Den første videnskabelige artikel om Rehaler-behandlingen blev i august 2018 udgivet i Cephalalgia som Open Access:

Øvrig information kan findes i Ph.D.-afhandlingen "Pulmonary gas exchange and blood gas tensions: new frontiers in imaging, diagnosis and treatment." (Johansen 2017) (se bl.a. nedenstående uddrag).

Du er som professionel desuden meget velkommen til at kontakte forskningschef Troels Johansen på for mere information

The rationale for using a partial rebreathing device (PRD) in migraine is threefold:

  1. To utilize CO2’s ability to reduce the nerve system’s excitability, in particular in order to counteract the migraine trigger phenomenon Cortical Spreading Depression (CSD) which has been shown to be remarkably sensitive to changes in CO2 and pH.  
  2. To utilize the uniquely strong vasodilatory properties of CO2 in order to increase blood flow and oxygen delivery to the brain early in the evolution of the attack, to counteract the impaired cerebral blood flow characteristic of the triggering and early stages of migraine attacks. Spikes in the extracellular K+ concentration are known to play an important role in the triggering of CSD (Pietrobon, Moskowitz, 2013); raising CBF increases oxygen and glucose delivery to the brain, in turn bolstering Na+/K+-ATPase activity and thereby decreasing the likelihood of such spikes.
  3. Results from clinical studies of CO2 as a treatment for migraine. Several clinical studies have shown a remarkable efficacy of inhaled CO2 in preventing and aborting attacks, but the methods of administering CO2 used have been problematic, some studies using bulky and expensive pressure cylinders and other studies using potentially dangerous closed rebreathing bags, i.e. with no supply of fresh air or oxygen to the patient while rebreathing.

In the following, the above three points will be addressed in turn.

1.1. CO 2  acidosis as an inhibitor of neuronal excitability  

As a general rule, raising CO2 and lowering pH depress the excitability of neurons, by the following mechanisms (Somjen, Tombaugh, 1998):

  • Increase of resting membrane potential 
  • Increase of firing threshold
  • Decrease of impulse conduction velocity 

It is likely that these phenomena play a role in the analgesic effects shown to result from raising CO2 (Marcussen, Wolff, 1950, Sikh, Agarwal, 1974, Dexter, 1982, Pradalier et al., 1984, Spierings, 2005a, Vause et al., 2007, Strider, Garrett & Durham, 2009, Tzabazis et al., 2010, Miller, 2011).

Not all mechanisms related to the neuronal effects of pH and CO2 have been elucidated, though it has been proposed that CO2-induced changes in neuronal functionality and excitability arise from a pH-dependent modulation of adenosine and ATP levels (Dulla et al., 2005), and it has been shown that the acid-sensing ion channel ASIC1a regulates neuronal excitability by means of a large sensitivity to extracellular pH (Miller, 2011, Ziemann et al., 2008).

It is of particular relevance for migraine that pH and CO2 have been shown to have a marked effect on the triggering threshold, propagation speed and duration of Cortical Spreading Depression (CSD). CSD is a transient, propagating wave of near-total depolarization, moving over the cerebral cortex at a speed of approximately 3 mm/min (see Figure 1). 


Figure 1: Progression of Cortical Spreading Depression (CSD). Arrows represent the direction of the moving CSD wave (the red area). The blue areas represent the brain areas which the CSD wave has passed, now having a markedly reduced perfusion. Figure from  (Olesen et al., 2009) .

CSD has been implicated as a very likely trigger for migraine attacks (Kelman, 2011, Ward, 2012, Charles, 2013), and has been convincingly shown to be the underlying phenomenon behind visual migraine auras (Lauritzen, 1994).  Following CSD, a large energy expenditure is required to re-establish normal electrochemical gradients of Na+, K+, Cl- and Ca2+, leading to increased oxygen consumption (Ayata, Lauritzen, 2015). This is accompanied by a CSD-effected decoupling of perfusion and oxygen consumption, resulting in a marked hypo-perfusion and oxygen desaturation for up to sixty minutes in the wake of CSD (Brennan et al., 2009), indicated by the blue coloration in Figure 1.

