Acute effect of acetazolamide on exercise haemodynamics in patients with pulmonary arterial and chronic thromboembolic pulmonary hypertension: a randomised controlled trial

Abstract

Background

The aim of this study was to investigate the acute effect of acetazolamide on pulmonary haemodynamics during exercise in patients with pulmonary vascular disease (PVD).

Methods

Patients with PVD diagnosed as pulmonary arterial (PAH) or chronic thromboembolic pulmonary hypertension (CTEPH) underwent right-heart catheterisation with haemodynamic measurements at rest and during stepwise incremental cycling exercise 60 min after receiving intravenous acetazolamide (500 mg) or saline placebo in accordance with a double-blind, randomised-controlled, crossover design. The main outcomes were the difference between pulmonary vascular resistance (PVR) and its components mean pulmonary artery pressure (mPAP), cardiac output (CO) and pulmonary arterial wedge pressure (PAWP), during exercise, assessed by a mixed linear regression analysis.

Results

A total of 24 patients (n=7 PAH, n=17 CTEPH; n=17 male) were included (mean±sd age 59±14 years). Treatment with acetazolamide compared with saline placebo showed the following average marginal effects during exercise: unchanged end-exercise power (mean difference −0.8 W, 95% CI −5.7–4.1 W; p=0.740), reduced mPAP (mean difference −1.7 mmHg, 95% CI −2.9– −0.5 mmHg; p=0.007), tendency to reduced CO (mean difference −0.3 L·min⁻¹, 95% CI −0.7–0.1 L·min⁻¹; p=0.097), unchanged PVR (mean difference −0.1 Wood units (WU), 95% CI −0.3–0.2 WU; p=0.694), unchanged PAWP (mean difference 0.0 mmHg, 95% CI −0.2–0.3 mmHg; p=0.783) and unchanged mPAP/CO slope (mean difference 0.1 WU, 95% CI −1.0–1.3 WU; p=0.839).

Conclusion

Intravenous acetazolamide was well tolerated and resulted in a significant but small decrease in mPAP, while CO, PVR and the pressure–flow slope during exercise were unchanged.

Shareable abstract

Treatment with intravenous acetazolamide during exercise is well tolerated and does not cause any significant changes in PVR but results in a significant but small decrease in mPAP, counterbalanced by a slight decrease in CO https://bit.ly/4heysmv

Introduction

Patients with pulmonary vascular disease (PVD) defined by precapillary pulmonary hypertension (PH) and diagnosed as pulmonary arterial hypertension (PAH) and chronic thromboembolic hypertension (CTEPH) typically present with insidious dyspnoea, fatigue, progressive exercise intolerance and occasionally syncope [1]. Haemodynamics during right-heart catheterisation (RHC) at rest and particularly during exercise are of prognostic value in patients with PH and individuals with dyspnoea and exercise PH may precede resting PH [2–6]. Chronically elevated pulmonary artery pressure (PAP) and vascular resistance (PVR) result in reduced pulmonary blood flow, increased strain to the right ventricle, reduced ventriculo-arterial coupling and, ultimately, right-heart failure [7]. Despite the current advances in medical therapy to alleviate symptoms and improve haemodynamics, most PVD patients continue to have a high burden of symptoms, especially during exercise, and experience reduced quality of life and a poor prognosis [1, 8]. Therefore, it remains of vital interest to investigate pharmaceutical agents that may improve pulmonary haemodynamics, particularly during exercise.

Acetazolamide is a carbonic anhydrase inhibitor, which induces metabolic acidosis through increased renal bicarbonate excretion, which in turn leads to hyperventilation and improved oxygenation, an effect that is mainly used to prevent high-altitude-related illnesses [9, 10]. Acetazolamide reduces hypoxic pulmonary vasoconstriction in humans [11], but it has also proven effective to reduce sleep-related breathing disorders in patients with cardiopulmonary disease including PVD [12–14]. Acetazolamide has recently been shown to have a further beneficial effect as an adjunct diuretic agent in patients with acute decompensated heart failure [15]. In the PH field, some in vitro and in vivo experiments in animals have suggested a direct vasodilator mechanism on the pulmonary vasculature during hypoxia, independent of renal bicarbonate excretion [16]. In a study of patients with COPD travelling to 3100 m above sea level, the hypoxia-induced rise in PAP after a night at high altitude was significantly blunted by acetazolamide prophylaxis before ascent [17]. A study investigating high-altitude pulmonary oedema found that acetazolamide had no effect on mean pulmonary artery pressure (mPAP) at 36–48 h and caused no reduction in high-altitude pulmonary oedema occurrence, suggesting that its effect in attenuating hypoxic pulmonary vasoconstriction might be short lived [18]. However, we have recently shown that intravenous acetazolamide in PVD patients does not significantly change resting haemodynamics, which points against the acute resting pulmonary vasodilator properties of acetazolamide in normoxia, but revealed a beneficial effect on hypoxia-induced PVR increase [19]. The effect of acetazolamide on exercise haemodynamics in patients with PVD has not been studied so far. Studies at altitude, however, in healthy patients and patients with COPD, revealed conflicting effects of acetazolamide on exercise performance [20, 21].

