Nocturnal hypoxemia severity influences the effect of CPAP therapy on renal renin–angiotensin–aldosterone system activity in humans with obstructive sleep apnea

Abstract

Study Objectives

Nocturnal hypoxemia (NH) in obstructive sleep apnea (OSA) is associated with renal renin–angiotensin–aldosterone system (RAAS) up-regulation and loss of kidney function. Continuous positive airway pressure (CPAP) therapy is associated with RAAS down-regulation, though the impact of NH severity remains unknown. We sought to determine whether NH severity alters the effect of CPAP on renal hemodynamics and RAAS activity in humans.

Methods

Thirty sodium-replete, otherwise healthy, OSA participants (oxygen desaturation index ≥ 15 h⁻¹) with NH (SpO₂ < 90% ≥ 12%/night) were studied pre- and post-CPAP (>4 h/night∙4 weeks). NH severity was characterized as moderate (mean SpO₂[MSpO₂] ≥ 90%; N = 15) or severe (MSpO₂ < 90%; N = 15). Glomerular filtration rate (GFR), renal plasma flow (RPF), and filtration fraction (FF) were measured at baseline and in response to angiotensin-II (3 ng/kg/min∙30 min, 6 ng/kg/min∙30 min), a marker of RAAS activity.

Results

Pre-CPAP, baseline renal hemodynamics did not differ by NH severity. Pre-CPAP, severe NH participants demonstrated blunted GFR (Δ30 min, −9 ± 4 vs 1 ± 3 mL/min, p = 0.021; Δ60 min, −5 ± 5 vs 8 ± 5 mL/min, p = 0.017) and RPF (Δ30 min, −165 ± 13 vs −93 ± 19 mL/min, p = 0.003; Δ60 min, −208 ± 18 vs −112 ± 22 mL/min, p = 0.001; moderate vs severe) responses to angiotensin-II. Post-CPAP, severe NH participants demonstrated maintained GFR (112 ± 5 vs 108 ± 3 mL/min, p = 0.9), increased RPF (664 ± 35 vs 745 ± 34 mL/min, p = 0.009), reduced FF (17.6 ± 1.4 vs 14.9 ± 0.6%, p = 0.009), and augmented RPF responses to Angiotensin-II (Δ30 min, −93 ± 19 vs −138 ± 16 mL/min, p = 0.009; Δ60 min, −112 ± 22 vs −175 ± 20 mL/min, p = 0.001; pre- vs post-CPAP), while moderate participants were unchanged.

Conclusions

Correction of severe, but not moderate, NH with CPAP therapy was associated with improved renal hemodynamics and decreased renal RAAS activity in humans with OSA.

nocturnal hypoxemia, obstructive sleep apnea, CPAP, renin–angiotensin–aldosterone system, renal hemodynamics

Statement of Significance

Nocturnal hypoxemia is a common consequence of untreated obstructive sleep apnea which has been associated with renin–angiotensin–aldosterone system activation and loss of kidney function. OSA participants with severe nocturnal hypoxemia demonstrated improved renal hemodynamics and renin–angiotensin–aldosterone system activity as indicated by the reduction in glomerular pressure and augmented responses to angiotensin-II challenge post-CPAP therapy, while moderately hypoxemic participants were unchanged. This study highlights the importance of nocturnal hypoxemia in OSA-mediated kidney and vascular disease and particularly the potential greater renal protective benefit of CPAP therapy in those patients with more severe nocturnal hypoxemia. In the era of precision-based medicine, the impact of identifying and targeting this OSA phenotype with severe nocturnal hypoxemia deserves further study.

Introduction

Nocturnal hypoxemia (NH) due to obstructive sleep apnea (OSA) is common in chronic kidney disease (CKD) [1–3]. Severe NH is associated with loss of kidney function [4, 5], hypertension [6], and the risk of death in patients with kidney disease [7]. There is growing evidence that intra-renal tissue hypoxia is the final common pathway in the development of end-stage kidney disease (ESKD) [8]. Animal and human studies have demonstrated that the chronic intermittent hypoxia (CIH) caused by OSA is associated with glomerular hyperfiltration [9–11] and activation of the renal renin–angiotensin–aldosterone system (RAAS) [9, 12–14], which predisposes to the development of both kidney [15, 16] and cardiovascular disease [16–21]. We have previously demonstrated that NH severity influences the magnitude of renal RAAS activation independent of obesity in patients with OSA [22].