It has been shown that by lowering pH (either by raising PCO₂ or lowering the bicarbonate concentration) a marked inhibitory effect is achieved on the triggering, propagation and duration of CSD (Gardner-Medwin, 1981, Tombaugh, 1994, Tong, Chesler, 2000). 

In the field of epilepsy, a large number of in vitro, animal and human studies have shown that it is possible to inhibit and abort epileptiform activity in the brain by increasing the inspired CO2 level (Lennox, Gibbs & Gibbs, 1938, Mitchell, Grubbs, 1956, Fried, Fox & Carlton, 1990, Tolner et al., 2011, Ohmori et al., 2013).  

In the light of the studies cited, there are strong electrophysiological arguments for raising CO2 in migraine, particularly in the early part of attacks when CSD is still progressing.

1.2. Cerebral blood flow, migraine and CO 2

Despite advances in imaging and other diagnostic techniques, the physiological “smoking gun” triggering migraine has not yet been found, though recently the correlation with CSD has been established. Even so, brain imaging studies have consistently shown that the early stages of migraine attacks are characterized by uniform or localized cerebral vasoconstriction and ischemia (Charles, 2013, Olesen et al., 1990, De Benedittis, 1999, Sanchez Del Rio et al., 1999, Calandre et al., 2002, Denuelle et al., 2008, Hansen et al., 2011, Koreshkina et al., 2013). On this basis, hypoperfusion has been implicated as a migraine trigger (Kelman, 2011, Brennan, Charles, 2010), possibly because of local ischemia triggering CSD (Ayata, Lauritzen, 2015).

At baseline (i.e. between attacks), there are indications of either a lower cerebral blood flow (CBF) among migraine patients compared to controls (De Benedittis, 1999, Calandre et al., 2002), a higher CBF (Loehrer et al., 2015)  or a more heterogeneous picture of brain areas of both hypo –and hyperperfusion at baseline (Arkink et al., 2012). 

Thirty years ago, the most widespread theory of migraine was that the transition from the aura or prodromal phase to the pain phase coincided with transition from cerebral vasoconstriction to vasodilation (De Benedittis, 1999). However, a famous study by Olesen and colleagues (Olesen et al., 1990) showed that hypoperfusion is often present more than an hour into, or even throughout, the pain phase – a finding which has been reported by several studies since then (Sanchez Del Rio et al., 1999, Denuelle et al., 2008, Bednarczyk et al., 1998). Furthermore, it has repeatedly been shown that cerebral vasodilation in itself does not cause pain (Tfelt-Hansen, 2010) .

Schematically, the evolution of cerebral perfusion in migraine attacks may be represented as shown in Figure 2 (the dashed red lines representing the more variable or doubtful perfusion-patterns).


Figure 2: Cerebral perfusion timeline in migraine attacks, based on a synthesis of the available evidence. Dashed lines represent variable or uncertain perfusion patterns.  

From the literature, it seems evident that hypoperfusion more than hyperperfusion is the characteristic finding in migraine, hyperperfusion seemingly being an epi-phenomenon to the actual pain (Charles, 2013). Still, it is very important to note that CBF may be heterogeneous, and even a normal total CBF may not preclude the existence of localized ischemic regions that could be responsible for triggering migraine (Brennan, Charles, 2010, Friberg et al., 1987, Young, Van Vliet, 1992, Tamura et al., 2011).  

It has been known for more than 50 years that CO2 is the most efficient cerebral vasodilator in existence (Reivich, 1964, Olesen, Paulson & Lassen, 1971, Madden, 1993), an evolutionary adaption to ensure adequate blood and oxygen supply to the brain when the metabolic demands (and thus CO2 production) are highest. The response curve of CBF to CO2 is sigmoidal and approximately shaped as shown in Figure 3 (Data from (Claassen et al., 2007)).


