CO₂ and Cancer Evidence Guide
Focused evidence from carbogen, transcutaneous CO₂ and CO₂-saturated solution studies
Over the course of the growth and development of Conscious Breathing — through product development, applied research, clinical questions, and the research behind two books — we have assembled a large scientific library around carbon dioxide physiology. This focused cancer guide narrows that material to studies in which CO₂ itself was part of the intervention.
The question here is deliberately simple: when researchers added CO₂ to an oxygen-rich breathing gas, delivered CO₂ directly across tissue, or infused a CO₂-saturated solution into the tumor circulation, what happened? The aim is to show the CO₂-specific evidence cleanly, while keeping human clinical findings distinct from animal and translational work.
Evidence labels describe study type, not a formal quality grade. Human, translational and animal studies are identified explicitly. The “Why it matters” column interprets the experiment without implying that preclinical findings have already been demonstrated clinically.
A useful way to read the evidence: carbogen studies test CO₂ inside an oxygen-rich breathing mixture, while the direct-treatment studies apply CO₂ locally or intra-arterially. Human evidence is strongest for tumor oxygenation and radiotherapy; direct antitumor effects remain preclinical.
1. Inhaled CO₂-containing gases: human and translational oncology
Carbogen is an oxygen-rich breathing mixture containing CO₂. Because it changes both gases at once, most clinical trials test the combined strategy rather than CO₂ alone. The most informative studies either vary the CO₂ fraction, directly measure tumor oxygenation, or show where the clinical benefit is concentrated.
| Study | Design / CO₂ manipulation | Key result | Why it matters |
|---|---|---|---|
| ANIMAL • MOUSE • MAMMARY TUMORS Kruuv, Inch & McCredie, 1967 — Cancer | C3H mice bearing C3HBA isotransplants/isoimplants; 20-min exposures to air, 100% O₂, 95% O₂ + 5% CO₂, or 90% O₂ + 10% CO₂. Tumor blood flow and intratumoral oxygenation were measured directly. | Relative tumor blood flow increased 20% with 5% CO₂ and 50% with 10% CO₂ versus air (P<0.01); 100% O₂ was not significantly different from air. CO₂-containing gases improved oxygenation most in anoxic tumor regions. | A clean separation experiment: oxygen remained very high while perfusion changed with the CO₂ dose, showing that oxygen alone did not reproduce the vascular response. |
| TRANSLATIONAL • MOUSE + HUMAN • PROSTATE CANCER Alonzi et al., 2009 — British Journal of Cancer | Susceptibility-weighted MRI before and during carbogen breathing in 23 DU145 xenograft mice, 17 PC3 xenograft mice and 17 men with prostate cancer. | Tumor R₂* fell 6.4% in DU145 and 5.8% in PC3 xenografts. Fourteen human tumors were evaluable; 9/14 (64%) showed significant R₂* reductions, with a mean decrease of 21.6% (P=0.0005), consistent with improved oxygenation. | Shows the same carbogen-responsive oxygenation signal in experimental prostate tumors and in a substantial subset of human prostate cancers. |
| HUMAN • COMPARATIVE • TUMOR OXYGENATION Powell et al., 1999 — Radiotherapy and Oncology | 16 patients; tumor pO₂ measured with an Eppendorf histograph during hyperoxic gas containing either 5% or 2% CO₂. | Median tumor pO₂ increased in every tumor. The fraction of measurements ≤10 mmHg fell from 47% to 29% with 5% CO₂ and from 55% to 17% with 2% CO₂. | Directly shows that CO₂ concentration within the hyperoxic mixture is a treatment variable; the lower 2% CO₂ mixture still produced strong tumor oxygenation. |
