Carbon dioxide overload, detected in human blood, suggests a potentially toxic atmosphere within 50 years

AbstractAnthropogenic activities are increasing the amount of carbon dioxide (CO2) in the atmosphere. There is mounting experimental evidence that lifetime exposure to these increasing atmospheric CO2 levels can negatively impact the normal physiology of organisms. However, directly assessing this in humans is very difficult. We analysed serum bicarbonate (HCO3−), calcium (Ca) and phosphorus (P) levels from the U.S. National Health and Nutrition Examination Survey (NHANES) from 1999 to 2020 as indirect proxies for atmospheric CO2 exposure. Over this period, average bicarbonate levels in this population show an increasing trend which parallels rising atmospheric CO2 concentrations. Both Ca and P have decreased steadily over the same period. If these trends continue, blood bicarbonate values could be at the limit of the accepted healthy range in half a century, and Ca and P will be at the limit of their healthy ranges by the end of this century. Studies indicate that, after this time, elevated atmospheric carbon dioxide, leading to CO2 accumulation in the body, has the potential to cause a range of adverse health effects. These findings highlight the urgent need for significant reductions in anthropogenic CO2 emissions to safeguard public health. Similar content being viewed by others IntroductionIn aerobic organisms, including humans, CO2 is produced as a by-product of cellular respiration and needs to be removed from the body as a waste product via expiration (Raven et al., 2007). CO2 crosses the cell plasma membrane, enters the blood and almost all of it (90–95%) diffuses into the red blood cells, where it is rapidly hydrated to hydrogen (H+) and bicarbonate (HCO3−) by the enzyme carbonic anhydrase (CA) (Arlot-Bonnemains et al., 1985; Lan et al., 2025). HCO3− in blood is the most important means of transport for CO2 throughout the body (Sherwood, 2013) and when that blood reaches the lungs, the reaction is reversed. HCO3− in the lungs combines with H+ produced by the oxygenation of deoxyhemoglobin to produce water and CO2, which is exhaled as a waste product (Lan et al.,

2025). Carbonic anhydrase thus allows a large pool of otherwise slowly reacting plasma HCO3− to be utilized in CO2 excretion (Arlot-Bonnemains et al., 1985). A quantitative relationship exists between plasma HCO3− and levels of CO2, in the blood whereby when CO2 increases, so does HCO3− and vice versa. This is evident in multiple clinical situations (Martinu et al., 2003; Ueda et al., 2009). Importantly, abnormal CO2 retention, or impaired CO2 elimination, which is seen in diseases (Mendez et al., 2019; Palmer & Clegg, 2023), impaired ventilation, or excessive CO2 inhalation (Robertson, 2006), can result in reduced blood pH. The body then attempts to buffer this ‘acidosis’ by various mechanisms, including increased ventilation (if possible), increased renal excretion of acid and conservation of filtered HCO3− (Sherwood, 2013), retention of calcium, phosphates and other substances (Gray et al., 1973) and nervous system stimulation to counteract the direct effects of pH changes on heart contractility and vasodilation (Burton, 1978; Eckenhoff & Longnecker, 1995). Calcium and phosphate (PO₄³⁻) play important supporting roles in maintaining blood acid-base balance, working alongside the HCO3− buffer system. Phosphate can accept or donate hydrogen ions to help stabilize pH, and when blood becomes acidic, calcium and phosphate can be released from bone to help neutralize excess acid (Salcedo-Betancourt & Moe, 2024).This leads to the question as to whether increases in atmospheric CO2 pose a threat to human health. There are now many studies that review health effects in the range 600–5,000 ppm CO2 (Azuma et al., 2018; Bierwirth, 2025; Carr et al., 2025; Jacobson et al., 2019) although there is still a paucity of studies investigating the effects of long-term exposure at relevant levels of CO2.

As such, the U.S. National Health and Nutrition Examination Survey (NHANES) study, which recorded blood chemistry parameters, such as HCO3−, calcium (Ca), and phosphorus (P), from about 7,000 people every two years between 1999 and 2020, provides a unique opportunity to assess potential secular trends in human blood biochemistry that may be a result of a changing air composition.Atmospheric carbon dioxide throughout human evolutionThe ancestor of modern humans is thought to have evolved between 5 and 8 million years ago, with the first Homo sapiens appearing in the fossil record around 150,000 years ago (Wood, 1996). Although not precise, it appears that throughout most, if not all, of the ensuing period of human evolution, levels of CO2 in the atmosphere remained relatively stable around 300 parts per million (ppm). These data were derived from a combination of studies of relict features including air trapped in ice cores (Barnola et al., 1987), the composition of fossil plankton (Zachos et al., 2001) and Carbon-13 (13C) content in fossil plant material (Cui et al., 2020). However, since the advent of widespread industrialisation, atmospheric CO2 levels have exponentially increased (Fig. 1). In just the last ~ 50 years it has risen from < 340 ppm (in 1980), to > 420 ppm in 2025 (Lan et al., 2025). Atmospheric CO2 is currently increasing at more than 2 ppm each year, largely due to humanity’s activities, such as the burning of fossil fuels (Eggleton, 2012).Fig. 1Atmospheric carbon dioxide concentrations (in ppm) over the last 800,000 years, based on measurements of air trapped in Antarctic ice (Lüthi et al., 2008; Rubino et al., 2019), and direct measurements made at the Mauna Loa Observatory (1958 to present) (Keeling et al.,