Continuous positive airway pressure (CPAP) therapy is an effective treatment for OSA [23]. CPAP therapy is also associated with down-regulation of renal RAAS activity [9] and attenuation of glomerular hyperfiltration [9, 11], an important and early marker of increased risk of death, cardiovascular, and pulmonary disease [24]. However, the extent to which the severity of NH influences the effect of CPAP on these outcomes remains unknown.

While there is no direct method for assessing intra-renal RAAS activity in humans, the renal and vascular RAAS may be evaluated indirectly by assessing the renal and systemic hemodynamic responses to a physiologic stressor from an angiotensin-II (AngII) infusion, after the RAAS has been suppressed by high salt intake [25–28]. The basis for distinction between the renal and vascular RAAS is due to differences in the local tissue RAAS [29], whereby the renal hemodynamic response to AngII infusion reflects renal RAAS activity [25–27] and the blood pressure (BP) response to AngII infusion reflects vascular RAAS activity [28]. In a study designed to evaluate the effect of CPAP on kidney function, we expanded recruitment to examine whether NH severity impacts the effect of CPAP therapy on renal and systemic hemodynamics and RAAS activity by measuring the response to AngII, a validated surrogate measure of RAAS activity [9, 25–28]. We hypothesized that NH severity influences the effect of CPAP on renal and systemic hemodynamics and the RAAS in OSA patients with normal kidney function.

Methods

Study participants

Eligible participants were 18–70 years old with moderate-severe OSA and significant intermittent NH (defined later) [6]. Participants were referred for suspected OSA to the Foothills Medical Centre (FMC) Sleep Centre from the community patients and respiratory homecare companies (Healthy Heart Sleep Company, Dream Sleep Respiratory Services Ltd, and RANA Respiratory Care Company) in Calgary, AB, Canada, between June 2011 and June 2014. All participants underwent a medical history, physical examination, and laboratory screening. Exclusion criteria included cardiovascular, cerebrovascular, lung, and kidney disease, uncontrolled or resistant hypertension (BP >140/90 despite maximal use of three or more antihypertensive agents of different classes [30]), diabetes mellitus, current treatment for OSA, current smoking, pregnancy, use of non-steroidal anti-inflammatory medications or exogenous sex hormones. The study protocol was approved by the Conjoint Health Research Ethics Board at the University of Calgary. Written informed consent was obtained from all study participants in accordance with the Declaration of Helsinki.

OSA and nocturnal hypoxemia

Participants performed an unattended, home sleep apnea test (HSAT; Remmers Sleep Recorder [RSR] Model 4.2, Saga Tech Electronic, Calgary, AB, Canada) [31, 32]. The monitor consists of an oximeter to record oxyhemoglobin saturation (SpO₂), a pressure transducer to record nasal airflow, a microphone to record snoring, and a body position sensor. The oximeter provides the data for an automated scoring algorithm, which calculates the oxygen desaturation index (ODI) based on the number of episodes of oxyhemoglobin desaturation ≥4% per hour of monitoring. Nocturnal oxygen saturation was sampled at 1 Hz. The RSR has been validated by comparison to attended polysomnography [31, 32].

Sleep apnea was defined as an ODI ≥15 h⁻¹ as this reflects moderate-severe sleep apnea which is likely to be clinically significant [31, 32]. The RSR has a sensitivity of 98% and specificity of 88% for an ODI ≥15 h⁻¹ [31]. Portable monitoring was performed following current guidelines and recommendations [33]. The data were reviewed by a sleep medicine physician (PJH) who confirmed that the estimated ODI was accurate and diagnostic of OSA. Nocturnal hypoxemia was defined as SaO₂ <90% for ≥12% of the duration of nocturnal monitoring, which has been used previously [6]. To investigate the impact of nocturnal hypoxemia severity on renal RAAS activity, we categorized OSA subjects into two groups based on a priori criteria related to their mean SpO₂ during the overnight monitoring test: severe nocturnal hypoxemia was defined as mean SpO₂ <90% while moderate hypoxemia was defined as mean SpO₂ ≥90%. Figure 1 shows representative HSAT recordings from two patients, one with moderate NH and the other with severe NH. Following review of the HSAT by a sleep medicine physician (PJH), patients either proceeded to an auto-CPAP titration or had further testing with overnight polysomnography (PSG) and arterial blood gas (ABG) analysis to guide treatment with CPAP or bilevel positive airway pressure (BPAP). PSG was done if the HSAT oximetry profile was suggestive of nocturnal hypoventilation such as sustained hypoxemia (defined as persistent oxygen desaturation below 90%). A split night (diagnostic/treatment) PSG was done to determine if the patient had hypoventilation (by measurement of ABG while awake and transcutaneous CO₂ during sleep), and to determine if the patient required treatment with CPAP or BPAP.