In the present analysis, therefore, we focused on the effect of acetazolamide versus placebo on pulmonary haemodynamics during exercise in patients with PVD due to PAH or CTEPH undergoing exercise RHC. The study was specifically designed to evaluate the hypothesis that acetazolamide modifies pulmonary haemodynamics, in particular, pulmonary vascular resistance (PVR), independently of the subacute effects of the drug on renal bicarbonate excretion, acid–base status and ventilation.

Methods

Study design and participants

This study was part of a larger randomised, placebo-controlled, double-blind, crossover trial assessing the acute effects of acetazolamide versus placebo on pulmonary haemodynamic parameters in patients with PVD undergoing RHC for clinical reasons [19].

The study protocol comprised supine resting haemodynamic measurements by RHC at three consecutive timepoints, as previously described: [19] a) baseline measurements, b) 60 min after administration of 500 mg of acetazolamide intravenously (Diamox, OM Pharma Suisse SA) and c) 60 min after saline placebo intravenous injections (with b) and c) in a double-blinded, randomised order). After their resting measurements at timepoints b) and c) were taken, patients underwent exercise testing in a 45° upright position on a cycle ergometer at increments of 10–20 W every 3 min until exhaustion. This analysis focuses on exercise haemodynamic measurements that have not been published thus far.

The study comprised patients referred for clinically indicated RHC fulfilling the following inclusion criteria: signed informed consent, age 20–80 years, either sex and diagnosis of precapillary PH according to 2015 guidelines (mPAP ≥25 mmHg along with a pulmonary arterial wedge pressure (PAWP) ≤15 mmHg) [22]. Patients were excluded if one of the following was present: PH due to left-heart disease (PAWP at rest >15 mmHg), at least moderate chronic obstructive or restrictive pulmonary disease (forced expiratory volume in 1 s ≤60% pred, forced vital capacity ≤60% pred), severe daytime hypercapnia (arterial carbon dioxide tension (P_aCO2) >6.5 kPa), liver disease, non-correctable electrolyte disturbances or severe chronic kidney and liver disease, pregnancy or breastfeeding, and known allergic response to acetazolamide and other known carbonic anhydrase inhibitors (methazolamide, dichlorphenamide, thiazide diuretics and sulfonamide loop diuretics) [19].

The study complies with the amended Declaration of Helsinki, was approved by the local ethical authorities (Cantonal Ethics Identifier: KEK-ZH-2016-00089) and was prior registered at ClinicalTrials.gov (NCT02755259).

Randomisation and blinding

Patients were randomised in balanced blocks of four to one of the two sequences (placebo/acetazolamide or acetazolamide/placebo) by a computer program. Participants and investigators were blinded to the administered drug until the conclusion of data analysis [19].

Assessments

Right-heart catheterisation

A Swan–Ganz catheter (Swan–Ganz CCOmbo V, Edwards Lifesciences, Irvine, CA, USA) was placed in the pulmonary artery via the right jugular vein under sonographic control. Transducers were set at the midthoracic level and zeroed to atmospheric pressure [23]. The following parameters were measured and averaged over several respiratory cycles: heart rate (HR), PAP (systolic, diastolic and mean), PAWP and right atrial pressure (RAP) [24]. Cardiac output (CO) was calculated by direct Fick using pulmonary and peripheral artery pulse oximetry after calibration with respective blood gases as previously described in order to account for rapidly changing haemodynamic parameters during exercise, and values were repeatedly measured heartbeat by heartbeat and registered in LabChart (version 8.1.16; ADInstruments). Oxygen uptake was continuously measured by a metabolic measurement unit (Geratherm SA, Germany) using standard ergospirometry methods [25, 26]. PVR was calculated as PVR=(mPAP−PAWP)/CO. The pressure–flow slope was calculated as (exercise mPAP−resting mPAP)/(exercise CO−resting CO). Measurements of haemodynamic parameters were obtained when patients were in a 45° upright position at rest and at the end of every completed 3-min increment until exhaustion.

Blood gas analysis and oximetry

Pulse oximetric oxygen saturation was continuously monitored at the fingertip and at the end of the catheter tip. A radial artery catheter was placed for continuous blood pressure monitoring and arterial blood gas analysis (aBGA) at rest and at end-exercise. Arterial and mixed venous blood gas samples were drawn from the arterial line and the distal port of the Swan–Ganz catheter, respectively, and immediately analysed (ABL90 FLEX blood gas analyser, Radiometer GmbH, Germany), as previously described [19].

Outcomes

The primary outcomes were the average marginal effect of treatment with acetazolamide versus placebo on mPAP, CO, PAWP, and the herewith calculated PVR and mPAP/CO slope during stepwise incremental cycling exercise.

Secondary outcomes included the average marginal effects of treatment with acetazolamide adjusted by clinically relevant baseline characteristics (PH class, sex, age) and aBGA values (arterial oxygen tension (P_aO2) and P_aCO2). Further outcomes included the change in haemodynamic parameters and aBGA values during exercise.