| HUMAN • PHASE III RCT • BLADDER CANCER Hoskin et al., 2010 — Journal of Clinical Oncology | 333 patients with locally advanced bladder cancer; radiotherapy alone versus radiotherapy plus carbogen and nicotinamide. | Three-year overall survival was 59% with carbogen/nicotinamide versus 46% with radiotherapy alone (P=0.04); relapse-free survival was 54% versus 43% (P=0.06). Risk of death was 14% lower with the combined strategy. | Large randomized human evidence that a CO₂-containing hypoxia-modification strategy can improve clinically meaningful outcomes when added to radiotherapy. It does not isolate CO₂ from oxygen or nicotinamide. |
| HUMAN • PHASE III RCT • LARYNGEAL CANCER Janssens et al., 2012 — Journal of Clinical Oncology | 345 patients with cT2-4 squamous-cell laryngeal cancer; accelerated radiotherapy (AR) versus accelerated radiotherapy plus carbogen and nicotinamide (ARCON), with a hypoxia-marker side study. | Five-year local control was 78% with AR versus 79% with ARCON (P=0.80); regional control was 86% versus 93% (P=0.04). In hypoxic tumors, regional control was 55% with AR versus 100% with ARCON (P=0.01); toxicity was similar. | The added benefit was concentrated where the biological target — tumor hypoxia — was present, supporting selection by tumor oxygenation rather than assuming every tumor should respond equally. |
| ANIMAL • MOUSE • NSCLC XENOGRAFT Lin et al., 2019 — BioMed Research International | A549 non-small-cell lung cancer xenografts; mice underwent room-air and 95% O₂ + 5% CO₂ carbogen conditions, with PET plus autoradiography/hypoxia staining. | In autoradiography, tumor ¹⁸F-FDG uptake was 4.19 ± 1.32%ID/g with carbogen versus 9.67 ± 4.35%ID/g with air (P<0.01). Pimonidazole binding also decreased in hypoxic regions. | Links carbogen-mediated reoxygenation to a rapid change in both tumor hypoxia and the glucose-uptake signal used clinically to image many cancers. |
| HUMAN • PHYSIOLOGY • GLIOBLASTOMA Fatigante et al., 1994 — Acta Oncologica | Six patients with histologically proven glioblastoma; arterial blood gases, inspired/expired gas concentrations and vital signs measured at baseline and during carbogen breathing. | Arterial oxygenation rose during carbogen, reaching its highest values after 10 min and remaining stable through 15 min. Nitrogen washout was incomplete in all six patients at 15 min. | A useful physiological boundary study: the internal response to an inspired gas mixture evolves over time, so actual gas exchange needs to be measured rather than inferred from the cylinder concentration alone. |
Swipe horizontally to view the complete table.
2. Direct transcutaneous CO₂ treatment of tumors
A research program at Kobe University developed a transcutaneous method in which a hydrogel is placed over the tumor and the area is exposed to CO₂. Across several tumor models, the experiments repeatedly measured changes in hypoxia signaling, mitochondria, apoptosis, growth and metastasis.
| Study | Design / CO₂ manipulation | Key result | Why it matters |
|---|---|---|---|
| ANIMAL • MOUSE • HUMAN MFH XENOGRAFT Onishi et al., 2012 — PLoS ONE | Human malignant fibrous histiocytoma (MFH) xenografts in nude mice; transcutaneous CO₂ versus control air, twice weekly for 2 weeks. | CO₂ reduced tumor volume by 48% versus control (P<0.01), increased PGC-1α and mitochondrial proliferation, and increased caspase-3/9, cytoplasmic cytochrome c and mitochondrial Bax. Body weight did not differ significantly. | Foundational direct-tumor study: slower growth occurred alongside increased mitochondrial machinery and activation of the intrinsic apoptotic pathway. |