Figure 1.

Representative home sleep apnea test recordings (10 min) from two patients showing intermittent hypoxemia due to obstructive sleep apnea, with moderate nocturnal hypoxemia [(A) Mean SpO₂ 91.3%] and severe nocturnal hypoxemia [(B) Mean SpO₂ 77.4%].

Study protocol

The study protocol for assessment of renal and vascular RAAS activity is well established [9, 25–28]. Subjects were studied sodium-replete and instructed to consume >200 mmol/day of sodium for 3 days before each study day to ensure maximum RAAS suppression [34]. Subjects were studied in the supine position in a temperature-controlled, quiet room in the morning after an 8 h fast to account for circadian variations in the RAAS [35]. All subjects provided a second morning spot urine for determination of urinary sodium to verify adherence with the high salt diet [36]. All pre-menopausal female subjects were studied 14 days after the first day of the last menstrual period, determined by counting days and measuring 17β-estradiol levels [37]. Menopause was defined as secondary amenorrhea ≥12 months [38]. Subjects on medications that interfere with RAAS activity were switched to the calcium-channel blocker (CCB) amlodipine to achieve adequate BP control 2 weeks prior to each study day, as these agents are considered to have a neutral effect on the RAAS [39, 40]. Amlodipine was taken daily each morning including the morning of the assessment.

At 8 am, an 18-gauge peripheral venous cannula was inserted into each antecubital vein (one for infusion and one for blood sampling). Each participant was given loading doses of 50 mg/kg of Inutest (Clinalfa, Austria) and 8 mg/kg para-aminohippurate (PAH; Merck, Canada), followed by constant infusions of Inutest at 30 mg/min and PAH at 12 mg/min for 90 min to establish baseline glomerular filtration rate (GFR) and effective renal plasma flow (RPF), respectively. Filtration fraction (FF), a surrogate marker of glomerular pressure, was calculated as GFR/RPF. After a 90 min equilibration period renal hemodynamics (GFR, RPF, FF) and systemic hemodynamics (BP) were measured at baseline and in response to AngII challenge (3 ng/kg/min∙30 min, 6 ng/kg/min∙30 min), as an index of renal and vascular RAAS activity [9, 25–28], respectively, followed by a 30 min recovery period. Blood samples were collected at baseline and every 30 min thereafter throughout the study period. Blood pressure was recorded at baseline and then every 15 min by an automatic recording device during the AngII infusion, and after the 30 min recovery period. A standard BP cuff (Dinamap; Critikon) was placed on the right arm and the mean of two readings taken by the same Registered Nurse (DYS) were recorded. Mean arterial pressure (MAP) was calculated as 1/3 systolic BP (SBP) + 2/3 diastolic (DBP).

CPAP therapy

After the first study day and completion of sleep diagnostic testing, subjects were treated with CPAP as per published guidelines [23]. All subjects underwent an auto-CPAP trial to determine individual CPAP requirement. Initial auto-CPAP settings were 16/6 cm H₂O and were automatically titrated according to the CPAP unit titration algorithm to optimize therapy. If airflow limitation or nocturnal hypoxemia were not fully corrected after the auto-CPAP trial, subjects were switched to fixed CPAP, which was estimated from the CPAP level at the 95th percentile. Adherence to CPAP therapy was monitored by electronic download from the unit each month. Once satisfactory CPAP adherence was achieved (defined as CPAP use for >4 h per night on >70% nights for 4 weeks) [23] and correction of OSA and nocturnal hypoxemia was confirmed by a repeat HSAT while using CPAP, subjects underwent reassessment of renal hemodynamics and renal RAAS during a second study day, identical to the pre-CPAP assessment.