FiForProfessionalsPage Cerebral blood flow (normalized by baseline CBF) as a function of arterial   CO 2   tension. Solid line represents adults without lung disease. Data from (Claassen et al., 2007).

As indicated by Figure 3, the response curve is approximately linear at small PaCO₂ changes from baseline. One study of 45 individuals (not differentiated in terms of migraine) (Pollock et al., 2009) found that the slopes for the CBF changes were approximately:

Men: an increase in %CBF of 7.5% for each increase of 1 mmHg PaCO₂

Women:  an increase in %CBF of 6.3% for each increase of 1 mmHg PaCO₂

Reversely, a decrease in the arterial oxygen tension (PaO2) will also increase CBF, but only at quite low PaO2 levels.

For migraine patients in particular, several studies have examined the vascular responsiveness of CBF to breath hold tests (a surrogate of PaCO₂), though both increased and decreased response curves (compared to healthy controls) have been found with this technique (Dora, Balkan, 2002, Akin, Bilensoy, 2006, Chan et al., 2009). 

1.3. CO 2 -treatment of headache: past studies

A literature search (inter alia in Embase and PubMed) identified five studies reporting data on the treatment of headache by use of CO2 (Marcussen, Wolff, 1950, Sikh, Agarwal, 1974, Dexter, 1982, Pradalier et al., 1984, Spierings, 2005a, Spierings, 2005b). Four of these concerned migraine pain and one post-spinal headache. The studies are summarized below:

Study 1:  Effects of carbon dioxide-oxygen mixtures given during preheadache phase of the migraine attack; further analysis of the pain mechanisms in headache (Marcussen, Wolff, 1950).


The authors administered two different gases to patients: A) 10% CO2 in air and B) 10% CO2 in O2. The gases were randomized. The patients were in different stages of migraine attack: 1) the pre-pain phase (characterized by what would today be called aura symptoms) 2) a phase where aura symptoms overlapped with headache and 3) a phase in which only headache was present. 

The gases were administered by face mask between one and three times, for (only) 5 minutes at a time, followed by 5-15 minutes of rest (necessary due to the strong hyperpnea induced by the high inspired CO2 level). 

In the five cases in which the gas was administered in the pre-pain phase, CO2-in-O2 completely abolished the aura symptoms and the expected headache did not occur. CO2-in-air resulted in the same clearing of aura symptoms, though they returned approx. five minutes after the gas was discontinued (possibly due to a rebound effect in the form of a hypocapnic “undershoot”). 

In cases where headache was already present, the authors report that “the results were so unpredictable that no conclusions could be drawn”. 

Study 2: Post spinal headache. A preliminary report on the effect of inhaled carbon dioxide (Sikh, Agarwal, 1974).


The study does not concern migraine but is nonetheless relevant to the subject at hand. The authors treated post-spinal-block headache (i.e. headache presumably due to low CNS pressure) by administering one of two gases: A) 5.6% CO2 in air (40 patients) or B) pure oxygen (12 patients). The gases were administered for 10 minutes once every 24 hours. In spite of this very short duration of treatment, the difference between the two groups was significant: 98% of the CO2 group was relieved of pain after three days, while only 58% of the O2 group was relieved (though a considerable number in each group is likely to have been due to spontaneous improvement). 

Study 3:  Rebreathing aborts migraine attacks (Dexter, 1982).


The author reports in brief form on approx. 30 rebreathing experiments on six patients (two with “classic” (i.e. aura) migraine and four with “common” (i.e. non-aura) migraine). Rebreathing was done with a simple polyethylene plastic bag, from which the patient breathed as long as possible “with only a small intake of fresh air”. 

The two aura patients started rebreathing during the aura phase and were able to avoid progression to headache in all or most of their attacks (it is not clear which).  These patients rebreathed between 20 and 30 minutes, one experiencing dyspnea afterwards and the other lapsing into unconsciousness during one rebreathing treatment, possibly due to the very high CO2 levels attained with this type of (potentially dangerous) closed-circuit rebreathing technique.   

