Acetazolamide during acute hypoxia improves tissue oxygenation in the human brain

Abstract

Low doses of the carbonic anhydrase inhibitor acetazolamide provides accelerated acclimatization to high-altitude hypoxia and prevention of cerebral and other symptoms of acute mountain sickness. We previously observed increases in cerebral O₂ metabolism (CMR_O2) during hypoxia. In this study, we investigate whether low-dose oral acetazolamide (250 mg) reduces this elevated CMR_O2 and in turn might improve cerebral tissue oxygenation (Pti_O2) during acute hypoxia. Six normal human subjects were exposed to 6 h of normobaric hypoxia with and without acetazolamide prophylaxis. We determined CMR_O2 and cerebral Pti_O2 from MRI measurements of cerebral blood flow (CBF) and cerebral venous O₂ saturation. During normoxia, low-dose acetazolamide resulted in no significant change in CBF, CMR_O2, or Pti_O2. During hypoxia, we observed increases in CBF [48.5 (SD 12.4) (normoxia) to 65.5 (20.4) ml·100 ml⁻¹·min⁻¹ (hypoxia), P < 0.05] and CMR_O2 [1.54 (0.19) to 1.79 (0.25) μmol·ml⁻¹·min⁻¹, P < 0.05] and a dramatic decline in Pti_O2 [25.0 to 11.4 (2.7) mmHg, P < 0.05]. Acetazolamide prophylaxis mitigated these rises in CBF [53.7 (20.7) ml·100 ml⁻¹·min⁻¹ (hypoxia + acetazolamide)] and CMR_O2 [1.41 (0.09) μmol·ml⁻¹·min⁻¹ (hypoxia + acetazolamide)] associated with acute hypoxia but also reduced O₂ delivery [6.92 (1.45) (hypoxia) to 5.60 (1.14) mmol/min (hypoxia + acetazolamide), P < 0.05]. The net effect was improved cerebral tissue Pti_O2 during acute hypoxia [11.4 (2.7) (hypoxia) to 16.5 (3.0) mmHg (hypoxia + acetazolamide), P < 0.05]. In addition to its renal effect, low-dose acetazolamide is effective at the capillary endothelium, and we hypothesize that local interruption in cerebral CO₂ excretion accounts for the improvements in CMR_O2 and ultimately in cerebral tissue oxygenation during hypoxia. This study suggests a potentially pivotal role of cerebral CO₂ and pH in modulating CMR_O2 and Pti_O2 during acute hypoxia.

acute exposure to hypoxia at high altitude can induce symptoms of acute mountain sickness (AMS). Acclimatization to hypoxia is accelerated and AMS can be prevented by prophylactic treatment with the carbonic anhydrase inhibitor acetazolamide (reviewed by Ref. 37). One intriguing effect of acetazolamide is a small (∼ 5%) reduction in whole body oxygen consumption (V̇o₂) in both human and animal studies (5, 18, 31, 40) and similar effects on cerebral O₂ metabolism (CMR_O2) (49). These studies focused on high-dose acetazolamide (typically 750–1,000 mg) that achieves almost complete inhibition of carbonic anhydrase. While the primary effect of acetazolamide is on acid base balance by producing a mild renal metabolic acidosis and an increase in ventilation (22), it also inhibits carbonic anhydrase in a dose-dependent fashion at other sites throughout the body, which might have additional effects related to high-altitude adaptation (37, 38). Considerably lower dose acetazolamide (250 mg) in which there is only selective or partial inhibition of carbonic anhydrase activities is effective to prevent AMS (1, 19, 24, 42); thus complete blockade of all carbonic anhydrase activity (from high-dose acetazolamide) is not required to achieve clinical benefit.

In recent studies we observed increases in CMR_O2 during acute and sustained hypoxia (34, 35). The hypothesis we were testing in the present study is whether the metabolism-lowering effects of acetazolamide might counteract the metabolism-increasing effects of hypoxia. Specifically, we tested if low-dose acetazolamide (which is know to be clinically effective in AMS prevention) mitigates the elevated CMR_O2 associated with cerebral hypoxia in the acute setting (before any onset of AMS) and if this in turn might improve cerebral tissue oxygenation (Pti_O2).

Using MRI, we measured cerebral blood flow (CBF), and venous O₂ saturation in the superior sagittal sinus (Sv_O2), and from these we determined CMR_O2 and cerebral tissue Pti_O2 responses to acetazolamide during 6 h of normobaric hypoxia.

