Minding the Gap Beyond Kidney Disease: Utility of the Anion Gap in Metabolic Syndrome

Obesity and metabolic syndrome (MetS) are common health problems, affecting more than 30% of Americans and increasing the risk of cardiovascular disease and mortality (1). Obesity has recently been identified as a risk factor for low serum bicarbonate (2) and elevated anion gap (AG) (3), regardless of CKD status. Table 1 shows different studies looking at the association between AG and obesity; however, they have not explored MetS directly, which is characterized by high triglyceride and fasting glucose levels and an increase in BP and central obesity and low HDL (1). In the CKD population, metabolic acidosis has been associated with insulin resistance, defective bone mineral metabolism, worsening kidney function, and increased mortality (4) (Figure 1). However, the effect of metabolic acidosis in non-CKD counterparts and the risk factors contributing to its presentation are less well understood. Notably, a cross-sectional analysis of the Chronic Renal Insufficiency Cohort Study showed an association between high waist circumference (WC) and lower serum bicarbonate levels; yet, this was again observed among patients with CKD (5).

Table 1.

Previously demonstrated associations between obesity/metabolic syndrome and acid-base disturbances

Condition or MetS Feature	Outcome Assessed	Association	Population	Reference
BMI	Bicarbonate	↓	Patients with and without CKD	Lambert and Abramowitz, 2021 (2)
BMI	Anion gap Bicarbonate	↑ ↓	Patients with and without CKD	Lambert and Abramowitz, 2021 (3)
BMI	Bicarbonate	↓	Patients with eGFR >60 ml/min per 1.73 m2	Driver et al., 2014 (18)
BMI	Anion gap Bicarbonate	↑ ↓	Patients without cardiovascular disease by report	Abramowitz et al., 2012 (19)
Diabetes	Bicarbonate	↓	Patients with eGFR >60 ml/min per 1.73 m2	Driver et al., 2014 (18)
Diabetes	Anion gap	↑	Patients without cardiovascular disease by report	Abramowitz et al., 2012 (19)
Diabetes	Bicarbonate	↓	Women who developed diabetes during study	Mandel et al., 2012 (20)
BP (systolic)	Anion gap Bicarbonate	↑ ↓	Patients without chronic diseases by report	Taylor et al., 2007 (13)
BP	Anion gap Bicarbonate	↑ ↓	Patients without cardiovascular disease by report	Abramowitz et al., 2012 (19)
HDL	Anion gap	↑	Patients with coronary artery disease with and without CKD	Yang et al., 2017 (15)
Insulin resistance	Anion gap Bicarbonate	↑ ↓	Adults without diabetes or other chronic diseases	Farwell and Taylor, 2008 (14)
Waist circumference	Bicarbonate	↓	Patients with CKD	Raphael et al., 2014 (5)

A positive or negative association between condition/metabolic syndrome feature and anion gap or bicarbonate is indicated by an upward or downward arrow, respectively. BMI, body mass index.

Figure 1.

Effects of metabolic acidosis in patients with CKD. Created with Biorender.com.

In this issue of Kidney 360, Lambert and colleagues conducted a retrospective analysis of clinical and biochemical parameters, including MetS features, serum bicarbonate, and AG values in more than 40,000 patients from the National Health and Nutrition Examination Survey (NHANES) (6). The authors had three major goals. First, they sought to demonstrate whether WC and number of MetS features were associated with low serum bicarbonate and/or elevated anion gap. Second, they aimed to determine whether metabolic disease, and not greater body mass index (BMI), accounted for these associations. Third, they examined whether these effects remained after adjusting for possible confounding factors such as insulin resistance, diabetes, hypertension, and CKD.

The authors demonstrated that increased WC and a greater number of MetS features were associated with lower bicarbonate and higher AG. By excluding patients with a BMI of ≥30 kg/m² and by examining associations with BMI after adjusting for WC or MetS, they demonstrated that BMI was not independently associated with these findings. When patients with diabetes and hypertension were excluded, associations of WC with AG were generally absent (with the exception of WC >111.5 cm), but associations of the number of MetS features with AG (other than for those with all five MetS features) and bicarbonate levels persisted. Findings were generally preserved in patients without CKD, defined as eGFR ≥90 ml/min per 1.73 m² and no albuminuria.

It is known that anion gap metabolic acidosis, normal anion gap metabolic acidosis, or both are commonly seen in patients with CKD, even in early stages and mild disease (4). Because of the aforementioned negative effects on CKD progression, bone health, and mortality in patients with CKD, metabolic acidosis has been evaluated in patients with kidney disease. However, data on acidosis and AG in non-CKD populations are limited.

Mechanisms for the development of metabolic acidosis in patients with MetS and no CKD have been proposed and include elevated lactate and small elevations in unmeasured metabolites. Regarding the former, MetS patients have been shown to have higher fasting and postprandial blood lactate, which is also chronically elevated in patients with obesity, dyslipidemia, and abnormal blood glucose (7). Suggested mechanisms include lactate production by adipose tissue (8) and impaired lactate clearance in conditions such as nonalcoholic fatty liver disease and nonalcoholic steatohepatitis (9). Regarding unmeasured anions, small elevations in serum pyruvate and other amino acids, and other intermediary metabolites (10), have been described in patients with obesity. Furthermore, greater dietary acid load is associated with lower serum bicarbonate in the general population (11), and impaired gastrointestinal alkali absorption has been seen in some MetS populations such as those with diabetes (12).