Laboratory measurements

Blood samples collected for PAH and inulin determinations were immediately centrifuged at 3,000 rpm for 10 min at 4°C. Plasma was separated, placed on ice, and stored at −70°C. PAH and inulin were measured in serum by colorimetric assays using anthrone and N-(1-naphthy)-ethylenediamine, respectively. The mean of two baseline clearance periods represent RPF and GFR. A radioimmunoassay was utilized for plasma renin activity (PRA; DiaSorin Clinical Assays, Stillwater, MN, USA). In brief, angiotensin-I (AngI), the primary product of PRA was generated at 37°C from endogenous renin and renin substrate at pH 6.0. The integrity of the generated AngI was maintained by inhibition of proteolytic activity using EDTA and phenylmethylsufonyl fluoride in the generation system. The accumulated AngI reflects PRA under these controlled conditions. The AngI generated was determined by radioimmunoassay using competitive binding principles, where the antibody was immobilized onto the lower inner wall of coated tubes. AngII plasma levels were measured by standard laboratory immunoassay techniques (Quest Diagnostics; San Juan Capistrano, CA, USA). Aldosterone was also measured using a radioimmunoassay assay (Inter Medico; Markham, ON, Canada). Serum creatinine was quantified by enzymatic colorimetric assay techniques (Roche/Hitachi Creatinine Plus). Fasting glucose was determined by a hexokinase-UV assay (Roche Diagnostics, Germany). Urinary albumin (UAE) and total protein excretion (UTPE) was determined by a turbidimetric endpoint assay using benzethonium chloride (Roche Total Protein Urine/CSF Gen. 3, Roche). Urinary sodium was determined by an indirect potentiometry assay using an ion-selective electrode (Roche Cobras Integra Sodium, Roche). Catecholamines were quantified by liquid chromatography with enzymatic colorimetric assay techniques.

Sample size and power calculation

Our previous work in OSA patients on a high-salt diet demonstrated a decrease in RPF of 182 ± 98 mL/min (mean ± SE) pre-CPAP and 219 ± 112 mL/min (mean ± SE) post-CPAP in response to AngII infusion, respectively [9]. Based on these data and anticipating a 20% absolute difference in the change in RPF response to AngII after CPAP therapy between the moderate and severe hypoxemia groups, we estimated that 15 subjects would be required in each group with a two-sided alpha of 0.05% and 85% power.

Analyses

Data are reported as mean ± standard error or number (percentage), where appropriate. The primary outcome was the change (Δ) in renal hemodynamics (GFR, RPF, and FF), at baseline and in response to a vasoconstrictor (AngII) challenge as a measure of renal RAAS activity [25–27] with CPAP, stratified by NH status. Secondary outcomes were differences in BP (reflecting vascular RAS activity [28]), PRA, and aldosterone by NH status at baseline and in response to AngII with CPAP. We utilized non-parametric testing in our study. Comparisons between moderate and severe hypoxemia were conducted using the Mann–Whitney Test while pre- and post-CPAP comparisons were made using the Wilcoxon Signed-Rank Test. To examine the response to AngII on each study day we conducted a Friedman Test for outcomes with multiple measurements. Post-hoc analyses with the Wilcoxon Signed-Rank Test were conducted with a Bonferroni correction applied resulting in a significance level set at p < 0.0167 for outcomes with multiple measurements. Sensitivity analyses were conducted to exclude subjects with controlled hypertension, sleep hypoventilation, and persistent nocturnal hypoxemia while on CPAP, as these subjects may have subclinical kidney impairment from pre-existing hypertension despite normal blood pressure. We also performed exploratory sex-stratified analyses in keeping with Sex and Gender Equity in Research (SAGER) guidelines [41]. All statistical analyses were performed with statistical software package SPSS V.26.0 (IBM, Amonk, NY, USA). All analyses were two-tailed with a significance level of 0.05, apart from outcomes with multiple measurements where the significance level was set at p < 0.0167.

Results

Study enrollment

Forty-six participants with newly diagnosed OSA and NH (25 moderate, 21 severe) completed study day 1 pre-CPAP (Figure S1). One severe NH participant ingested a single dose of candesartan (AngII-receptor blocker) the morning of study day 1. This subject was studied in an identical fashion post-CPAP, including candesartan ingestion, to allow for comparison of pre-post-CPAP results. This subject was included in the final analysis, but excluded in a sensitivity analysis, as the outcomes of interest were the changes in renal hemodynamics, at baseline and in response to AngII with CPAP. Sixteen participants (10 moderate, 6 severe) withdrew from the study (unable to obtain adequate vascular access, N = 2; unable to adhere to CPAP, N = 14). Thirty participants (15 moderate, 15 severe) completed the study and were included in the final analyses.