METHODS

Subjects

Six healthy, nonsmoking, sea-level residents age 35 ± 9 yr (2 females, 4 males) participated in the study. Ethical approval for these studies was granted by the Human Research Protection Program of the University of California, San Diego. Participants were informed of the experimental procedures and possible risks involved in the study, and written informed consent was obtained before participation.

Study Design

The impact of acetazolamide on CBF, CMR_O2, and Pti_O2, was evaluated during normoxia and following 6 h of normobaric hypoxia. Baseline MRI measurements were made during normoxia of steady-state cerebral blood flow (CBF) and steady-state transverse relaxation (T₂) in superior sagittal sinus blood [from which we calculated venous oxygen saturation (Sv_O2)]. Additional measurements were also made of hemoglobin (Hb) concentration, hematocrit (Hct), and pulse oximeter measurements of arterial hemoglobin O₂ saturation (Sa_O2). From these measurements we determined steady-state CMR_O2 (35). By incorporating these data into a diffusion model of O₂ transport, we calculated the partial pressure of tissue oxygen in the brain (Pti_O2) (4). The measurements were repeated following 6 h of acute normobaric hypoxia at Sa_O2 ∼85% (Fig. 1). These normoxic and hypoxic measurements were repeated on a subsequent day, following prophylaxis with oral acetazolamide. We chose the lowest clinically effective dose of acetazolamide for prophylaxis (250 mg; Refs. 1, 19, 24, 42) administered as a divided dose of 125 mg ∼2 h before the baseline MRI scan and 125 mg ∼3 h before the hypoxic MRI scan.

Fig. 1.

MRI measures of cerebral blood flow (CBF) and Sv_O2 were made during normoxia along with pulse oximeter measures of Sa_O2. From these we calculated cerebral O₂ metabolism (CMR_O2). Subjects then breathed a hypoxic gas mixture that was continually adjusted to maintain Sa_O2 at ∼85% for 6 h (Fi_O2 = ∼0.12–0.13). Subjects breathed a premixed hypoxic gas mixture within the MRI (dashed line) containing 90 Torr O₂, and CBF, Sv_O2, and Sa_O2 measures were repeated. Measurements were repeated on a subsequent day following a split dose of acetazolamide.

Acute Hypoxia Paradigm

Within the MRI scanner subjects breathed a premixed 90-Torr hypoxic mixture (12.5% O₂, balance N₂) via a close-fitting low-deadspace nonrebreathing mask (Hans Rudolph 7900/2600, Kansas City, MO). Thirty minutes were included before the MRI measurements for subjects to acclimate to the mask. For the 6-h hypoxic exposure outside the scanner, the close fitting masks were poorly tolerated (and precluded subjects from drinking or eating). Instead, subjects breathed nitrogen-enriched air via a nonrebreathing loose-fitting Venturi mask (Hudson, Temecula, CA), and gas flow rate was adjusted to maintain a stable Sa_O2 at ∼85% (Fi_CO2 ∼0.12-0.13). Since we did not have ambulatory O₂ monitoring to specifically target Fi_O2, we monitored Sa_O2 as a surrogate. The main motivation for this additional setup was to improve subject comfort and compliance. The lighter weight mask was better tolerated by subjects for prolonged periods and improved Sa_O2 stability during the hypoxic exposure.

Physiological Measurements

Arterial O₂ saturation was continuously measured using a Nonin 3100 Wrist Pulse Oximeter (during the 6-h hypoxia exposure between MRI measurements) and a Nonin 8600FO MRI-compatible pulse oximeter (Nonin Medical, Plymouth, MN; during MRI measurement). Inspired and expired partial pressures of O₂ and CO₂ were continuously monitored during MRI measurements using a Perkin Elmer 1100 medical gas spectrometer (Perkin Elmer, Waltham, MA). Hct and Hb were determined from a venous blood sample; Hb was measured with an IL-682 co-oximeter (Instrument Laboratories, Lexington, MA). Hematocrit was determined from direct measurements of packed cell height in a capillary tube following centrifuging. To ensure that the hypoxic and cardiorespiratory state remained consistent during the steady-state measurements, Sa_O2 and Pet_CO2 were compared during CBF and venous T₂ MRI measurements. Imaging sequences were repeated if Sa_O2 deviated by more than ±2%, or Pet_CO2 deviated by ±2 Torr between these scans.

MRI Measurements

Cerebral blood flow.