Statistical analysis

Mixed linear model regressions were performed in our intention-to-treat population on the above-mentioned haemodynamic variable outcomes with adjustment for treatment (acetazolamide versus placebo), order of administration plus the interaction term of treatment and order (to account for carry-over treatment effects), and exercise condition (rest versus end-exercise) or exercise performance (in W) as fixed effects. Individual subjects were modelled as a random factor. In a second step, regressions were adjusted for relevant baseline values and measured aBGA values as additional fixed effects. To account for potential premature fatigue limitation during the second exercise period, end-exercise haemodynamic values in each individual subject were compared at isotime (thus at the lower maximal work rate of the two exercise tests) and are henceforth termed end-exercise.

Single missing data points were not replaced or imputed, and implausible outliers were omitted. The regression data are displayed as average marginal effects of each predictor. Continuous variables are summarised as mean±sd. Statistical significance was assumed when 95% confidence intervals did not overlap 0 with a corresponding p-value of <0.05. All statistical analyses were performed using R Studio (version 2023.03.1+446, R Studio, Inc., San Francisco, CA, USA).

Results

Screening and baseline characteristics

In total, 24 patients with PVD were included and randomised in this study. The study screening process has been described previously (figure 1) [19]. Most of the participating patients were male (71%), 29% had PAH (57% male) and 71% had CTEPH (76% male), most were in World Health Organization functional class II (63%) and the mean±sd age was 59±14 years (table 1). Baseline haemodynamics were mPAP 36±11 mmHg, PVR 5.0±2.7 Wood units (WU), CO 5.5±1.8 L·min⁻¹ and PAWP 11±2 mmHg.

FIGURE 1.

Patient flow. RHC: right-heart catheterisation; PH: pulmonary hypertension.

TABLE 1.

Baseline characteristics

Patients with pulmonary vascular disease, n	24
Female	7 (29)
WHO functional class
I	2 (8)
II	15 (63)
III	7 (29)
Idiopathic pulmonary arterial hypertension, WHO group 1.1	7 (29)
Chronic thromboembolic pulmonary hypertension, WHO group 4	17 (71)
Pulmonary hypertension medications
Endothelin receptor antagonists	10 (42)
Phosphodiesterase-5 inhibitors	6 (25)
Prostacyclin receptor agonists	2 (8)
Calcium channel antagonists	4 (17)
>1 pulmonary hypertension medication class	6 (25)
Other relevant medications
Diuretics	10 (42)
Anticoagulants	20 (83)
Age, years	59±14
Body mass index, kg·m−2	27.8±4.5
Body surface area, m2	2.0±0.2
6-min walk distance, m	531±111

Data are presented as n (%) or mean±sd unless otherwise stated. WHO: World Health Organization.

Primary outcomes

There was no significant difference in maximum exercise power between acetazolamide (75±35 W) and placebo (76±37 W). When receiving acetazolamide, patients had a PVR of 5.1±3.2 WU at rest and 5.2±2.2 WU at end-exercise, whereas after treatment with saline placebo, patients’ PVR was 5.2±3.1 WU at rest and 5.3±2.6 WU at end-exercise (table 2). A mixed model regression analysis (table 3) showed no significant difference in the average marginal effect on PVR by treatment with acetazolamide compared with saline placebo during exercise (−0.1 WU, 95% CI −0.3–0.2 WU; p=0.694). PVR did not significantly change with increasing exercise power in W (0.001 WU, 95% CI −0.004–0.006 WU; p=0.711). Notably, there was no difference in PVR attributable to the order of treatment (0.1 WU, 95% CI −2.1–2.4 WU; p=0.903). Analysis of PVR components during exercise revealed a significant decrease in mPAP after treatment with acetazolamide (−1.7 mmHg, 95% CI −2.9– 0.5 mmHg; p=0.007) and a trend towards reduction of CO (−0.3 L·min⁻¹, 95% CI −0.7–0.1 L·min⁻¹; p=0.097), whereas PAWP (0.0 mmHg, 95% CI −0.2–0.3 mmHg; p=0.783) was no different from that after treatment with saline placebo. mPAP and CO showed a marked increase per W during exercise, while PAWP also showed a minor, but significant, increase (figure 2). Total pulmonary resistance (mPAP/CO), unlike PVR, showed a decrease during exercise but no association with acetazolamide treatment (figure 2). The pressure–flow slopes did not differ significantly (0.1 WU, 95% CI −1.0–1.3 WU; p=0.839) (figure 2). There were no order effects on PVR, mPAP, CO and PAWP.

TABLE 2.