Baseline characteristics

Baseline characteristics stratified by NH severity are presented in Table 1. All participants were non-diabetic, non-smoking, sodium-replete, with normal kidney function and BP < 140/90, and without awake hypoxemia. There was a similar distribution of self-identified pre- and post-menopausal women between groups. Both groups demonstrated intermittent hypoxemia, with a lower nadir in oxygen desaturation in the severe NH group (p < 0.001) (Figure 1). Despite the differences in NH severity, both awake SpO₂ (p = 0.13) and ODI (p = 0.13) were similar between the two groups. Eleven participants with severe NH and one participant with moderate NH had overnight polysomnography, which confirmed a diagnosis of OSA. In addition, three participants with severe NH met criteria for mild sleep hypoventilation (increase in transcutaneous PCO₂ on transition from wakefulness to sleep of 12–13 mmHg). Furthermore, eight participants with severe NH had an arterial blood gas, which did not show awake hypoventilation (PaCO₂ range 34–45 mmHg). BMI (p = 0.001), urinary albumin (p = 0.008) and protein (p = 0.050) excretion, and norepinephrine (p = 0.031) were higher in those with severe NH. There were no differences in BP or circulating RAAS components between groups.

Table 1.

Baseline characteristics by nocturnal hypoxemia severity pre- and post-CPAP therapy

	Pre-CPAP		Post-CPAP
	Moderate	Severe	Moderate	Severe
N	15	15	—	—
Age, years	48 ± 3	49 ± 3	—	—
Sex, N (% female)	4 (27)	6 (40)	—	—
Menopause, N (% of women)	2 (50)	3 (50)	—	—
Race, N (% Caucasian)	10 (67)	12 (87)	—	—
Hypertension, N (%)¶	3 (20)	7 (47)	—	—
Awake SpO2, %	95.1 ± 0.4	94.0 ± 0.6	—	—
BMI, kg/m2	32 ± 1	38 ± 2†	33 ± 1	39 ± 2†
Serum creatinine, µmol/L	72 ± 4	76 ± 4	75 ± 4	79 ± 4
Urinary Na+, mmol/day	383 ± 29	334 ± 28‡	355 ± 34	341 ± 35‡
Fasting glucose, mmol/L	4.7 ± 0.2	4.9 ± 0.1	4.8 ± 0.1	4.9 ± 0.2
Hemoglobin A1c, %	5.8 ± 0.1	6.0 ± 0.1	5.8 ± 0.1	5.6 ± 0.2
ODI, h−1	41.3 ± 4.9	55.1 ± 6.7	4.0 ± 0.7*	2.8 ± 0.6*
SpO2 <90, % monitoring time	23.1 ± 2.3	73.9 ± 4.8†	3.1 ± 1.1*	13.7 ± 5.2* §
Mean SpO2, %	91.1 ± 0.2	85.9 ± 1.1†	92.8 ± 0.4*	92.0 ± 0.5* §
Minimum nocturnal SpO2, %	75.0 ± 0.9	63.8 ± 2.9†	85.7 ± 1.0*	83.4 ± 1.8* §
Oximetry monitoring time, h	7.6 ± 0.4	7.2 ± 0.4	6.8 ± 0.3	7.3 ± 0.4§
SBP, mmHg	126 ± 3	129 ± 3	120 ± 3*	124 ± 3
DBP, mmHg	78 ± 2	78 ± 3	72 ± 2*	73 ± 2*
MAP, mmHg	94 ± 2	95 ± 3	88 ± 2*	90 ± 2
Heart rate, bpm	65 ± 3	71 ± 2	62 ± 2	66 ± 3
UAE, mg/day	4.4 ± 0.7	28.6 ± 10.7 †‡	4.6 ± 1.1	18.1 ± 6.8 †‡
UTPE, mg/day	61 ± 4	100 ± 20‡	49 ± 4*	81 ± 12 †‡
PRA, ng/mL/h	0.86 ± 0.14	1.44 ± 0.40║	0.79 ± 0.11	1.08 ± 0.22║
Angiotensin-II, ng/L	17 ± 1	23 ± 3	16 ± 1	23 ± 5
Aldosterone, pmol/L	157 ± 24	196 ± 38	108 ± 12*	110 ± 15*
Norepinephrine, nmol/L	2.3 ± 0.2	3.3 ± 0.4†	1.8 ± 0.2*	2.1 ± 0.2*
Epinephrine, pmol/L	73 ± 6	72 ± 10	75 ± 9	79 ± 15
Dopamine, pmol/L	101 ± 0	111 ± 8	100 ± 0	107 ± 6