CBF was measured in gray matter in the cerebrum using a PICORE QUIPPS 2 arterial spin labeling (ASL) technique (TE = 9.1 ms, TR = 2.5 s, TI₁ = 700 ms, TI₂ = 1,500 ms, six 6-mm slices, 5 min). Additional images were collected to determine proton density of cerebrospinal fluid and coil sensitivity profile to allow absolute quantitation of CBF from the MRI signal (28, 44, 46).

Venous T₂ relaxation.

T₂ was measured in the superior sagittal sinus using a TRUST (T₂ relaxation under spin tagging) MRI technique (25) with single shot spiral readout (TE = 2.8 ms, TR = 8 s, TI = 1.2 s, 4 echoes at effective TE 0, 40, 80, and 160 ms, 10-mm slice, 80 mm tag, 4.5 min). Images were acquired just superior to the torcula. The T₂ value was then used to determine venous O₂ saturation.

Data Analysis

Cerebral blood flow.

Raw ASL data for each subject were corrected for the expected changes in T₁ relaxation of blood based on the measured Sa_O2 (33). Images were corrected for cardiorespiratory physiological noise (29) and field inhomogeneities (27). Resting CBF was averaged across the 5 min of data collection in cerebral gray matter using a mask generated from a separate high-resolution Fast Spoiled GRASS (FSPGR) T₁-weighted 3D anatomical MRI (TE = 4.2 ms, TR = 10.1 ms, TI = 450 ms, bandwidth = 20.83 kHz, field of view = 25 × 25 × 16 cm, matrix 256 × 256 × 128, ∼1 × 1 × 1.3 mm resolution, 5.5 min). Cerebral gray matter was automatically segmented in the FSPGR scan using FAST software (FMRIB Software Library, Oxford, UK).

Venous oxygen saturation (Sv_O2).

Cerebral venous O₂ saturation was derived from the T₂ relaxation of venous blood in the superior sagittal sinus. This uses a simplified Luz-Meiboom model, with a quadratic dependence of T₂ on O₂ saturation and hematocrit, which provides a validated method for determining Sv_O2 (26, 48). Calibration scaling constants were determined from a prior normative group (35).

Cerebral oxygen delivery (do₂).

Cerebral oxygen delivery was calculated based on CBF and O₂ content of the arterial blood (Ca). We assumed negligible O₂ dissolved in plasma and approximated the Ca by 1.36 × [Hb] × Sa_O2 (6).

Cerebral oxygen metabolism (CMR_O2).

Cerebral O₂ metabolism was calculated using the Fick equation:

$${\text{CMRO}}_2=\text{CBF}\cdot \text{OEF}\cdot \text{Ca}$$ (1)

where CBF is the cerebral blood flow, OEF is the O₂ extraction fraction between arterial and venous blood, (Sa_O2-Sv_O2)/Sa_O2, and Ca is the O₂ content of arterial blood.

Since the arterial O₂ content is approximated by the product of Sa_O2 and the maximum O₂ content of fully saturated blood, this can be rewritten as:

$${\text{CMRO}}_2=\text{CBF}\cdot \frac{\%{\mathrm{S}\mathrm{a}}_{{\mathrm{O}}_{\mathrm{2}}}-\%{\mathrm{S}\mathrm{v}}_{{\mathrm{O}}_{\mathrm{2}}}}{100}\left[\text{Hb}\right]$$ (2)

In this notation, Hb concentration is expressed in units of molar equivalents of oxygen carried per liter of blood when hemoglobin is fully saturated (meq/l) (6). CBF was calculated from the arterial spin labeled MRI measurements, Sa_O2 from a pulseoximeter, Sv_O2 from the sagittal sinus T₂ MRI measures, and Hb concentration from a blood sample. CBF is in units of ml·100 ml⁻¹·min⁻¹, CMR_O2 is in μmol·ml⁻¹·min⁻¹ (assuming 1 ml tissue/g, this is numerically equivalent to μmol·g⁻¹·min⁻¹), and Sv_O2 and Sa_O2 are in percentages.

Cerebral tissue oxygenation (Pti_O2).

The calculation of cerebral tissue Pti_O2 is based on the theoretical framework proposed by Buxton (4) but modified to include an arterial hemoglobin saturation <100% (i.e., hypoxic). This model allows us to calculate Pti_O2 from CBF and CMR_O2 given a baseline (normoxic) and challenge (hypoxic) state.