| ANIMAL • MOUSE • OSTEOSARCOMA Harada et al., 2013 — Experimental Cell Research | 25 mice bearing highly metastatic LM8 osteosarcoma tumors; CO₂-treated (n=13) versus room-air control (n=12), 10 min twice weekly for 3 weeks. | Final tumor volume was 1,018.3 ± 177.3 mm³ with CO₂ versus 2,755.3 ± 435.9 mm³ untreated (P<0.01), about 63% lower. Pulmonary metastatic area fell to 15% of untreated levels (P<0.01), with lower HIF-1α, MMP-2 and MMP-9 and improved intratumoral hypoxia. | Connects tumor reoxygenation with both marked primary-tumor suppression and a large reduction in metastatic spread. |
| ANIMAL • MOUSE • HUMAN ORAL SCC XENOGRAFT Takeda et al., 2014 — PLoS ONE | Human oral squamous-cell carcinoma xenografts in mice; transcutaneous CO₂ versus control. | Final tumor volume was 67.9 ± 24.1 mm³ with CO₂ versus 376.1 ± 104.1 mm³ in controls (P<0.05). Lymph-node metastasis occurred in 1/8 treated animals versus 5/8 controls, alongside lower HIF-1α/VEGF/MMP signaling and greater mitochondrial apoptosis. | Replicates the antitumor and antimetastatic pattern in a different cancer type and metastatic route. |
| ANIMAL • MOUSE • THREE HUMAN XENOGRAFT MODELS Ueha et al., 2017 — Oncology Reports | Breast-cancer, osteosarcoma and MFH/UPS xenografts; treatment duration, frequency and interval were varied to optimize transcutaneous CO₂. | Apoptosis was significantly induced with exposures ≥10 min, and tumor volume significantly decreased with exposures >5 min. Twice- versus five-times-weekly treatment did not change the effect; 3- and 4-day intervals were more effective than 2- or 5-day intervals. | Demonstrates a reproducible exposure-response window and shows that treatment timing, not simply more frequent exposure, influences the antitumor effect. |
| ANIMAL • MOUSE • BREAST-CANCER BONE METASTASIS Takemori et al., 2018 — Oncology Reports | Human MDA-MB-231 breast-cancer cells implanted into mouse tibia; 100% CO₂ or control treatment twice weekly for 2 weeks. | CO₂ suppressed metastatic bone destruction and reduced HIF-1α, osteoclast-differentiation/osteolytic factors and multinucleated TRAP-positive osteoclasts. | Extends the work beyond primary-tumor volume to a clinically important consequence of metastatic disease: destruction of surrounding bone. |
| ANIMAL • MOUSE • HUMAN ORAL SCC XENOGRAFT Yatagai et al., 2021 — BioMed Research International | HSC-3 oral squamous-cell carcinoma xenografts; control, transcutaneous CO₂, cisplatin or combination treatment, twice weekly for four treatments. | CO₂-treated and combination groups showed lower PD-L1, PD-L2 and galectin-9 protein expression than control/cisplatin groups. Tumor volume also fell significantly with CO₂, cisplatin and combination treatment; the combination group showed the smallest increase. | Suggests that the CO₂-associated change in the tumor environment reaches immune-suppressive signaling as well as oxygenation and metabolism. |
| ANIMAL • MOUSE • HUMAN MFH XENOGRAFT Onishi et al., 2012 — Journal of Cancer Science & Therapy | 24 mice with human MFH xenografts; four groups (n=6 each): control, transcutaneous CO₂, doxorubicin, or CO₂ + doxorubicin, treated twice weekly for 2 weeks. | At 2 weeks, tumor volume was 3,960 ± 166 mm³ control, 2,013 ± 166 CO₂, 1,927 ± 91 doxorubicin and 1,106 ± 105 combination; the combination was significantly lower than all other groups (P<0.01). HIF-1α/VEGF signaling fell with CO₂. | Provides a direct bridge from reduced tumor hypoxia to greater chemotherapy response, with the four-arm design separating CO₂, drug and combination effects. |
Swipe horizontally to view the complete table.