For the non-aura patients, it is not completely clear in which phase the rebreathing started (pain or pre-pain). Nonetheless, 12 out of 21 attacks were stopped in these patients.  In one patient, all four attacks were stopped, while in another only two out of six attacks were stopped – indicating a possible difference in efficacy between attack types.  

Study 4:  Trial treatment of migraine attack by rebreathing of expired air (Pradalier et al., 1984).


The authors used rebreathing from a plastic bag (apparently no intake of fresh air at all), in 50 patients having a total of 113 attacks. The rebreathing was performed for between five and 15 minutes, 15 of the patients being unable to rebreathe more than five minutes due to symptoms of CO2 acidosis (feelings of suffocating, dizziness or nausea). The article seems to indicate that CO2-treatment was applied only in cases where the pain phase of the migraine attack had started.

In 26% of the attacks, the treatment was effective in relieving the pain. 

Study 5:

  • Non-inhaled, intranasal carbon dioxide for the abortive treatment of migraine headache: efficacy, tolerability and safety (Spierings, 2005a)
  • Abortive treatment of migraine headache with non-inhaled, intranasal carbon dioxide: a randomized, double-blind, placebo-controlled, parallel-group study (Spierings, 2005b)


This study used non-inhaled carbon dioxide in the nasal cavity as a way to treat fully developed migraine headache, based on the theory that CO2 can deactivate the trigeminal system by hyperpolarizing its nerve fibres (Tzabazis et al., 2010) – though this physiological interpretation has been challenged (Jürgens, Reetz & May, 2013). 152 patients with migraine headache where treated for approximately 19 minutes distributed over two hours, with either CO2 or a dummy device (Spierings, 2005b).

The 2-hour efficacy in abolishing headache was 36.4% in the CO2 group and 10.0% in the dummy group.

1.4. PRD as migraine treatment

Based on the data available, it seemed possible that increasing inspired CO2 before the pain phase of the attack could be an effective way of preventing attacks. The studies reporting data on this approach (Marcussen, Wolff, 1950, Dexter, 1982) used very high inspired CO2-tensions, for which reason the treatment was necessarily aborted after a short while. It is remarkable that even very short treatment durations had as great an effect as reported by Marcussen et al. in preventing the headache phase. 

Marcussen reported a better effect of the CO2-oxygen mixture than the CO2-air mixture, a finding which is of relevance for the discussion of using a partial rebreathing device. It is clear that a partial rebreathing device does not give as high inspired oxygen fractions as a gas mixture of 10% CO2 in oxygen (i.e. an inspired O2 fraction (FIO2) of 90%, compared to 16-20% with the PRD depending on the device setting), but the device does allow for much longer, even continuous, use, which may make up for this deficiency. It is indeed possible to construct a device with a higher FiO2 (e.g. using gas bottles) but this would necessarily be a more complicated, costly and cumbersome device, likely not practical and/or affordable for many patients. 

Importantly, such high FiO2 levels may not be necessary. From the point of view of increasing oxygen delivery to the brain (brain oxygen delivery = cerebral blood flow multiplied by arterial oxygen concentration (  = CBF x CaO2)), a partial rebreathing device employs a trade-off between 

A)    increased CBF due to the PaCO₂ increase 

B)    decreased arterial oxygen concentration (CaO2) due to the nature of rebreathing,  

C)    Bohr’s effect modulating the O2-Hb affinity, and in extreme cases: 

D)    CBF increases induced by low PaO2

Since the curve of  CaO2 as a function of PaO2 is sigmoidal (Johansen, 2017), a decrease in inspired oxygen from 149 mmHg (sea level) to 120 mmHg or even 90 mmHg will have only a small effect on CaO2.