For hypoxia experiments, the measured quantities for the model are fractional arterial hemoglobin saturation, Y_a (= %Sa_O2/100) (from pulseoximeter measurements), fractional venous hemoglobin saturation, Y_v (= %Sv_O2/100) (from TRUST MRI), and the ratio (f) of CBF during hypoxia to CBF in the baseline normoxic state (from ASL MRI measurements).

The mass balance equation for the ratio (r) of CMR_O2 in the hypoxic state to CMR_O2 in the baseline state is:

$$r=f\frac{Y_{\text{a}}-Y_{\text{v}}}{Y_{\text{a,0}}-Y_{\text{v,0}}}$$ (3)

The mean capillary Po₂ (Pc_O2) is taken as the Po₂ given by the Hill equation for a saturation halfway between arterial and venous:

$${\text{Pc}}_{{\text{O}}_{\text{2}}}={\text{P}}_{50}{\left(\frac{Y_{\text{a}}+Y_{\text{v}}}{2-Y_{\text{a}}-Y_{\text{v}}}\right)}^{1/h}$$ (4)

A second expression for the metabolic rate ratio r, based on blood/tissue O₂ diffusion is given in Eq. 5 in terms of mean capillary Pc_O2 and tissue Pti_O2 during hypoxia and during baseline normoxia (4).

$$r=\frac{{\text{Pc}}_{{\text{O}}_{\text{2}}}-{\text{Pti}}_{{\text{O}}_{\text{2}}}}{{\text{Pc}}_{{\text{O}}_{\text{2,0}}}-{\text{Pti}}_{{\text{O}}_{\text{2,0}}}}$$ (5)

Equation 3 is used to calculate the corresponding CMR_O2 ratio and Eq. 4 is used to calculate the Pc_O2 in normoxic and hypoxic condition. Equation 5 is then used to calculate Pti_O2.

There are several model parameters that must be assumed in addition to the measured parameters. These are baseline tissue Po₂ (Pti_O2,0), the Po₂ for 50% saturation of hemoglobin (P₅₀), and the Hill exponents (h). These assumed values are Pti_O2,0 = 25 mmHg, P₅₀ = 26 mmHg, and h = 2.8 (4).

Statistical Analysis

Data were analyzed with repeated measures ANOVA of our primary outcome variables (CMR_O2, CBF, do₂, Pet_CO2, Sa_O2, Sv_O2, Pti_O2), with two repeated measure groupings (hypoxic exposure, acetazolamide prophylaxis), each at two levels (normoxia, 6-h hypoxia, no acetazolamide, acetazolamide) (StatView 5.0.1; SAS Institute, Cary, NC). This paired approach allowed each subject to act as his or her own control, which provided greater statistical power to detect small changes. Post hoc power was calculated using G*Power (G*Power 3.1.3; Heinrich Heine University, Dusseldorf, Germany). Data are expressed as mean (SD). Changes were significant at P < 0.05 two tailed.

RESULTS

All six subjects completed the study. The results are summarized in Tables 1 and 2.

Table 1.

Primary outcome variables

	SaO2, %	SvO2, %	PetCO2, mmHg	CBF, ml·100 ml−1·min−1	do2 mmol/min	OEF	CMRO2, μmol·ml−1·min−1	PtiO2, mmHg	SaO2-SvO2
Normoxia
no-Rx (n = 6)
Mean (SD)	97.6 (0.6)	57.8 (3.1)	38.6 (2.3)	48.5 (12.4)	6.36 (0.91)	0.41 (0.03)	1.54 (0.19)	25.0 (−)	39.8 (3.1)
CI	97.1–98.1	55.3–60.3	36.8–40.5	38.6–58.5	5.63–7.09	0.38–0.43	1.39–1.69	25.0–25.0	37.3–42.3
Rx-Az (n = 6)
Mean (SD)	97.7 (1.1)	58.8 (4.3)	38.1 (2.4)	49.7 (19.0)	6.26 (1.29)	0.40 (0.04)	1.48 (0.27)	25.7 (5.0)	38.9 (4.1)
CI	96.8–98.5	55.3–62.2	36.2–40.0	34.5–64.9	5.22–7.29	0.36–0.43	1.26–1.70	21.7–29.7	35.6–42.2
6-h Hypoxia
no-Rx (n = 6)
Mean (SD)	79.2 (4.4)	44.3 (4.0)	36.3 (3.9)	65.5 (20.4)	6.92 (1.45)	0.44 (0.05)	1.79 (0.25)	11.4 (2.7)	34.9 (4.8)
CI	75.6–82.7	41.1–47.5	33.2–39.5	49.2–81.7	5.76–8.08	0.40–0.48	1.60–1.99	9.3–13.6	31.0–38.7
Rx-Az (n = 6)
Mean (SD)	81.3 (5.5)	46.3 (6.4)	33.4 (4.7)	53.7 (20.7)	5.60 (1.14)	0.43 (0.07)	1.41 (0.09)	16.5 (3.0)	35.0 (5.5)
CI	76.9–85.7	41.2–51.4	29.7–37.2	37.1–70.3	4.69–6.51	0.38–0.48	1.34–1.48	14.1–18.9	30.6–39.4
Statistics (two-way ANOVA)
P hypoxia	0.0002	0.0002	0.02	0.04	NS	NS	NS	0.0001	0.04
P Rx	NS	NS	NS	0.05	0.01	NS	0.02	0.04	NS
P hypoxia × Rx	NS	NS	NS	0.06	NS	NS	0.06	NS	NS