Exercise, haemodynamic and blood gas data at rest and at end-exercise after treatment with placebo or acetazolamide

Treatment	Placebo		Acetazolamide
Exercise condition	Rest	End-exercise	Rest	End-exercise
Exercise data
Exercise power, W		76±37		75±35
Main haemodynamic parameters of interest
Pulmonary vascular resistance, WU	5.2±3.1	5.3±2.6	5.1±3.2	5.2±2.2
Mean pulmonary arterial pressure, mmHg	37±12	64±14	38±12	63±13*
Cardiac output, L·min−1	5.9±1.7	11.1±3.8	7.0±3.0	11.0±3.4*
Pulmonary arterial wedge pressure, mmHg	11±3	12±3	11±3	12±3
Pressure–flow slope#, WU		5.4±4.6		5.4±4.1
Additional haemodynamic parameters
Heart rate, beats per min	76±10	119±17	77±11	119±19
Systolic blood pressure, mmHg	132±17	188±51	130±17	180±32
Mean blood pressure, mmHg	92±11	122±57	89±12	110±15
Diastolic blood pressure, mmHg	72±10	90±61	70±9	77±9
Systolic pulmonary arterial pressure, mmHg	57±21	101±19	59±21	98±19*
Mean pulmonary arterial pressure, mmHg	37±12	64±14	38±12	63±13*
Diastolic pulmonary arterial pressure, mmHg	24±9	39±10	25±9	37±12*
Right atrial pressure, mmHg	6±3	15±6	7±4	15±6
Oxygen saturation, %	89±19	90±5	93±3	90±5*
Mixed venous oxygen saturation, %	62±8	34±13	62±7	34±12
Blood gas parameters
Haemoglobin, g·L−1	143±12	145±13	144±14	149±13
Lactate, mmol·L−1	1.1±0.5	2.2±1.3	1.1±0.4	1.7±0.6*
HbO2, %	91±3	89±6	91±4	89±5
Arterial pH	7.420±0.030	7.401±0.044	7.404±0.023	7.388±0.028*
PaO2, kPa	8.6±1.6	8.2±1.7	9.0±1.8	8.4±1.7
PaCO2, kPa	4.4±0.5	4.3±0.6	4.4±0.5	4.5±0.6
HbvO2, %	63±7	33±16	64±6	34±13
PvO2, kPa	4.6±0.5	2.8±0.5	4.7±0.4	3.1±0.7*
PvCO2, kPa	4.96±0.49	6.01±0.60	5.08±0.42	5.85±0.89
CaO2, mL·L−1	177±17	117±86	178±18	149±70
CvO2, mL·L−1	122±19	35±39	125±16	42±38
HCO3−, mmol·L−1	22.5±1.5	21.5±1.4	21.8±1.4	21.1±1.4*

Data are presented as mean±sd. WU: Wood unit; Hb_O2: arterial oxygenated haemoglobin; P_aO2: arterial oxygen tension; P_aCO2: arterial carbon dioxide tension; Hb_vO2: mixed venous oxygenated haemoglobin; P_vO2: mixed venous oxygen tension; P_vCO2: mixed venous carbon dioxide tension; C_aO2: arterial oxygen content; C_vO2: mixed venous oxygen content; HCO₃⁻: bicarbonate. ^#: change in mean pulmonary arterial pressure divided by change in cardiac output. *: p<0.05.

TABLE 3.

Mixed regression models of haemodynamic variables adjusted for treatment, order of treatment and exercise power or condition

Haemodynamic variable	Independent variable	Average marginal effect (95% CI)	p-value
Exercise power, W	Order	−13.333 (−41.817–15.151)	0.359
	Acetazolamide	−0.833 (−5.746–4.079)	0.740
PVR, WU	Exercise power, W	0.001 (−0.004–0.006)	0.711
	Order	0.140 (−2.094–2.373)	0.903
	Acetazolamide	−0.058 (−0.347–0.231)	0.694
mPAP, mmHg	Exercise power, W	0.267 (0.246–0.287)	<0.001*
	Order	−2.237 (−14.021–9.546)	0.710
	Acetazolamide	−1.698 (−2.928– −0.468)	0.007*
CO, L·min−1	Exercise power, W	0.056 (0.049–0.062)	<0.001*
	Order	−0.273 (−1.607–1.060)	0.688
	Acetazolamide	−0.333 (−0.727–0.061)	0.097
PAWP, mmHg	Exercise power, W	0.010 (0.006–0.014)	<0.001*
	Order	0.105 (−1.889–2.099)	0.918
	Acetazolamide	0.036 (−0.223– −0.295)	0.783
TPR, WU	Exercise power, W	−0.006 (−0.012– −0.001)	0.029*
	Order	−0.113 (−2.514–2.287)	0.926
	Acetazolamide	−0.108 (−0.447–0.231)	0.533
Pressure–flow slope#, WU	End-exercise	0.453 (−0.675–1.582)	0.431
	Order	−0.977 (−3.877–1.923)	0.509
	Acetazolamide	0.119 (−1.029–1.266)	0.839

“Order” indicates treatment with acetazolamide coming first. PVR: pulmonary vascular resistance; WU: Wood unit; mPAP: mean pulmonary arterial pressure; CO: cardiac output; PAWP: pulmonary arterial wedge pressure; TPR: total pulmonary resistance (mPAP/CO). ^#: change in mean pulmonary arterial pressure divided by change in cardiac output. *: p<0.05.

FIGURE 2.

a–e) Haemodynamic variables as data points depending on exercise power until end-exercise; f) Pressure–flow curve (the difference in mean pulmonary arterial pressure (mPAP) between end-exercise and rest divided by the difference in cardiac output (CO) between end-exercise and rest). The data points are colour-coded for treatment (red indicating saline placebo, turquoise indicating acetazolamide). A linear trend line colour-coded for each treatment has been laid over the corresponding data. The corresponding grey areas of each trend line indicate the range uncertainty depending on data density and distribution. p-values from the respective mixed model regression are displayed for each panel. a) Pulmonary vascular resistance (PVR). b) mPAP. c) CO. d) Pulmonary arterial wedge pressure (PAWP). e) Total pulmonary resistance (TPR).