On the other hand, the increase in PaCO₂ associated with such a drop in PaO2 will lead to a very large increase in CBF, as discussed in (Johansen, 2017).When computing the theoretical change in  during PRD use, we thus multiply a highly increased CBF with an only slightly decreased CaO2, yielding a significantly increased  , as illustrated in (Johansen, 2017).

With the PRD at a high-rebreathing setting in a 34-year-old male test person, we measured a PaCO₂ increase from 40.5 to 50.3 mmHg, with only a very small drop in CaO2 from 20.6 ml/dl to 20.3 ml/dl (all parameters measured by arterial blood gas samples). Based on the aforementioned CBF vs CO2 equation (Pollock et al., 2009), this would equate to an increase in   of more than 70% with the PRD compared to the baseline. If using carbogen (with a CO2 content producing the same increase in PaCO₂ as in the PRD example) the increase in   would be approx. 86%. In the experiments reported by Marcussen and Wolff using 10% CO2 in oxygen, the PaCO₂ is likely to have been significantly above 50 mmHg though, and possibly above the point at which the CBF response curve flattens out (approx. 60 mmHg, see Figure 3). 

As shown in (Johansen, 2017)   is at very high rebreathing ratios likely to eventually start falling due to the decrease in CaO2, but based on our experiences with the PRD, symptoms of respiratory acidosis would have made the user terminate the device use before that happens (which would be the case for carbogen as well if the PaCO₂ levels were equal). 

The pure-rebreathing method (employed by Pradalier and to some extent Dexter) seems very ill-advised and dangerous, since a steady state is not achieved and   will fall precipitously after an initial spike. 

The above calculations are predicated on the assumption that part of the mechanism by which CO2 prevents migraine attacks is by increasing   globally and locally during the pre-pain phase, preventing local ischemia that may trigger CSD and migraine. It seems likely that such a 70% increase in overall   (as calculated above) will serve to more than counteract any local ischemia present in the early parts of migraine attacks.   

Based on (Dexter, 1982, Pradalier et al., 1984, Spierings, 2005a) it seems likely that even when applied after the beginning of the pain phase, the device will be still able to relieve headache for a subset of patients and attacks. Based on (among others) Spierings’ results, it seems likely that this effect is at least partially mediated by modulation of nerve sensitivity, and not solely by vascular tone. 

Even so, we concluded that the best treatment option in a clinical trial would be to start the PRD use as early as possible in the evolution of the migraine attack, in order to counteract CSD and hypoxemia.  


Akin, A. & Bilensoy, D. 2006, "Cerebrovascular reactivity to hypercapnia in migraine patients measured with near-infrared spectroscopy", Brain research, vol. 1107, no. 1, pp. 206-214. 

Arkink, E.B., Bleeker, E.J.W., Schmitz, N., Schoonman, G.G., Wu, O., Ferrari, M.D., Van Buchem, M.A., Van Osch, M.J.P. & Kruit, M.C. 2012, "Cerebral perfusion changes in migraineurs: A voxelwise comparison of interictal dynamic susceptibility contrast MRI measurements", Cephalalgia, vol. 32, no. 4, pp. 279-288. 

Ayata, C. & Lauritzen, M. 2015, "Spreading depression, spreading depolarizations, and the cerebral vasculature", Physiological Reviews, vol. 95, no. 3, pp. 953-993. 

Bednarczyk, E.M., Remler, B., Weikart, C., Nelson, A.D. & Reed, R.C. 1998, "Global cerebral blood flow, blood volume, and oxygen metabolism in patients with migraine headache", Neurology, vol. 50, no. 6, pp. 1736-1740. 

Brennan, K.C., Chang, J.C., Shook, L.L., Biag, J.D., Toga, A.C., Arthur, W. & Charles, A.C. 2009, "Acute and long term hemoglobin desaturation during cortical spreading depression; A pincer effect of metabolic demand and deranged neurovascular coupling", Annals of Neurology, vol. 66, pp. S76. 

Brennan, K.C. & Charles, A. 2010, "An update on the blood vessel in migraine", Current opinion in neurology, vol. 23, no. 3, pp. 266-274. 