3. CO₂ combined with radiation or chemotherapy
The most clinically relevant preclinical question is not only whether CO₂ changes a tumor on its own, but whether changing the tumor environment can make established treatment work better. The studies below use comparative designs that separate CO₂, conventional treatment and the combination.
| Study | Design / CO₂ manipulation | Key result | Why it matters |
|---|---|---|---|
| ANIMAL • MOUSE • HUMAN ORAL SCC XENOGRAFT Iwata et al., 2018 — Oncology Reports | Four-arm study: control, radiotherapy, transcutaneous CO₂, or the combination in human oral squamous-cell carcinoma xenografts. | At the end of the experiment, combination-group tumor volume was 23% of control, 38% of radiotherapy alone and 37% of CO₂ alone. CO₂ and radiotherapy alone each also reduced tumor volume versus control; white-cell counts fell with radiotherapy but not with CO₂ alone. | Shows a greater combined antitumor effect than either treatment alone while keeping the systemic hematologic effect of radiotherapy separate from the local CO₂ intervention. |
| ANIMAL • MOUSE • HUMAN MFH XENOGRAFT Onishi et al., 2014 — International Journal of Oncology | 24 mice randomized to control, transcutaneous CO₂, X-ray irradiation, or combination; treatments twice weekly for 2 weeks, with parallel in-vitro oxygen-condition experiments. | Combination-group tumor volume was 28% of control, 42% of CO₂ alone and 47% of X-ray alone (P<0.05). Apoptosis and ROS increased while HIF-1α decreased in the combination group. | Confirms in a second radiotherapy model that improving the hypoxic tumor environment can increase the measured effect of radiation. |
| ANIMAL • RABBIT • VX2 LIVER TUMOR Katayama et al., 2017 — International Journal of Oncology | 40 rabbits (10/group) with VX2 liver tumors; proper-hepatic-artery infusion of saline, CO₂-saturated solution, cisplatin, or the combination. | HIF-1α-positive hypoxic cells were 246 ± 118 per area in controls, 21 ± 10.4 with CO₂ (~91% lower), 152 ± 40.7 with cisplatin (~38% lower) and 16 ± 7.3 with the combination. Day-7 relative tumor volume was 210.3 ± 55.1% with combination versus 269.6 ± 45.2% with cisplatin, and was also below the expected additive value (P<0.05). | Important deep-tumor delivery experiment: intra-arterial CO₂ strongly altered the hypoxic fraction, while combining it with cisplatin produced greater tumor-growth inhibition than cisplatin alone. |
| ANIMAL • MOUSE • HUMAN ORAL SCC XENOGRAFT Yatagai et al., 2021 — BioMed Research International | Four groups: control, transcutaneous CO₂, cisplatin, or combination; tumor growth, body weight and immunosuppressive factors were compared. | Tumor volume fell significantly in the CO₂, cisplatin and combination groups, with the combination showing the smallest increase. Body weight fell significantly in the cisplatin and combination groups but not in the CO₂ group; CO₂-containing groups also showed lower immunosuppressive-factor expression. | Shows that CO₂ and cisplatin do not simply reproduce the same biological effect: CO₂ altered tumor signaling and growth without the body-weight loss seen in cisplatin-treated animals. |
Swipe horizontally to view the complete table.
4. Preserving the host without feeding the tumor
One study applied CO₂ away from the tumor and asked a different question: can skeletal muscle be protected in a cancer-bearing animal without accelerating the tumor?
| Study | Design / CO₂ manipulation | Key result | Why it matters |
|---|---|---|---|
| ANIMAL • MOUSE • ORAL SCC / MUSCLE ATROPHY Sasaki et al., 2024 — PLoS ONE | 14 tumor-bearing mice (7/group); human oral SCC implanted dorsally, with CO₂ applied to the legs twice weekly for 4 weeks rather than to the tumor. | Fat-free mass was 1.11 ± 0.15 with CO₂ versus 0.94 ± 0.10 in controls (P<0.05), and quadriceps fiber cross-sectional area was 1.75 ± 0.21 versus 1.00 ± 0.17 (P<0.01). Tumor-volume differences were not significant at any measured time point. | Shows a separate host-tissue effect: local CO₂ preserved skeletal muscle in cancer-bearing mice without a detectable increase in remote tumor growth. |
Swipe horizontally to view the complete table.