Haemodynamic outcomes adjusted for PH class, sex and age

Adjustment for clinically relevant baseline predictors did not relevantly change the above-mentioned haemodynamic associations. In our study population, female patients showed a higher mPAP (14.1 mmHg, 95% CI 0.4–27.9 mmHg; p=0.044) and higher PVR (3.7 WU, 95% CI 1.2–6.2 WU; p=0.003) but no difference in CO or PAWP during exercise. Compared with those with PAH, patients with CTEPH showed higher PVR (0.9 WU, 95% CI 0.1–1.7 WU; p=0.028) and a trend towards higher mPAP (4.3 mmHg, 95% CI −0.4–9.0 mmHg; p=0.076), albeit with lower CO (−0.8 L·min⁻¹, 95% CI −1.3– −0.2 L·min⁻¹; p=0.004), while showing no significant difference in PAWP. Age was not a significant predictor for any of the haemodynamic parameters (table 4).

TABLE 4.

Mixed regression models of haemodynamic variables adjusted for treatment, order of treatment, exercise intensity and clinically important baseline values

Haemodynamic variable	Independent variable	Average marginal effect (95% CI)	p-value
PVR, WU	Exercise power, W	0.001 (−0.004–0.006)	0.662
	Order	0.464 (−1.278–2.206)	0.602
	Acetazolamide	−0.055 (−0.343–0.234)	0.711
	Age, years	0.007 (−0.079–0.092)	0.878
	Female	3.695 (1.220–6.171)	0.003*
	CTEPH versus PAH	0.912 (0.1000–1.723)	0.028*
mPAP, mmHg	Exercise power, W	0.267 (0.247–0.287)	<0.001*
	Order	−2.279 (−12.950–8.392)	0.676
	Acetazolamide	−1.697 (−2.927– −0.468)	0.007*
	Age, years	0.146 (−0.369–0.662)	0.578
	Female	14.149 (0.366–27.932)	0.044*
	CTEPH versus PAH	4.259 (−0.446–8.963)	0.076
CO, L·min−1	Exercise power, W	0.056 (0.049–0.062)	<0.001*
	Order	−0.439 (−1.640–0.762)	0.474
	Acetazolamide	−0.337 (−0.730–0.057)	0.094
	Age, years	0.054 (−0.005–0.112)	0.071
	Female	0.027 (−1.539–1.593)	0.973
	CTEPH versus PAH	−0.773 (−1.304– −0.242)	0.004*
PAWP, mmHg	Exercise power, W	0.010 (0.006–0.014)	<0.001*
	Order	−0.057 (−2.028–1.914)	0.955
	Acetazolamide	0.036 (−0.223–0.295)	0.783
	Age, years	0.045 (−0.051–0.141)	0.360
	Female	−1.265 (−4.064–1.533)	0.376
	CTEPH versus PAH	−0.477 (−1.395–0.441)	0.308
TPR, WU	Exercise power, W	−0.006 (−0.011–0.000)	0.036*
	Order	−0.039 (−2.132–2.054)	0.971
	Acetazolamide	−0.106 (−0.445–0.233)	0.540
	Age, years	−0.014 (−0.116–0.087)	0.781
	Female	2.505 (−0.203–5.212)	0.070
	CTEPH versus PAH	1.172 (0.249–2.096)	0.013*

“Order” indicates treatment with acetazolamide coming first. PVR: pulmonary vascular resistance; WU: Wood unit; mPAP: mean pulmonary arterial pressure; CO: cardiac output; PAWP: pulmonary arterial wedge pressure; TPR: total pulmonary resistance; CTEPH: chronic thromboembolic pulmonary hypertension; PAH: pulmonary arterial hypertension. *: p<0.05.

Exploratory analyses on influence of arterial blood gas values and side effects

P_aO2, mixed venous oxygen tension (P_vO2) and pH were reduced at end-exercise, while lactate levels were elevated under placebo, indicating exercise-induced lactic acidosis and hypoxaemia. Treatment with acetazolamide resulted in lower pH and bicarbonate, indicative of metabolic acidosis, while P_aCO2 and mixed venous carbon dioxide tension, including rest and end-exercise, were unchanged. We detected an order effect on pH when acetazolamide was administered first but not on other aBGA values (table 5). Adjustment of haemodynamic variables during exercise for P_aO2 revealed a negative association with PVR and mPAP (trend) but no association with CO or PAWP. Similarly, adjustment for P_aCO2 showed a marked, negative association with PVR, mPAP and CO, while no change in PAWP occurred (table 6 and figure 3).

TABLE 5.