Calandre, E.P., Bembibre, J., Arnedo, M.L. & Becerra, D. 2002, "Cognitive disturbances and regional cerebral blood flow abnormalities in migraine patients: Their relationship with the clinical manifestations of the illness", Cephalalgia, vol. 22, no. 4, pp. 291-302. 

Chan, S.-., Tam, Y., Lai, C.-., Wu, H.-., Lam, Y.-., Wong, P.-. & Kwong, K.K. 2009, "Transcranial Doppler study of cerebrovascular reactivity: Are migraineurs more sensitive to breath-hold challenge?", Brain research, vol. 1291, pp. 53-59. 

Charles, A. 2013, "The evolution of a migraine attack - A review of recent evidence", Headache, vol. 53, no. 2, pp. 413-419. 

Claassen, J.A., Zhang, R., Fu, Q., Witkowski, S. & Levine, B.D. 2007, "Transcranial Doppler estimation of cerebral blood flow and cerebrovascular conductance during modified rebreathing", Journal of applied physiology (Bethesda, Md.: 1985), vol. 102, no. 3, pp. 870-877. 

De Benedittis, G. 1999, "CBF changes during headache-free periods and spontaneous/induced attacks in migraine with and without aura: A TCD and SPECT comparison study", Journal of neurosurgical sciences, vol. 43, no. 2, pp. 141-147. 

Denuelle, M., Fabre, N., Payoux, P., Chollet, F. & Geraud, G. 2008, "Posterior cerebral hypoperfusion in migraine without aura", Cephalalgia, vol. 28, no. 8, pp. 856-862. 

Dexter, S.L. 1982, "Rebreathing aborts migraine attacks", British medical journal, vol. 284, no. 6312, pp. 312. 

Dora, B. & Balkan, S. 2002, "Exaggerated interictal cerebrovascular reactivity but normal blood flow velocities in migraine without aura", Cephalalgia, vol. 22, no. 4, pp. 288-290. 

Dulla, C.G., Dobelis, P., Pearson, T., Frenguelli, B.G., Staley, K.J. & Masino, S.A. 2005, "Adenosine and ATP link PCO₂ to cortical excitability via pH", Neuron, vol. 48, no. 6, pp. 1011-1023. 

Friberg, L., Olsen, T.S., Roland, P.E. & Lassen, N.A. 1987, "Focal ischaemia caused by instability of cerebrovascular tone during attacks of hemiplegic migraine. A regional cerebral blood flow study", Brain, vol. 110, no. 4, pp. 917-934. 

Fried, R., Fox, M.C. & Carlton, R.M. 1990, "Effect of diaphragmatic respiration with end-tidal CO₂ biofeedback on respiration, EEG, and seizure frequency in idiopathic epilepsy", Annals of the New York Academy of Sciences, vol. 602, pp. 67-96. 

Gardner-Medwin, A.R. 1981, "Possible roles of vertebrate neuroglia in potassium dynamics, spreading depression and migraine.", The Journal of experimental biology, vol. 95, pp. 111-127. 

Hansen, J.M., Schytz, H.W., Larsen, V.A., Iversen, H.K. & Ashina, M. 2011, "Hemiplegic migraine aura begins with cerebral hypoperfusion: Imaging in the acute phase", Headache, vol. 51, no. 8, pp. 1289-1296. 

Johansen, T. 2017, Pulmonary gas exchange and blood gas tensions: new frontiers in imaging, diagnosis and treatment. PhD thesis., Department of Clinical Medicine, Aarhus University. 

Jürgens, T.P., Reetz, R. & May, A. 2013, "No relevant modulation of TRPV1-mediated trigeminal pain by intranasal carbon dioxide in healthy humans.", The journal of headache and pain, vol. 14, no. 1, pp. 33. 

Kelman, L. 2011, "The biological basis of headache", Expert Review of Neurotherapeutics, vol. 11, no. 3, pp. 363-378. 