The analysis at hand informs the available literature mainly by demonstrating that obesity and MetS likely have independent effects on AG and bicarbonate outside of the CKD population, which to our knowledge has not been shown previously.

The most important limitation, acknowledged by the authors, is that NHANES data do not include blood gases or pH values, limiting a comprehensive acid-base evaluation. Using low serum bicarbonate levels as a surrogate for metabolic acidosis is not always appropriate. As Lambert and colleagues point out, if any underlying respiratory abnormality were present in the study population at large, it would more likely be respiratory acidosis associated with obesity hypoventilation, which would drive bicarbonate higher rather than lower. However, having pH values would allow for a more comprehensive assessment. The authors partly accounted for this by adjusting for a BMI ≥30 kg/m² to overcome ventilatory abnormalities seen at higher BMIs.

Importantly, the authors used different formulas to assess AG and had more impressive outcomes with full AG compared with the traditional and corrected AG, which are the AGs used in practice. Additionally, although MetS features and WC were associated with trends in bicarbonate and AG, mean bicarbonate was around 25 mEq/L and traditional anion gap was around 11, which does not indicate a state of acidosis. In fact, only 4%–11% of patients had a serum bicarbonate <22 mEq/L. One can see how these interpretations can become challenging in clinical practice, especially when laboratory variability also plays a role.

Additionally, the authors focused on the number of MetS features and WC in their analysis, rather than highlighting the independent effects of each MetS feature. Individual MetS features have been linked to various AG and bicarbonate levels in other studies (Table 1); for example, higher systolic BPs and insulin resistance have been associated with higher AG and lower bicarbonate (13, 14), and lipid profiles characterized by lower LDL and higher HDL levels have been associated with higher AG, presumably from statin use (15). It would be interesting to see whether BP, fasting glucose levels, hypertriglyceridemia, and low HDL were independently associated with bicarbonate and AG in the study population at hand.

The mechanism driving low bicarbonate levels in this study is yet to be elucidated. Because patients with self-described liver disease were excluded from the analysis, low lactate clearance from nonalcoholic fatty liver disease and nonalcoholic steatohepatitis is less likely a contributing factor. Furthermore, the authors attempted to correct for dietary acid but were only able to do this by relying upon dietary recall data, limiting accuracy. They also did not assess what medications patients were taking, many of which—metformin, salicylates, alcohols, and antimicrobials, among others—are associated with acidosis (16). The exclusion of patients with diabetes and hypertension partly avoids the confounding effects of offending agents commonly prescribed in these populations. However, the problem arises when people with higher BMIs have other conditions treated with such drugs, such as metformin for prediabetes or spironolactone for acne or polycystic ovarian syndrome.

As most nephrologists will attest, bicarbonate supplementation in CKD is often done despite the poor quality of supporting data. In patients with CKD, sodium bicarbonate is often supplemented to maintain a serum bicarbonate ≥22 to prevent CKD progression and death; however, the evidence on this is lacking. Some studies have shown relative preservation of renal function in CKD patients who received sodium bicarbonate supplementation compared with those who did not, but these have mainly been small, single-center studies, and the bicarbonate target of ≥22 is somewhat arbitrary, even on the basis of the studies that do exist (4). In acute settings, it is common practice to supplement sodium bicarbonate for patients with and without CKD who have a pH <7.1, but evidence for this practice is lacking as well, with mortality benefit only shown in patients with AKI. Additionally, this practice is driven by pH rather than bicarbonate (17). Research on the effect of bicarbonate supplementation in patients with chronic, mild acidosis without CKD—the population highlighted in the current study—is almost nil and remains to be explored.

In conclusion, low serum bicarbonate levels may not provide an accurate estimate of metabolic acidosis, and incorporating pH values would help assess the presence of other acid-base abnormalities. Furthermore, adding serum lactate levels, patients’ medications, and dietary patterns into future analyses may help elucidate possible etiologies of metabolic acidosis and AG in non-CKD populations. Finally, the study highlights the need for additional research on the role of bicarbonate supplementation in non-CKD patients with acidosis, and perhaps even in patients with CKD, given that its supplementation is often done with little evidence to support it.

The authors should be commended for bringing to light metabolic abnormalities that may affect far more patients than we previously knew. Given the prevalence of obesity and MetS, elucidating risk factors for negative health outcomes in this population could be far reaching.

Disclosures

C. Tamargo reports ownership interest in Amazon, Amedisys, AstraZeneca, Ayro, General Motors, Microsoft, Uber, and Vulcan. The remaining author has nothing to disclose.

Funding

None.

Author Contributions

C. Tamargo and C.E. Cervantes conceptualized the study, wrote the original draft, and reviewed and edited the manuscript, and C.E. Cervantes provided supervision.