Primary outcome variables are summarized as mean (SD) and 95% confidence intervals (CI) for each experimental session and are defined as follows: sao2≈, arterial O₂ saturation; Sv_O2, venous O₂ saturation; Pet_CO2, end-tidal partial pressure of CO₂; CBF, cerebral blood flow; do₂, O₂ delivery; OEF, oxygen extraction fraction; CMR_O2,cerebral oxygen metabolic rate, Pti_O2, O₂ partial pressure in cerebral tissues. The data were analyzed using two-way repeated measures ANOVA with 2 grouping variables, each at 2 levels: 1) normoxia vs. 6-h hypoxia; and 2) treatment with acetazolamide (Rx-Az) vs. no treatment (no-Rx). The P values for the main effect of hypoxia (P hypoxia), main effect of acetazolamide (P Rx), and hypoxia × acetazolamide interaction (P hypoxia × Rx) are reported only for statistical significant results (P < 0.05). Additional post hoc statistical analysis is included in the relevant figure legends. Note: Pti_O2 for normoxia, no-Rx are reference values (4) so no SD are reported.

Table 2.

Primary outcome variables from Table 1 presented as difference measures between normoxia vs. hypoxia conditions and between no-acetazolmide vs. acetazolamide prophylaxis

	SaO2, %	SvO2, %	PetCO2, mmHg	CBF, ml·100 ml−1·min−1	do2, mmol/min	OEF	CMRO2, μmol·ml−1·min−1	PtiO2, mmHg	SaO2-SvO2
Pair-wise difference (normoxia − hypoxia)
no-Rx (n = 6)
Mean (SD)	18.4 (4.1)	13.5 (5.4)	2.3 (4)	−16.9 (12.8)	−0.6 (1.3)	0 (0.1)	−0.3 (0.2)	13.6 (2.7)	4.9 (7.2)
CI	12.4–23.9	6.4–19.3	−2.8–7.5	−30.7–3.3	−2.3–1.2	−0.1–0.1	−0.6–0.1	10.3–18.6	−6.9–13.3
Rx-Az (n = 6)
Mean (SD)	16.4 (5.2)	12.4 (6.5)	4.7 (2.5)	−4 (9.6)	0.7 (0.8)	0 (0.1)	0.1 (0.2)	9.2 (4.5)	3.9 (6.6)
CI	8.1–22.3	4.2–22.3	1.3–7.7	−20.6–7.1	−0.3–1.9	−0.1–0.1	−0.2–0.4	3.4–14	−0.8–16.8
Pair-wise difference (no-acetazolamide − acetazolamide)
Normoxia (n = 6)
Mean (SD)	−0.1 (0.9)	−0.9 (5.8)	0.5 (2.4)	−1.2 (6.6)	0.1 (0.5)	0 (0.1)	0.1 (0.2)	−0.7 (4.5)	0.9 (6.2)
CI	−1.3–1.7	−9.7–6	−2.7–3.8	−14.5–4.7	−0.6–0.7	−0.1–0.1	−0.3–0.4	−7.3–6.3	−6.1–11.4
Hypoxia (n = 6)
Mean (SD)	−2.1 (2.9)	−2 (5.6)	2.9 (3.4)	11.8 (9.4)	1.3 (1.1)	0 (0.1)	0.4 (0.2)	−5.1 (2.8)	−0.1 (6.5)
CI	−4.3–3.2	−7.1–8.3	−0.9–9.2	0.9–24.5	0.3–2.5	−0.1–0.1	0.2–0.6	−8.2–0.3	−9.7–10.3

Primary outcome variables are summarized as mean (SD) and 95% confidence intervals (CI) for each experimental session.

