Reply from Ryan L. Hoiland and Philip N. Ainslie

In the letter to the editor by Pun (2017), several points of concern are raised: (1) that there may have been a difference between the three cyclo-oxygenase (COX) inhibitors we used relative to their effective suppression of prostaglandin production, (2) that carbon dioxide may have itself had an effect on our results, and (3) that the relative balance between COX1 and COX2 inhibition and the resulting effect on the balance between vasodilatory and vasoconstrictor prostaglandins of each drug must be considered. Although these questions were clearly addressed in our paper, and related reference list, we re-address these concerns below. Pun suggests we needed to account for the arterial blood gas parameters during our study. We used end-tidal forcing, which has notably been shown to produce very minor end-tidal to arterial gradients for CO2 (Tymko et al. 2016). Further, the stimulus was matched between trials; therefore, the cause for concern relative to arterial blood gas stimulus is misplaced. Additionally, it is suggested that ‘carbon dioxide itself might have altered blood gas parameters and the vasomotion as well’. Relative to the potential for vasomotion, of which we have published our stance on multiple times (Ainslie & Hoiland, 2014; Hoiland & Ainslie, 2016), the use of duplex ultrasound allows us to quantify large extra-cranial cerebral artery vasomotion (Thomas et al. 2015), as was the purpose of the study under discussion (Hoiland et al. 2016). This approach arguably allows for more parsimonious interpretation relative to previous studies (Ainslie & Hoiland, 2014), such as those using transcranial Doppler (e.g. Beaudin et al. 2014). Relative to the variable COX1 and COX2 selectivity between non-steroidal anti-inflammatory drugs (Park & Bavry, 2014), it is apparent that ketorolac and naproxen are both slightly more selective to COX2 while indomethacin is slightly more selective for COX1. While this may appear a cause for concern when comparing these drugs, the same lack of effect of COX inhibition on cerebral vascular CO2 reactivity has also been highlighted in studies using other drugs possessing a greater COX1 versus COX2 selectivity similar to indomethacin (INDO). For example, ibuprofen does not influence cerebral vascular CO2 reactivity in animal models, whereas INDO does (Chemtob et al. 1991), while aspirin (acetylsalicylic acid) possesses no effect in humans (Markus et al. 1994). Therefore, contrary to the suggestion by Pun, it does not seem plausible that differences in COX1/COX2 selectivity impact the effectiveness of COX inhibitors on CO2 reactivity. The influence of several COX inhibitors that have been used to investigate cerebral vascular regulation by CO2 is summarized in Table 1. Given this marked difference between INDO and other COX inhibitors, we found that INDO is a potent inhibitor of cyclic AMP-dependent protein kinase (Kantor & Hampton, 1978; Goueli & Ahmed, 1980), which is integral to the regulation of smooth muscle tone (Adelstein & Conti, 1978). Given the potent inhibition of cAMP by indomethacin, it is difficult to contest that this mechanism is not inhibiting vasodilatation during hypercapnia; however, it may be more reasonable to contest that cAMP inhibition may not represent the entire difference in effect between drugs (INDO, naproxen, ketorolac). In this regard, some points raised by Pun may be relevant in explaining a portion of the difference in drug effects (albeit a relatively small portion in our opinion, if at all). However, previous research has indicated the doses we used for INDO and naproxen produce similar, and marked, inhibition of prostaglandin synthesis (Eriksson et al. 1983), while a study using a lower INDO dose (0.8 mg kg−1; Table 1) showed consistent reductions in cerebral vascular CO2 reactivity (Wennmalm et al. 1984). Therefore, the rationale to assume the difference in action may be due to dosage, as Pun suggests, appears unfounded. Further, under this assumption that prostaglandin synthesis inhibition is not appreciably different between the drugs under discussion, differences in the potential upregulation of other arachidonic acid end products (20-hydroxyeicosatetraenoic acid and lipoxygenase generating vasoactive mediators) would also appear to be quite unlikely. Considering the importance of exploring longer duration dosing as Pun suggests, these data have been previously published – they highlight that INDO continues to reduce cerebral vascular CO2 reactivity following 1 week of 0.8 mg kg−1 three times daily (Eriksson et al. 1983). As summarized in our original article (Hoiland et al. 2016), and reiterated here, INDO must be acting via a permissive mechanism(s) unrelated to COX inhibition given other COX inhibitors do not affect cerebral vascular CO2 reactivity in healthy humans (Table 1). Given the evidence on INDO's potent inhibition of cAMP-dependent protein kinase (Kantor & Hampton, 1978; Goueli & Ahmed, 1980), and cAMP's integral action in regulating vascular tone (Adelstein & Conti, 1978), cAMP dependent protein kinase inhibition seems a strong candidate for this permissive action. Discussion of these topics is imperative in furthering our understanding of drug-mediated changes in vascular function, and highlights that the potential for effects in addition to any drugs primary route of action need to be considered when developing an experimental paradigm. The authors declare no conflict of interest, financial or otherwise. P.N.A. receives funding from the Gouvernement du Canada/Natural Sciences and Engineering Research Council of Canada (Conseil de Recherches en Sciences Naturelles et en Génie du Canada) and a Canada Research Chair in cerebrovascular physiology. Ryan Hoiland is supported by a NSERC post-graduate scholarship.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.