Mixed regression models of arterial blood gas values adjusted for treatment, order of treatment and exercise condition

aBGA parameter	Independent predictor variable	Average marginal effect (95% CI)	p-value
PaO2, kPa	End-exercise	−0.748 (−1.160– −0.336)	<0.001*
	Order	0.702 (−0.498–1.902)	0.252
	Acetazolamide	0.309 (−0.093–0.711)	0.132
PaCO2, kPa	End-exercise	−0.007 (−0.108–0.095)	0.899
	Order	0.236 (−0.156–0.628)	0.237
	Acetazolamide	0.060 (−0.040–0.159)	0.239
Arterial pH	End-exercise	−0.015 (−0.020– −0.010)	<0.001*
	Order	−0.023 (−0.043– −0.003)	0.022*
	Acetazolamide	−0.014 (−0.020– −0.009)	<0.001*
Lactate, mmol·L−1	End-exercise	0.739 (0.493–0.985)	<0.001*
	Order	−0.587 (−0.863– −0.312)	<0.001*
	Acetazolamide	−0.228 (−0.473–0.017)	0.069
PvO2, kPa	End-exercise	−1.671 (−1.844– −1.498)	<0.001*
	Order	0.000 (−0.348–0.348)	0.999
	Acetazolamide	0.186 (0.022–0.349)	0.026*
PvCO2, kPa	End-exercise	0.909 (0.675–1.144)	<0.001*
	Order	0.175 (−0.171–0.522)	0.321
	Acetazolamide	0.023 (−0.202–0.248)	0.843

“Order” indicates treatment with acetazolamide coming first. aBGA: arterial blood gas analysis; P_aO2: arterial oxygen tension; P_aCO2: arterial carbon monoxide tension; P_vO2: mixed venous oxygen tension; P_vCO2: mixed venous carbon dioxide tension. *: p<0.05.

TABLE 6.

Mixed regression models of haemodynamic variables adjusted for treatment, order of treatment, exercise condition and arterial blood gases

Haemodynamic variable	Independent variable	Average marginal effect (95% CI)	p-value
PVR, WU	End-exercise	−0.230 (−0.844–0.388)	0.462
	Order	0.704 (−0.950–2.357)	0.404
	Acetazolamide	0.367 (−0.146–0.879)	0.161
	PaO2, kPa	−0.469 (−0.789– −0.149)	0.004*
	PaCO2, kPa	−2.515 (−3.715– −1.315)	<0.001*
mPAP, mmHg	End-exercise	26.441 (23.775–29.107)	<0.001*
	Order	3.014 (−7.178–3.225)	0.563
	Acetazolamide	0.042 (−2.276–2.359)	0.972
	PaO2, kPa	−1.378 (−2.862–0.106)	0.069
	PaCO2, kPa	−7.326 (−13.037– −1.615)	0.012*
CO, L·min−1	End-exercise	5.504 (4.592–6.417)	<0.001*
	Order	−0.006 (−1.770–1.758)	0.995
	Acetazolamide	−0.355 (−1.175–0.465)	0.397
	PaO2, kPa	−0.254 (−0.681– −0.173)	0.243
	PaCO2, kPa	1.822 (0.371–3.273)	0.014*
PAWP, mmHg	End-exercise	0.793 (−0.076–1.662)	0.074*
	Order	0.536 (−1.479–2.551)	0.602
	Acetazolamide	0.411 (−0.357–1.179)	0.294
	PaO2, kPa	−0.195 (−0.645–0.254)	0.394
	PaCO2, kPa	−1.311 (−2.982–0.360)	0.125
TPR, WU	End-exercise	−0.943 (−1.653– −0.233)	0.009*
	Order	1.119 (−0.764–3.002)	0.244
	Acetazolamide	0.422 (−0.194–1.038)	0.180
	PaO2, kPa	−0.482 (−0.844– −0.119)	0.009*
	PaCO2, kPa	−2.444 (−3.753– −1.134)	<0.001*

“Order” indicates treatment with acetazolamide coming first. PVR: pulmonary vascular resistance; WU: Wood units; mPAP: mean pulmonary arterial pressure; CO: cardiac output; PAWP: pulmonary arterial wedge pressure; TPR: total pulmonary resistance; P_aO2: arterial oxygen tension; P_aCO2: arterial carbon dioxide tension. *: p<0.05.

FIGURE 3.

Pulmonary vascular resistance (PVR) is shown as data points depending on a) the arterial oxygen tension (P_aO2) and b) the arterial carbon dioxide tensions (P_aCO2), colour-coded for treatment (red indicating saline placebo, turquoise indicating acetazolamide), with the left panel showing data at rest and the right panel showing data at end-exercise. A linear trend line colour-coded for each treatment has been laid over the corresponding data. The corresponding grey areas of each trend line indicate the range uncertainty depending on data density and distribution. WU: Wood unit.

No patient reported any typical carbonic anhydrase inhibitor side-effects [19].

Cardiometabolic measurements

Maximum oxygen uptake (V′_O2max) did not show any significant difference between the treatment arms. The minute ventilation (V′_E) to carbon dioxide output (V′_CO2) ratio tended to be increased after treatment with acetazolamide versus placebo, while the end-expiratory absolute end-tidal carbon dioxide tension (P_ETCO2) and fractional carbon dioxide (F_ECO2) tended to be decreased. Patients treated with acetazolamide showed increased subjective breathing sensation measured with the Borg CR10 scale. All other cardiometabolic parameters were unchanged (supplementary table 1).