Normoxia vs. Hypoxia

To investigate the physiological changes between normoxia and hypoxia, measurements of Sa_O2, Sv_O2, Sa_O2-Sv_O2, OEF, Pet_CO2, CBF, CMR_O2, and do₂ were first compared between normoxic vs. hypoxic conditions. Subjects showed decreased Sa_O2, Sv_O2, and increased ventilatory drive during hypoxia, with decreased Pet_CO2 (Fig. 2). Both CBF and CMR_O2 were significantly increased during hypoxia. There was no significant change in do₂ during hypoxia. The net effect of this was a significant decline in estimated Pti_O2 during hypoxia (Fig. 3).

Fig. 2.

Change in hemoglobin oxygen saturation (A) in arterial (solid symbols) and venous blood (open symbols) and change in end-tidal Pco₂ (B) during normoxia and following 6-h acute hypoxia, with and without acetazolamide treatment. During hypoxia there was a significant decrease in both arterial and venous hemoglobin oxygen saturation (P = 0.0002, P = 0.0002). Pet_CO2 also showed a small decline (9%, P = 0.02, post hoc paired t-test). During hypoxia, both venous and arterial hemoglobin oxygen saturation showed a small (but non-significant) improvement with acetazolamide prophylaxis. There was a trend towards a decrease in Pet_CO2 with acetazolamide prophylaxis (P = 0.089 relative to no acetazolamide). Error bars indicate ± 1 SD. Dashed lines are 95% confidence limits.

Fig. 3.

Changes in CBF (A), cerebral tissue Po₂ (B), O₂ delivery (C), and CMR_O2 (D) during normoxia and following 6 h acute hypoxia, with and without acetazolamide prophylaxis. Following hypoxia, there was a 26% increase in CBF relative to normoxia (P = 0.02, post hoc paired t-test). This was accompanied by a 16% increase in CMR_O2 (P = 0.03, psot hoc paired t-test). Since O₂ delivery was not significantly changed (P = 0.328), this resulted in a 54% decrease in cerebral tissue Pti_O2 (P < 0.001, post hoc paired t-test). When acetazolamide was administered before the hypoxic exposure, the hypoxic CBF elevation was prevented, and CBF remained unchanged relative to normoxia levels (P = 0.36). The hypoxic rise in CMR_O2, was also prevented and in fact showed a small decrease of 4.7% (although not significantly different from normoxia, P = 0.42). Acetazolamide also reduced the variance in CMR_O2 during hypoxia between subjects. O₂ delivery during hypoxia was also lower with acetazolamide treatment (P = 0.013). The decline in cerebral Pti_O2 during hypoxia was partially mitigated in the presence of acetazolamide (P = 0.007 relative to no acetazolamide). This improvement in cerebral tissue Pti_O2 by acetazolamide during hypoxia was primarily due to its impact in reducing CMR_O2, which outweighed the decline in O₂ delivery. Error bars indicate ± 1 SD. Dashed lines are 95% confidence limits.

Effect of Acetazolamide in Normoxia and Hypoxia

To assess the effect of acetazolamide, data were analyzed with and without acetazolamide prophylaxis for normoxia and hypoxia conditions.

In normoxia, acetazolamide had no significant effect on all physiological variables measured. In hypoxia, prophylactic acetazolamide had no significant effect on Sa_O2, Sv_O2, Sa_O2-Sv_O2, but there was a trend towards decreased Pet_CO2 (P = 0.089, post hoc t-test) (Fig. 2). Acetazolamide prevented the hypoxia-related increases in both CMR_O2 and CBF, which resulted in a decrease in do₂ (Fig. 3). The relative reduction in CMR_O2 exceeded the reduction in do₂, thus improving cerebral Pti_O2 during hypoxia (Fig. 3B).

DISCUSSION

Hypoxia stimulates ventilatory drive via its effect on the peripheral chemoreceptors. This stimulated increase in alveolar ventilation also reduces Pa_CO2 (itself a potent ventilatory stimulant); thus the resulting hypocapnia and respiratory alkalosis limit the potential magnitude of any hypoxia-driven increase in ventilation and hence limit the potential to increase arterial oxygenation. The established view of the therapeutic effect of acetazolamide in hypoxia is a renally mediated metabolic acidosis with an enhancement in minute ventilation and oxygenation that would otherwise be limited by hypocapnia (22, 37).

