Critique of ‘Control of arterial PCO2 by somatic afferents’

In a recent study published in The Journal of Physiology, Haouzi & Chenuel (2005) have rekindled an age-old debate about the exercise hyperpnoea conundrum by counterposing two major groups of competing hypotheses. According to the authors, ‘The first group of mechanisms suggests that no specific exercise stimuli are required to explain the – coupling during ERC... Such a system, optimized with information on the mechanical properties of the respiratory system, has even[emphasis added] been suggested to account for the entire response, not only to ERC but also to a voluntary exercise (Poon, 1987) (the limits of this hypothesis are presented in the Discussion).’ While we are flattered by the authors' generosity in singling out Poon (1987) as the pinnacle of the first group of hypothesized mechanisms to be disregarded, it would be only fair that the limits of the second group – particularly the hereby-disputed ‘somatic afferents’ argument being held by the authors against the first group – be subjected to equal scrutiny. The experimental design of Haouzi & Chenuel (2005) relied on two complex approaches in anaesthetized sheep: (1) an extracorporeal gas exchanger to ‘isolate’ cephalic P (cP) (for both carotid and central chemoreceptors) from systemic P; and (2) ERC to simulate muscular exercise. Both are problematic. The data in Figs 4 and 7 suggest that increased by 4–5 l min−1 on average during ERC, with cP supposedly remaining constant throughout the 5 min experimental period. However, the online cP monitors used by the authors had a time constant of 20 s (compared with typical 100 ms in infrared CO2 analysers) and a resolution of 2% (i.e. ∼0.8 Torr at a nominal P of 40 Torr; Fig. 2) relative to analysed arterial blood samples. Such a long time constant and limited resolution of measurement in the face of changing make it difficult to as certain that cP was truly constant throughout the experimental period as Figs 4 and 7 purport to show. Indeed, a previous study by the same group using a similar extracorporeal gas exchanger in sheep (Haouzi et al. 2003) reportedly failed to totally isolate cP or suppress the ventilatory CO2 sensitivity. If anything, in 5 out of 7 animals the –cP relationship inexplicably increased 20-fold (to 14 l min−1 Torr−1 on average, sometimes even with a negative gain; see Fig. 7 in Haouzi et al. (2003)). In these majority of cases it should take as little as a 0.8 Torr increase in cP (i.e. within the resolution limit) during ERC to raise by > 11 l min−1, which is more than enough to account for the mere 4–8 l min−1 increase in during ERC (Figs 4 and 7). The data in Fig. 8 intimate a red herring implying a much ‘smaller’–cP slope than previously reported – by evaluating it under two distinct conditions where cP was either isolated or allowed to increase with systemic P (conditions 1 and 2; see legend to Fig. 8) even though Haouzi et al. (2003) found a 20-fold increase in slope only when cP was isolated (i.e. under same condition as during ERC). Thus, the possibility remains that any apparent increase in during ERC may reflect a peculiar chemoreflex gain increase under isolated cP rather than activation of somatic afferents per se. In fact, such anomalous adaptation of chemoreflex gain with differing experimental conditions is consistent with the chemoreflex's newfound plasticity property (Poon & Siniaia, 2000), effectively lending further support for the model of Poon (1987) instead of (Haouzi & Chenuel, 2005). In Fig. 9, it is suggested that occlusion of the inferior vena cava for 1 min during ERC caused an increase in despite a corresponding decrease in cP, whereas occlusion of the aortic artery caused a decrease in . Since both interventions should reduce venous return, the discrepant effects are ascribed by the authors to a difference in muscle peripheral vascular bed and corresponding activation of somatic group III and IV afferent fibres. This speculation neglects the possibility that venous occlusion and arterial occlusion may exert opposite effects on vertebral arterial haemodynamics, which could be a significant factor even at constant carotid pressure. A moderate decrease in brain blood flow resulting from venous occlusion alone could effectively stimulate the central chemical control of breathing (Chapman et al. 1979). With respect to the ERC approach, the authors contend that ‘We modified our technique for contracting the hindlimb muscles... to be as close as possible to the approach used by Cross et al. (1982[b]) in anaesthetized dogs.’‘A low-intensity current was used to minimize the direct stimulation of afferent fibres and to prevent nociceptive stimuli (Cross et al. 1982[b])’. Far from it. On the contrary, the two ERC approaches are incomparable as one uses skin electrodes applied to the thighs of sheep and the other uses needle electrodes directly inserted through the hamstring muscles in dogs. Moreover, the electric current density required for direct activation of afferent fibres and nociceptive stimuli is determined not only by current intensity but also effective electrode area (Alon et al. 1994), but little information about the latter is given. Considering the non-uniform spatial distribution of skin electrical conductance it is entirely possible – even with the use of ‘large electrodes’ and ‘low-intensity current’– for high current density to develop in discrete skin–electrode contact areas causing pain and discomfort (Poon et al. 1980). Indeed, because the thresholds for transcutaneous excitation of muscles or motor nerves are generally higher than for sensory nerves (Kantor et al. 1994), such skin electrodes for ERC are likely to inadvertently activate nociceptive or other afferent fibres and receptors unlike in Cross et al. (1982). Under such circumstances it is incumbent on the authors to directly rule out possible artifacts related to the ‘electrical nature of the stimulation’ rather than simply invoke Cross et al. (1982). Finally, it should be recognized that somatic afferent stimulation may indeed affect ventilation by entraining the respiratory rhythm (Potts et al. 2005), probably via a pathway from the ventral spinocerebellar tract to pneumotaxic neurons in dorsolateral pons (Song et al. 2006). Such a somatic–respiratory coupling mechanism may have more to do with coordination of movements than tracking metabolic rate per se, however.