Discussion

This is the first randomised, double-blind, placebo-controlled crossover study investigating the acute effects of acetazolamide on pulmonary haemodynamics during exercise in patients with PVD assessed by RHC. The present study was explicitly designed to investigate the acute effects on pulmonary haemodynamics previously postulated, which were assumed to be independent of the later-onset renal effects [11, 16]. The main finding of this study is that 60 min after administration of intravenous acetazolamide, versus placebo, the mPAP during exercise was decreased by an average of −1.7 mmHg, whereas the PVR remained unchanged, most probably as we found a tendency towards a concordantly decreased CO. Notably, there were no order effects in any of the haemodynamic variables, indicating an absence of significant carry-over effects of the study drug in the crossover model. PH class, age and sex did not influence these results.

Effect of acetazolamide on exercise haemodynamics

A steep increase in mPAP during exercise along with an insufficient rise in CO due to an increased PVR is one of the main mechanisms leading to exercise intolerance in patients with PVD. Therefore, identifying drugs that positively modulate pulmonary haemodynamics during exercise remains of vital importance for such patients [2–6]. Overall, previous data are conflicting on the effect of acetazolamide on pulmonary haemodynamics and especially during exercise, although the body of literature remains very slim [11, 17, 27]. In one of our previous studies in COPD patients travelling to altitude, we found a reduced PAP by echocardiography at rest after prophylaxis with acetazolamide compared with placebo assessed after one night at 3100 m [17]. In accordance with the aforementioned study (but in the present study, in patients with PVD and during exercise), we observed a statistically significant decrease in mPAP during exercise after acetazolamide versus placebo. However, as the CO tended to be decreased as well, acetazolamide versus placebo did not induce a change of PVR during exercise and the pressure–flow slope did not differ. In accordance with the previously published results from the same PVD collective, which focused on pulmonary haemodynamics at supine rest 60 min after acetazolamide versus placebo, in the present study, acetazolamide showed no effect on PVR at rest in a 45° upright position before cycling [19]. However, the previously published results showed that the PVR increase was smaller in response to hypoxia with acetazolamide versus placebo, which is in line with other studies that found a reduced hypoxic pulmonary vasoconstriction with acetazolamide [11, 19].

Our study proposes that, as PVR remained unchanged, acute treatment with acetazolamide does not induce a direct vasodilator effect in patients with PVD at rest, contrary to results obtained in animal studies [28, 29]. Whether a vasodilator effect in PVD patients is present during exercise is debatable. On the whole, the lower mPAP, which, however, is counterbalanced by a reduced CO in the present study and the unchanged pressure–flow curve over the course of exercise, probably suggests a neutral effect of acetazolamide on pulmonary vasculature. As therapeutical options are severely limited for PVD patients, the reduction in mPAP could still present an interesting focal point for future (long term) haemodynamic investigation. Why the CO tended to be lower during exercise 60 min after the administration of intravenous acetazolamide versus saline placebo remains unclear. A potential mechanism could be a decreased preload to the heart due to aa diuretic and/or vasodilatory effect of acetazolamide with consecutively reduced stroke volume and CO. As the surrogate measures of preload RAP and PAWP remained unchanged after treatment with acetazolamide, we assume substantial changes in preload to be unlikely. An alternative explanation could be a potential negative inotropic effect due to the acetazolamide-induced acidosis and direct inhibition of myocardial carbonic anhydrase. On a cardiometabolic level, V′_O2max was unchanged following treatment with acetazolamide, although V′_E/V′_CO2, P_ETCO2 and F_ECO2 showed a trend towards being reduced, indicating the beginning of the ventilatory action of acetazolamide, which was expected not to be fully present in accordance with the chosen study design.

The effect of carbonic anhydrase inhibitors on exercise performance has been mostly studied in healthy participants in hypoxic environments. Some studies revealed decreased performance due to increased dyspnoea perception by acidosis-induced hyperventilation, but pulmonary haemodynamics were not measured in these studies [30–32]. Further studies in mountaineers, one of which measured pulmonary artery pressure using Doppler echocardiography, showed no difference in peak exercise capacity in hypoxic environments with acetazolamide [21, 33], and a study in COPD patients receiving preventive acetazolamide likewise showed no difference in exercise performance [20]. Two studies in healthy mountaineers even suggested an age-related, compromised exercise capacity and acclimatisation after treatment with acetazolamide despite better oxygenation, although pulmonary haemodynamics were not directly investigated [32, 34]. Two reviews of the effects of acetazolamide on exercise in healthy individuals concluded that a reduction in maximum exercise capacity constituted a common finding at sea level, while the effects of acetazolamide under hypoxic conditions are multifactorial and unpredictable, and potentially more beneficial at higher altitudes [30, 31]. A trial of daily therapy with acetazolamide over 5 weeks concluded that the therapy did not increase 6-min walking distance as a measure of exercise capacity in patients with precapillary PH despite better oxygenation [27].

