Health and Nutrition Examination Survey: sample design, 2011-2014. Vital Health Stat 2. 2014;(162):1–33. [PubMed] [Google Scholar]
- 21. Gregg EW, Cheng YJ, Srinivasan M, et al. Trends in cause-specific mortality among adults with and without diagnosed diabetes in the USA: an epidemiological analysis of linked national survey and vital statistics data. Lancet. 2018;391(10138):2430–2440. [DOI] [PubMed] [Google Scholar]
- 22. Levey AS, Stevens LA, Schmid CH, et al. ; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) . A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604–612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Scully T, Ettela A, LeRoith D, Gallagher EJ. Obesity, type 2 diabetes, and cancer risk. Front Oncol. 2021;10:615375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Fais S, Venturi G, Gatenby B. Microenvironmental acidosis in carcinogenesis and metastases: new strategies in prevention and therapy. Cancer Metastasis Rev. 2014;33(4):1095–1108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Estrella V, Chen T, Lloyd M, et al. Acidity generated by the tumor microenvironment drives local invasion. Cancer Res. 2013;73(5):1524–1535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Clarke R, Shipley M, Lewington S, et al. Underestimation of risk associations due to regression dilution in long-term follow-up of prospective studies. Am J Epidemiol. 1999;150(4):341–353. [DOI] [PubMed] [Google Scholar]
- 27. Wan Z, Guo J, Pan A, Chen C, Liu L, Liu G. Association of serum 25-hydroxyvitamin D concentrations with all-cause and cause-specific mortality among individuals with diabetes. Diabetes Care. 2021;44(2):350–357. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Original data generated and analyzed during this study are included in this published article or in the data repositories listed in “References.”