Cerebral Hemodynamic and Metabolic Responses During Acute Hypoxia

In the current study, we observed that CMR_O2 increased moderately (16%) during 6 h of hypoxia with concurrent increase in CBF. This is consistent with previous studies showing that CMR_O2 and CBF increase during acute and sustained hypoxic hypoxia (34, 35, 47). In addition, we observed a significant decrease in tissue Pti_O2 during hypoxia (to 54% of normoxia level), and no change in OEF or O₂ delivery (do₂). This mirrors previous findings during prolonged (>10 h) hypoxia in human subjects and animal experiments (36, 45).

Acetazolamide Effect on CBF

Literature reports of the effect of acetazolamide on CBF show considerable variation (reviewed by Ref. 22), with increases (2, 3, 8), no change or decreases in CBF (16, 36). Investigators who reported increases in CBF only used higher doses of acetazolamide (750 mg to 1 g) given intravenously with measurements of CBF made within 15–30 min at the height of very high plasma concentrations of the drug rather than a more clinically applicable lower orally administered dose used in the current study and conventionally for AMS prophylaxis (250 mg, i.e., ∼3 mg/kg). In fact, Grossmann and Koeberle (13) observed a linear dose-response in CBF which increased from 5 mg/kg up to 20 mg/kg iv of acetazolamide administration, but no significant CBF change with lower doses (< 5 mg/kg even given iv).

We also found that low-dose acetazolamide did not change CBF during normoxia, but it did decrease the otherwise elevated CBF during acute hypoxia. Selective inhibition of carbonic anhydrase subtypes in different tissues likely contributes to these results; The more abundant type II carbonic anhydrase, which is mainly located in erythrocytes (20), is not inhibited at low doses of acetazolmide, so it is still fully functional in the vasculature. Type IV carbonic anhydrase on the endothelium of cerebral capillary beds (11), is completely inhibited by even very low doses of acetazolamide, leading to cerebral tissue CO₂ retention and increases the cerebrospinal fluid Pco₂ by ∼1–2 mmHg (21, 41). This is sufficient to drive up alveolar ventilation since central chemoreceptors are highly sensitive to CO₂ changes in the cerebrospinal fluid (22), which then result in ventilatory hypocapnia (36). Since erythrocyte type II carbonic anhydrase is still active, direct vasodilation is not a primary effect with low-dose acetazolamide (13), and the overall result is a decrease in CBF due to the hypocapnic vasoconstriction.

Acetazolamide Effect on CMR_O2

The sensitivity of CMR_O2 to hypoxic hypoxia and to the effects of low-dose acetazolamide suggests CO₂ and local tissue pH play an important role in regulating O₂ metabolism. Tissue CO₂ indirectly stimulates adenosine generation in the brain via a reduction in extracellular pH, which suppresses neuronal firing (7). Changes in pH from carbonic anhydrase inhibition have also been shown to reduce neuronal excitability (23). We hypothesize that the increased CMR_O2observed during hypoxia in the current study might also represent adenosine-mediated modulation in neuronal excitability. Hypoxia-induced hyperventilation will decrease Pa_CO2, and ultimately reduce Pco₂ in cerebral tissues, thus reducing CO₂-mediated neuronal suppression and increasing neural excitability and oxygen utilization. Low-dose acetazolamide selectively inhibits cerebral endothelial carbonic anhydrase, which leads to an increase in cerebral tissue Pco₂. This in turn will impact cerebral adenosine generation and limit CMR_O2. This mechanism is supported by recent studies demonstrating that the hypoxia-induced rise in CMR_O2 can be partially mitigated during isocapnic hypoxia by maintaining Pet_CO2 at normoxia levels (34). Elevating Pet_CO2 during hypoxia also mitigates many of the cerebral symptoms associated with acute mountain sickness (14, 17). The pH sensitivity of phosphofructo-1-kinase, an important rate-limiting step in glycolysis, offers an additional pathway for reducing O₂ consumption, (10).

Acetazolamide Effect on Pti_O2

Selective inhibition of carbonic anhydrase by low-dose acetazolamide also impacts cerebral tissue oxygenation. A secondary outcome of preventing the rise in CBF during hypoxia by low-dose acetazolamide is a decrease in O₂ delivery to the brain. On the face of it, this does not appear to be a positive effect of acetazolamide prophylaxis. However, the reduction in CMR_O2 by acetazolamide exceeds the reduction in O₂ delivery, which reduces the degree of cerebral tissue Pti_O2 decline otherwise present in hypoxia. In a study examining cerebral oxygenation during exercise in trekkers at high altitude, Vuyk et al. (43) observed a reduction in cerebral oxygenation measured by near-infrared spectroscopy (NIRS). Following acetazolamide the cerebral oxygenation increased. The increase in cerebral oxygenation exceeded the expected increase from just improved arterial oxygenation due to ventilation (37, 43). This finding supports our proposed mechanism of acetazolamide altering both the delivery and utilization of oxygen in hypoxia.