Acetazolamide has various other uses, such as in the prevention of altitude sickness [9, 10], benefits in high loop gain sleep-related breathing disorders occurring in patients with PVD [12–14], and especially in the treatment of decompensated heart failure when used as an adjunct diuretic [15]. This may be of interest for PVD patients who also have decompensated right-heart failure. The present study shows that the acute use of acetazolamide does not impair haemodynamics at rest and during exercise in patients with PVD and can be considered safe in this respect, especially as our patient collective did not experience any side-effects that could be attributed to carbonic anhydrase inhibitors [19]. Hence, we believe that it may be of interest to investigate the effect of acetazolamide on right-sided heart failure.

The effect of acetazolamide on arterial blood gases

The present study was designed to detect the vasodilatory effects of acetazolamide independently of renal bicarbonate excretion, due to the time and diuresis dependent nature of bicarbonate elimination. Our data show that 60 min after treatment with acetazolamide, there was a slight but significant drop in pH and bicarbonate levels compared with treatment with saline placebo, suggesting that bicarbonate excretion has already commenced by the time of measurement. As the pH levels are still within a normal physiological range, we presume the effect of bicarbonate loss on the observed haemodynamic effects to be minor. P_aO2 and P_aCO2 remained unchanged after treatment with acetazolamide. Conversely, we observed a significantly higher P_vO2 – this effect is contrary to the trend towards lower CO, which typically would result in lower P_vO2 due to higher tissue oxygen extraction. Why this was the case is unclear. Adjustment of haemodynamic variables for P_aO2 expectedly showed a strong negative association with both PVR and mPAP. It is known that PVR and thus mPAP increase with lower P_aO2 as a result of hypoxic pulmonary vasoconstriction [35, 36] and, conversely, that supplemental oxygen results in lower PVR and mPAP in patients with PVD, possibly due to partially reversible hypoxic vasoconstriction [37, 38]. It is notable, however, that P_aCO2 showed a negative association with PVR and mPAP independently of treatment with acetazolamide or placebo, which is in stark contrast to current knowledge that hypercapnia causes pulmonary vasoconstriction in humans and other mammals. We believe that this finding is an expression of the fact that patients with more severe PVD tend to hyperventilate more. This inverse relationship is present at rest and at end-exercise (figure 3b), and the absolute values of P_aCO2 are at the lower limit of normal or below and thus outside an expected hypercapnic stimulus range. Despite the statistical correction of the interaction of P_aO2 and P_aCO2, these findings confirm that P_aO2 and P_aCO2 are highly co-dependent and that hypoxic pulmonary vasoconstriction appears to override changes in P_aCO2. In summary, a true reversal of vasomotor reaction to hypercapnia is highly unlikely. Lastly, the lack of association between PAWP and P_aO2 and P_aCO2 in the absence of congestive heart failure is indicative that changes in blood gases primarily result in precapillary haemodynamic changes and do not affect postcapillary haemodynamics, independently of exercise [37, 39].

Limitations

Our study has the following limitations. Our sample of 24 patients was small, but still remarkable considering the rarity of PVD and the complexity of the study design. For ethical reasons, due to the invasive nature (RHC) of our study, we deliberately opted for a crossover design, which allowed the study interventions to be performed in one session and provided some statistical advantages. As a result, a long-term washout period could not be provided, although we performed statistical analysis to detect potential carry-over effects pertinent to the study drug being administered in order before placebo. Regarding these, we detected an order effect on the build-up of lactate and arterial pH, which was elevated in the second exercise period, most likely reflective of exercise-induced lactic acidosis. Importantly, however, our primary and secondary haemodynamic outcomes did not show any order effects. Although statistical adjustment cannot fully eliminate carry-over effects, we find it unlikely that significant undetected order effects affected the outcome of the present study. Finally, patients with CTEPH and female patients displayed worse pulmonary haemodynamics, with higher PVR and mPAP, and patients with CTEPH additionally showed a reduced CO. We believe that this is because the women in our study group predominantly had more severe PAH, while patients with CTEPH demonstrate more impaired haemodynamics in general. Unfortunately, owing to the rarity of the disease, selection bias cannot be fully excluded. These circumstances are mitigated by the fact that in our crossover design each patient was their own control, allowing for adjustment as a random effect in the mixed model, as interindividual variance carries greater weight in smaller sample sizes, and increasing the statistical power of our study.

Conclusion

Treatment with intravenous acetazolamide during exercise did not cause any significant changes in PVR but resulted in a significant but small decrease in mPAP, possibly due to a slight decrease in CO, while the pressure–flow slope remained unchanged. In total, the haemodynamic effect of acetazolamide on pulmonary vasculature was probably neutral, although the reduction mPAP remains of interest, considering the scarce therapeutical options in PVD patients. Furthermore, acetazolamide was well tolerated by the patients in our sample and appears to be safe to use in patients with PVD. Finally, as acetazolamide provides benefits in various clinical settings, including in acute left-sided heart failure, studying its use in right-sided heart failure could be warranted.

Supplementary material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material 00782-2024.SUPPLEMENT ^{(443KB, pdf)}

Data availability

The data that support the findings of this study are not openly available but are available from the corresponding author upon reasonable request.