BMI, body mass index; CPAP, continuous positive airway pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; ODI, oxygen desaturation index; PRA, plasma renin activity; SBP, systolic blood pressure; SpO₂, oxyhemoglobin saturation; UAE, urine albumin excretion; UTPE, urinary total protein excretion.

*p < 0.05 vs pre-CPAP.

†p < 0.05 vs moderate hypoxemia.

^‡N = 13.

^§N = 12.

^║N = 14.

¶Ten study participants (three moderate, seven severe) were on antihypertensive medications: moderate—ACE-inhibitor (n = 1), ARB (n = 2), beta-blocker (n = 1); severe—ACE-inhibitor (n = 1), ARB (n = 3), beta-blocker (n = 1), CCB (n = 1), and thiazide-like diuretic (n = 2). One participant with severe nocturnal hypoxemia was on dual therapy (ARB/diuretic) while another was on quadruple therapy (CCB, beta-blocker, ACE-inhibitor/diuretic). For the participant on quadruple therapy, three of the four antihypertensive agents were not at maximum doses.

All participants treated with CPAP demonstrated excellent adherence and efficacy reflected by download from their units (Table 2). Twenty participants completed the study on auto-CPAP (14 moderate, 6 severe), while 10 participants were converted to fixed-CPAP (1 moderate, 9 severe) to optimize therapy. All participants met acceptable HSAT criteria for correction of OSA (ODI < 10 h− ¹; Table 1) and all but five severe NH participants met criteria for correction of NH (SpO₂ ≥ 90% for <12% monitoring time); in four of these five participants the post-CPAP MSpO₂ was ≥90% and one participant had a MSpO₂ <90% (89% improved from 85% pre-CPAP). Two participants were unable to complete HSAT on therapy prior to reassessment. However, their CPAP downloads indicated adequate therapy reflected by an apnea–hypopnea index of 0.3 and 0.8 h⁻¹ with CPAP use >4 h/night 96% and 100% of the time used. There were no differences in ODI or NH parameters post-CPAP between groups.

Table 2.

CPAP adherence and efficacy

Parameter	Moderate (N = 15)	Severe (N = 15)	P-value*
Total days used, days	144 ± 16	142 ± 19	0.8
Average hours/night	6.4 ± 0.3	6.3 ± 0.4	0.9
30 days prior to reassessment
% time used	91 ± 3	96 ± 2	0.3
>4 h/night, % time	80 ± 5	88 ± 4	0.4
AHI, h−1	4.1 ± 1.2	1.5 ± 0.3	0.025

AHI, apnea-hypopnea index; CPAP, continuous positive airway pressure.

*Pre- vs post-CPAP.

Baseline characteristics pre- vs post-CPAP

BMI remained high among those with severe NH compared to moderate NH. Urine protein excretion decreased post-CPAP among those with moderate NH (p = 0.013), and both urine albumin (p = 0.011) and protein (p = 0.009) excretion were reduced compared to those with severe NH. Post-CPAP, moderate NH participants had reduced SBP (p = 0.033), DBP (p = 0.006), MAP (p = 0.007), norepinephrine (p = 0.009), and aldosterone (p = 0.011), while PRA was unchanged (p = 0.4). Among severe NH participants, DBP (p = 0.034), norepinephrine (p = 0.014), and aldosterone (p = 0.023) were reduced, while SBP (p = 0.19), MAP (p = 0.053), and PRA (p = 0.8) were maintained post-CPAP.