5. What the CO₂ studies themselves suggest mechanistically
The experiments do not all measure the same step, but the CO₂-specific studies repeatedly intersect at a small set of physiological and molecular changes.
Recurring experimental pattern. CO₂ delivery → improved tumor perfusion / oxygenation or reduced hypoxia markers → lower HIF-1α and related hypoxia signaling → greater mitochondrial / apoptotic activity → lower tumor growth, metastatic behavior, or greater sensitivity to conventional treatment in the studied models.
Measured changes across the program include improved tumor oxygenation; lower HIF-1α, VEGF and MMP signaling; increased PGC-1α and mitochondrial activity; cytochrome-c release and mitochondrial apoptosis; reduced metastatic spread; reduced PD-L1/PD-L2/galectin-9 protein expression in one model; and improved response to radiation or chemotherapy. No single experiment measured every link, so the full sequence is an assembled interpretation of repeated findings across separate studies.
6. The apparent paradox: tumors can already be high in CO₂
Solid tumors are not necessarily CO₂-poor. Measurements cited in the cancer chapter place tumor PCO₂ around 59–84 mmHg in some settings, above normal arterial values. The hypothesis therefore cannot simply be “tumors lack CO₂, therefore add CO₂.”
The distinction suggested by the treatment studies is between stagnant metabolic CO₂ accumulating in poorly perfused tissue and deliberately delivered CO₂ that changes local perfusion, oxygen unloading and hypoxic signaling. The direct-treatment experiments are intriguing because already-hypoxic tumors became less hypoxic after CO₂ exposure, with HIF-1α falling rather than rising. Route, timing and tissue state therefore appear to matter.
7. What has and has not been demonstrated
The clearest way to read the evidence is to separate what has been shown in people from what remains preclinical and from the clinical question that is still open.
| Study | Design / CO₂ manipulation | Key result | Why it matters |
|---|---|---|---|
| HUMAN EVIDENCE Carbogen / radiotherapy | Human tumor studies show that CO₂-containing carbogen can increase tumor oxygenation. Large randomized radiotherapy trials in bladder and laryngeal cancer have tested carbogen plus nicotinamide as a hypoxia-modification strategy. | The clinical evidence supports CO₂-containing gas as part of an adjunctive oxygenation strategy; it does not establish CO₂ alone as an anticancer treatment. | |
| PRECLINICAL EVIDENCE Multiple animal tumor models | Direct/local CO₂ has reduced tumor hypoxia or HIF-1α, increased mitochondrial/apoptotic signaling, reduced growth and metastasis in several models, altered immune-suppressive markers, and enhanced radiation or chemotherapy in comparative experiments. | This is the strongest evidence for a direct antitumor role, but it remains preclinical and comes largely from one research program. | |
| UNANSWERED CLINICAL QUESTION Direct/local CO₂ in human cancer | No clinical trial has shown that transcutaneous CO₂ shrinks established human tumors or improves cancer survival as a standalone treatment. Optimal route, dose, schedule, tumor selection and combination strategy remain unresolved. | The preclinical findings justify further testing; they do not justify replacing established oncology treatment. |
Swipe horizontally to view the complete table.
8. The strongest defensible reading
Current evidence-based interpretation. CO₂ can modify the oxygenation and signaling state of tumors, and in several preclinical models that change is followed by mitochondrial apoptosis, slower tumor growth, less metastasis, or greater sensitivity to conventional treatment. In humans, the clearest evidence is not standalone tumor treatment but carbogen-assisted modification of tumor oxygenation during radiotherapy. The unresolved question is whether the direct local effects repeatedly seen in animal tumors can be reproduced safely and usefully in human cancer.
References
Full citations for the studies in this focused guide, numbered independently from the book chapter. DOI links are intentionally omitted from this working version.
1. Kruuv JA, Inch WR, McCredie JA. Blood flow and oxygenation of tumors in mice. I. Effects of breathing gases containing carbon dioxide at atmospheric pressure. Cancer. 1967;20(1):51–59.