Koreshkina, I., Khalikov, D., Nazinkina, V., Amelin, V. & Osipova, V. 2013, "Brain perfusion permutations during migraine attacks without aura", Journal of Headache and Pain, vol. 14, no. 1. 

Lauritzen, M. 1994, "Pathophysiology of the migraine aura: The spreading depression theory", Brain, vol. 117, no. 1, pp. 199-210. 

Lennox, W.G., Gibbs, F.A. & Gibbs, E.I. 1938, "The Relationship in Man of Cerebral Activity to Blood Flow and to Blood Constituents", Journal of neurology and psychiatry, vol. 1, no. 3, pp. 211-225. 

Loehrer, E., Vernooij, M.W., Van Der Lugt, A., Hofman, A. & Ikram, M.A. 2015, "Migraine and cerebral blood flow in the general population", Cephalalgia, vol. 35, no. 2, pp. 190-198. 

Madden, J.A. 1993, "The effect of carbon dioxide on cerebral arteries", Pharmacology and Therapeutics, vol. 59, no. 2, pp. 229-250. 

Marcussen, R.M. & Wolff, H.G. 1950, "Effects of carbon dioxide-oxygen mixtures given during preheadache phase of the migraine attack; further analysis of the pain mechanisms in headache.", Archives of neurology and psychiatry, vol. 63, no. 1, pp. 42-51. 

Miller, J.W. 2011, "Stopping seizures with carbon dioxide", Epilepsy Currents, vol. 11, no. 4, pp. 114-115. 

Mitchell, W.G. & Grubbs, R.C. 1956, "Inhibition of audiogenic seizures by carbon dioxide", Science (New York, N.Y.), vol. 123, no. 3189, pp. 223-224. 

Ohmori, I., Hayashi, K., Wang, H., Ouchida, M., Fujita, N., Inoue, T., Michiue, H., Nishiki, T. & Matsui, H. 2013, "Inhalation of 10% carbon dioxide rapidly terminates Scn1a mutation-related hyperthermia-induced seizures", Epilepsy research, vol. 105, no. 1-2, pp. 220-224. 

Olesen, J., Burstein, R., Ashina, M. & Tfelt-Hansen, P. 2009, "Origin of pain in migraine: evidence for peripheral sensitisation", The Lancet Neurology, vol. 8, no. 7, pp. 679-690. 

Olesen, J., Friberg, L., Skyhoj Olsen, T., Iversen, H.K., Lassen, N.A., Andersen, A.R. & Karle, A. 1990, "Timing and topography of cerebral blood flow, aura, and headache during migraine attacks", Annals of Neurology, vol. 28, no. 6, pp. 791-798. 

Olesen, J., Paulson, O.B. & Lassen, N.A. 1971, "Regional cerebral blood flow in man determined by the initial slope of the clearance of intra-arterially injected 133Xe.", Stroke; a journal of cerebral circulation, vol. 2, no. 6, pp. 519-540. 

Pietrobon, D. & Moskowitz, M.A. 2013, "Pathophysiology of Migraine", Annual Review of Physiology, Vol 75, vol. 75, pp. 365-391. 

Pollock, J.M., Deibler, A.R., Whitlow, C.T., Tan, H., Kraft, R.A., Burdette, J.H. & Maldjian, J.A. 2009, "Hypercapnia-induced cerebral hyperperfusion: An underrecognized clinical entity", American Journal of Neuroradiology, vol. 30, no. 2, pp. 378-385. 

Pradalier, A., Baron, J.F., Dry, J. & Launay, J.M. 1984, "Trial treatment of migraine attack by rebreathing of expired air", Presse médicale (Paris, France : 1983), vol. 13, no. 31, pp. 1901. 

Reivich, M. 1964, "Arterial PCO₂ and cerebral hemodynamics", The American Journal of Physiology, vol. 206, pp. 25-35. 