2005)Temporal trend in human serum bicarbonate, calcium and phosphorus levelsAs levels of atmospheric CO2 increase, it is obvious that humans will have no option but to inhale more CO2. It is also known that increased blood loading with CO2 (hypercapnia) is physiologically correlated with HCO3− levels (Malte & Wang, 2024), due to the hydration of CO₂ to form carbonic acid, which dissociates into H⁺ and HCO₃⁻. To compensate, renal mechanisms increase HCO₃⁻ reabsorption and generation to buffer H⁺ and partially restore pH homeostasis (Alka & Casey, 2014). This potentially represents a health risk if HCO3− levels increase above the normal healthy range and/or the duration of increased HCO3− is excessive. In healthy adult humans, arterial HCO3− is typically between 22 and 26mEq/L (Larkin & Zimmanck, 2015), although a recent re-evaluation suggests that 22.1–28.3mEq/L is more appropriate for arterial blood, and up to 30mEq/L is appropriate for venous blood (Kraut & Madias, 2018).Although temporal population biochemical data relating to HCO3− are rare, a previous study (Zheutlin et al., 2014) examined data from the U.S. National Health and Nutrition Examination Survey (NHANES) from 1999 to 2012 (7 cycles) looking at the population average HCO3− levels in blood samples from a total of 33,546 adults (~ 5,000 per cycle). Between 1999 and 2012, there was an upward trend in serum HCO3− levels in the study population, with it increasing approximately 5%, from ~ 23.8mEq/L in 1999–2000 to ~ 25.0mEq/L in 2011–2012. This increase paralleled atmospheric CO2 levels which increased by a similar proportion over the same period (from ~ 369ppm to ~ 393ppm) (Lan et al.,

2025). The authors question whether “the increasing trend in serum bicarbonate found in our study is related to elevated ambient CO2 and climate change”.MethodsThe overarching NHANES protocol is ethically approved by the Ethics Review Committee of the National Center for Health Statistics, which involved obtaining informed consent from all participants. The NHANES project assesses health and nutrition in a representative sample of adults and children in the United States via interviews, physical examinations and laboratory tests.In the present study, we extend the 7-cycle trend analysis of Zheutlin et al. (2014) to include more recent and comprehensive NHANES data. Our analyses encompass an additional 4-cycles (up to 2019) and include all records from the NHANES database (i.e. from birth to 80 + years of age; ~7,000 records per cycle). From the database, we extracted data on levels of bicarbonate (mEg/L), calcium (mmol/L) and phosphorus (mmol/L) in participant serum. These data were averaged across all participants for each cycle, giving mean values at each time point. Atmospheric carbon dioxide levels at Mauna Loa, Hawaii over the same time period were obtained the Global Monitoring Laboratory dataset (Lan et al., 2025).ResultsOur key finding was that the increase identified by Zheutlin et al. (2014) has continued (Fig. 2). The most recent serum HCO3− levels measured in the NHANES population (which uses venous blood) was 25.3mEq/L (in 2019–2020), representing a ~ 7% increase from 1999. Similarly, this has paralleled atmospheric CO2 over the same period. As stated above, the upper healthy limit for HCO3− in venous blood can be taken as 30mEq/L (Kraut & Madias, 2018), although this value requires further scrutiny particularly given that the high HCO3− condition would likely be perpetual in the future as atmospheric CO2 continues to increase. Assuming a linear relationship and a ~ 0.34% increase rate per year, the calculated trendline in Fig.

2A predicts that the healthy maximum HCO3− level of 30mEq/L will be reached in the year 2076. Due to the large sample sizes, population level variation in serum HCO3− levels is low (SE based on unweighted data < 0.032 for all years).Fig. 2(A) Comparison between temporal trend in population serum bicarbonate (filled circles, left y-axis) in U.S. adults from the NHANES biochemistry database and measured atmospheric concentration CO2 (open boxes, dotted line, right y-axis) at Mauna Loa, Hawaii (Lan et al., 2025). The black line is the trendline for HCO3− with the formula y = 0.081x-138.15. (B) Temporal trends in population serum calcium (filled triangles, solid line left y-axis) and phosphorus (open diamonds, dotted line, right y-axis) in U.S. adults from the NHANES biochemistry database over the same period of timeSimilarly, Fig. 2B illustrates changes in serum calcium (Ca) and phosphorus (P) levels in the same population between 1999 and 2020. Mean serum Ca levels have decreased ~ 2% over this period, while P levels have decreased by ~ 7% (if the potentially spurious measurement in NHANES 1999–2000 is excluded). Total serum Ca levels for healthy adults are generally accepted to be between 2.1 and 2.6 mmol/L, with levels below this termed hypocalcemia (Bazydlo et al., 2014). For P the healthy range is 0.81 to 1.45 mmol/L (Bazydlo et al., 2014), below which hypophosphatemia occurs. Like serum HCO3−, due to the large sample sizes, population level variation in both Ca and P are low (SE based on unweighted data < 0.0012 for Ca and < 0.0027 for P for all years). If we also assume linear relationships for these parameters, the calculated trendlines in Fig.