2. Alonzi R, Padhani AR, Maxwell RJ, et al. Carbogen breathing increases prostate cancer oxygenation: a translational MRI study in murine xenografts and humans. British Journal of Cancer. 2009;100(4):644–648.
3. Powell MEB, Collingridge DR, Saunders MI, et al. Improvement in human tumour oxygenation with carbogen of varying carbon dioxide concentrations. Radiotherapy and Oncology. 1999;50(2):167–171.
4. Hoskin PJ, Rojas AM, Bentzen SM, et al. Radiotherapy with concurrent carbogen and nicotinamide in bladder carcinoma. Journal of Clinical Oncology. 2010;28(33):4912–4918.
5. Janssens GO, Rademakers SE, Terhaard CH, et al. Accelerated radiotherapy with carbogen and nicotinamide for laryngeal cancer: results of a phase III randomized trial. Journal of Clinical Oncology. 2012;30(15):1777–1783.
6. Lin Y, Bao YN, Huang CX, et al. The effect of carbogen breathing on 18F-FDG uptake in non-small-cell lung cancer. BioMed Research International. 2019;2019:2920169.
7. Fatigante L, Cartei F, Ducci F, et al. Carbogen breathing in patients with glioblastoma multiforme submitted to radiotherapy: assessment of gas exchange parameters. Acta Oncologica. 1994;33(7):807–811.
8. Onishi Y, Kawamoto T, Ueha T, et al. Transcutaneous application of carbon dioxide (CO₂) induces mitochondrial apoptosis in human malignant fibrous histiocytoma in vivo. PLoS ONE. 2012;7(11):e49189.
9. Harada R, Kawamoto T, Ueha T, et al. Reoxygenation using a novel CO₂ therapy decreases the metastatic potential of osteosarcoma cells. Experimental Cell Research. 2013;319(13):1988–1997.
10. Takeda D, Hasegawa T, Ueha T, et al. Transcutaneous carbon dioxide induces mitochondrial apoptosis and suppresses metastasis of oral squamous cell carcinoma in vivo. PLoS ONE. 2014;9(7):e100530.
11. Ueha T, Kawamoto T, Onishi Y, et al. Optimization of antitumor treatment conditions for transcutaneous CO₂ application: an in vivo study. Oncology Reports. 2017;37(6):3688–3694.
12. Takemori T, Kawamoto T, Ueha T, et al. Transcutaneous carbon dioxide application suppresses bone destruction caused by breast cancer metastasis. Oncology Reports. 2018;40(4):2079–2087.
13. Yatagai N, Hasegawa T, Amano R, et al. Transcutaneous carbon dioxide decreases immunosuppressive factors in squamous cell carcinoma in vivo. BioMed Research International. 2021;2021:5568428.
14. Onishi Y, Kawamoto T, Ueha T, et al. Transcutaneous application of carbon dioxide (CO₂) enhances chemosensitivity by reducing hypoxic conditions in human malignant fibrous histiocytoma. Journal of Cancer Science & Therapy. 2012;4(7):174–181.
15. Iwata E, Hasegawa T, Ueha T, et al. Transcutaneous carbon dioxide enhances the antitumor effect of radiotherapy on oral squamous cell carcinoma. Oncology Reports. 2018;40(1):434–442.
16. Onishi Y, Akisue T, Kawamoto T, et al. Transcutaneous application of CO₂ enhances the antitumor effect of radiation therapy in human malignant fibrous histiocytoma. International Journal of Oncology. 2014;45(2):732–738.
17. Katayama N, Sugimoto K, Okada T, et al. Intra-arterially infused carbon dioxide-saturated solution for sensitizing the anticancer effect of cisplatin in a rabbit VX2 liver tumor model. International Journal of Oncology. 2017;51(2):695–701.
18. Sasaki A, Takeda D, Kawai H, et al. Transcutaneous carbon dioxide suppresses skeletal muscle atrophy in a mouse model of oral squamous cell carcinoma. PLoS ONE. 2024;19(4):e0302194.