Sanchez Del Rio, M., Bakker, D., Wu, O., Agosti, R., Mitsikostas, D.D., Østergaard, L., Wells, W.A., Rosen, B.R., Sorensen, G., Moskowitz, M.A. & Cutrer, F.M. 1999, "Perfusion weighted imaging during migraine: Spontaneous visual aura and headache", Cephalalgia, vol. 19, no. 8, pp. 701-707. 

Sikh, S.S. & Agarwal, G. 1974, "Post spinal headache. A preliminary report on the effect of inhaled carbon dioxide", Anaesthesia, vol. 29, no. 3, pp. 297-300. 

Somjen, G.G. & Tombaugh, G.C. 1998, "pH modulation of neuronal excitability and central nervous system functions" in pH and Brain Function, eds. K. Kaila & B.R. Ransom, 1st edn, Wiley-Liss, , pp. 373-393. 

Spierings, E. 2005a, Non-inhaled, intranasal carbon dioxide for the abortive treatment of migraine headache: efficacy, tolerability and safety, San Diego edn, San Diego. 

Spierings, E. 2005b, "Abortive treatment of migraine headache with non-inhaled, intranasal carbon dioxide: a randomized, double-blind, placebo-controlled, parallel-group study", Headache, vol. 45, pp. 809. 

Strider, J.W., Garrett, F.G. & Durham, P.L. 2009, "Non-inhaled administration of 100% carbon dioxide represses capsaicin mediated activation of neurons and glia in the trigeminal nucleus caudalis", Cephalalgia, vol. 29, pp. 20-21. 

Tamura, N., Araki, N., Yamamoto, T., Nakazato, Y. & Itokawa, K. 2011, "German “sympathicovascular theory” for the pathogenesis of migraine", Cephalalgia, vol. 31, pp. 203. 

Tfelt-Hansen, P. 2010, "Intracranial vasodilation in migraine?", Annals of Neurology, vol. 67, no. 5, pp. 695-696. 

Tolner, E.A., Hochman, D.W., Hassinen, P., Otáhal, J., Gaily, E., Haglund, M.M., Kubová, H., Schuchmann, S., Vanhatalo, S. & Kaila, K. 2011, "Five percent CO₂ is a potent, fast-acting inhalation anticonvulsant", Epilepsia, vol. 52, no. 1, pp. 104-114. 

Tombaugh, G.C. 1994, "Mild acidosis delays hypoxic spreading depression and improves neuronal recovery in hippocampal slices", Journal of Neuroscience, vol. 14, no. 9, pp. 5635-5643. 

Tong, C.K. & Chesler, M. 2000, "Modulation of spreading depression by changes in extracellular pH", Journal of neurophysiology, vol. 84, no. 5, pp. 2449-2457. 

Tzabazis, A.Z., Niv, S.H., Manering, N.A., Klyukinov, M., Cuellar, J.M., Bhatnagar, A. & Yeomans, D.C. 2010, "Trigeminal antihyperalgesic effect of intranasal carbon dioxide", Life Sciences, vol. 87, no. 1-2, pp. 36-41. 

Vause, C., Bowen, E., Spierings, E. & Durham, P. 2007, "Effect of carbon dioxide on calcitonin gene-related peptide secretion from trigeminal neurons", Headache, vol. 47, no. 10, pp. 1385-1397. 

Ward, T.N. 2012, "Migraine diagnosis and pathophysiology", CONTINUUM Lifelong Learning in Neurology, vol. 18, no. 4, pp. 753-763. 

Young, D.B. & Van Vliet, B.N. 1992, "Migraine with aura: A vicious cycle perpetuated by potassium-induced vasoconstriction", Headache, vol. 32, no. 1, pp. 24-34. 

Ziemann, A.E., Schnizler, M.K., Albert, G.W., Severson, M.A., Howard III, M.A., Welsh, M.J. & Wemmie, J.A. 2008, "Seizure termination by acidosis depends on ASIC1a", Nature neuroscience, vol. 11, no. 7, pp. 816-822.