Measurement Limitation

Pet_CO2 might not accurately reflect the underlying Pa_CO2 after acetazolamide. Consistent with this notion is a recent study at high altitude, which showed the increase in Pa_CO2 (∼4 mmHg) is greater than that in Pet_CO2 (∼1 mmHg) after acetazolamide administration (9). The current study did not include separate arterial blood gas measurements. Thus our measurement of Pet_CO2 following acetazolamide may in fact underrepresent the actual changes in Pa_CO2.

Another potential limitation in our experimental design is the timing of the acetazolamide dose. We gave this as a divided dose to maintain a more constant plasma concentration and limit side effects. One-hundred and twenty-five milligrams were taken 2 h before the first (normoxic) MRI. Six hours later, a second 125-mg dose was taken ∼3 h before the second (hypoxic) MRI. Oral acetazolamide has an absorption half-life of ∼1 h, so maximum concentration may take up to ∼4 h after ingestion (30). Thus the degree of carbonic anhydrase inhibition could be less complete during the first MRI scan. Due to logistical constraints, the ordering of the normoxia/hypoxia MRI studies was fixed with the normoxia measurements occurring first. Thus this difference in carbonic anhydrase inhibition primarily impacted the normoxia measurements. The conclusions of this study thus focus on the more prolonged acetazolamide effects during the hypoxia measurements.

Our conclusions from these studies are impacted by the assumptions underlying our modeling of Pti_O2 (reviewed in Ref. 4). Our model for Pti_O2 assumed that increases in cerebral perfusion do not involve additional capillary recruitment (12). The Pti_O2 calculation also assumed a constant oxygen-hemoglobin dissociation curve (4); however, the acidosis due to acetazolamide could cause a significant right shift in the dissociation curve, and this in turn would increase the capillary Po₂ and the tissue Po₂. We observed ∼3 mmHg decline in Pet_CO2 with acetazolamide after 6 h of hypoxia, which is within the expected range for this low-dose (selective carbonic anhydrase inhibition) regimen (31). We did not make independent measurements of arterial pH in this study, but estimates of the effect on pH from the two 125-mg acetazolamide doses in normal human subjects would be a decrease from 7.4 to 7.33 during the second (hypoxic) MR measurement (39). We calculated that this decrease in pH would actually result in a higher Pti_O2 of 19.2 mmHg (15, 32). Thus assuming a constant oxygen dissociation curve does not affect our primary finding of an improvement in tissue Pti_O2 during hypoxia with acetazolamide prophylaxis, although the true magnitude of this effect may in fact be ∼16% greater than we present here.

Conclusion

Within the limits of our models discussed above, the primary findings of this study are that low-dose oral acetazolamide, when given prophylactically, attenuates the rise of CMR_O2 during hypoxia and thus improves cerebral tissue oxygenation. This suggests a potentially important role played by cerebral CO₂ and pH in modulating CBF, CMR_O2, as well as Pti_O2 during hypoxia.

This mode of action of acetazolamide in altering both cerebral O₂ delivery and O₂ utilization and hence preserving tissue oxygenation during hypoxia may be an important mechanism underlying the prophylactic effect of acetazolamide on acute mountain sickness and warrants further investigation.

GRANTS

This work was supported by National Institute of Neurological Disorders and Stroke Grants R01-NS-053934, R21-NS-075812, and R01-NS-036722.

DISCLAIMERS

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: K.W., Z.M.S., and D.J.D. performed experiments; K.W., Z.M.S., R.B.B., and D.J.D. analyzed data; K.W., Z.M.S., R.B.B., E.R.S., and D.J.D. interpreted results of experiments; K.W., Z.M.S., and D.J.D. prepared figures; K.W., Z.M.S., R.B.B., and D.J.D. drafted manuscript; K.W., R.B.B., E.R.S., and D.J.D. edited and revised manuscript; K.W., Z.M.S., R.B.B., E.R.S., and D.J.D. approved final version of manuscript; D.J.D. conception and design of research.