Baseline renal hemodynamics stratified by NH severity are reported in Table 3 and Figure 2. Pre-CPAP, baseline renal hemodynamics did not differ by NH severity. Post-CPAP, there were no changes in baseline GFR (p = 0.14), RPF (p = 0.3), or FF (p = 0.11) among moderate NH participants. However, in severe NH, CPAP was associated with maintained GFR (p = 0.9), increased RPF (p = 0.009), and reduced FF (p = 0.023).

Table 3.

Renal and systemic hemodynamic responses to angiotensin-II before and after CPAP therapy in obstructive sleep apnea participants with moderate (N = 15) and severe (N = 15) nocturnal hypoxemia

			Angiotensin-II
	Baseline		∆3 ng/kg/min		∆6 ng/kg/min
	Moderate	Severe	Moderate	Severe	Moderate	Severe
Pre-CPAP
GFR, mL/min	128 ± 10	112 ± 5	−9 ± 4‡	1 ± 3†	−5 ± 5	8 ± 5†
RPF, mL/min	691 ± 33║	664 ± 38	−165 ± 13‡║	−93 ± 19 †‡	−208 ± 18‡§║	−112 ± 22 †‡
FF, %	19.3 ± 1.8║	17.6 ± 1.4	4.0 ± 1.3‡║	3.1 ± 0.8‡	6.7 ± 1.3‡§║	5.0 ± 1.4‡§
SBP, mmHg	125 ± 3	129 ± 3	17 ± 4‡	14 ± 3‡	24 ± 4‡§	19 ± 3‡
DBP, mmHg	78 ± 2	78 ± 3	10 ± 1‡	7 ± 2‡	14 ± 2‡	10 ± 2‡§
MAP, mmHg	94 ± 2	95 ± 3	12 ± 2‡	9 ± 2‡	17 ± 2‡§	13 ± 2‡§
PRA, ng/mL/h	0.86 ± 0.14	1.44 ± 0.40║	−0.29 ± 0.04‡	−0.40 ± 0.11‡║	−0.50 ± 0.11‡§	−0.61 ± 0.14‡§║
Aldosterone, pmol/L	150 ± 25	196 ± 38	163 ± 26‡	216 ± 47‡	248 ± 36‡§	330 ± 60‡§
Post-CPAP
GFR, mL/min	119 ± 9	108 ± 3	−4 ± 3	−1 ± 2	−7 ± 3	−1 ± 3
RPF, mL/min	714 ± 46║	745 ± 34*	−174 ± 23‡║	−138 ± 16* ‡	−226 ± 28‡§║	−175 ± 20* ‡§
FF, %	17.6 ± 1.9║	14.9 ± 0.6*	4.5 ± 0.7‡║	3.3 ± 0.4‡	5.9 ± 0.5‡║	4.6 ± 0.6‡
SBP, mm Hg	120 ± 3*	124 ± 3	20 ± 4‡	10 ± 3‡	27 ± 4‡§	23 ± 4‡§
DBP, mm Hg	72 ± 2*	73 ± 2*	13 ± 2* ‡	9 ± 2‡	16 ± 1‡	12 ± 2‡
MAP, mmHg	88 ± 2*	90 ± 2	16 ± 2* ‡	10 ± 2 †‡	20 ± 2‡§	16 ± 2‡§
PRA, ng/mL/h	0.79 ± 0.11	1.08 ± 0.22║	−0.25 ± 0.07‡	−0.36 ± 0.11‡║	−0.43 ± 0.07‡§	−0.36 ± 0.18║
Aldosterone, pmol/L	108 ± 12*	110 ± 15*	163 ± 30‡	219 ± 42‡	256 ± 32‡§	330 ± 56‡§

AngII, angiotensin II; CPAP, continuous positive airway pressure; DBP, diastolic blood pressure; FF, filtration fraction; GFR, glomerular filtration rate; MAP, mean arterial pressure; PRA, plasma renin activity; SBP, systolic blood pressure.

*p < 0.05 vs pre-CPAP.

†p < 0.05 vs moderate hypoxemia.

^‡P < 0.0167 vs baseline.

^§p < 0.0167 vs AngII (3 ng/kg/min).

^║N = 14.

Figure 2.

Changes in baseline glomerular filtration rate (GFR, mL/min), renal plasma flow (RPF, mL/min), and filtration fraction (FF, %) in obstructive sleep apnea participants with moderate and severe nocturnal hypoxemia before and after CPAP therapy. (A) Glomerular filtration rate in moderate nocturnal hypoxemia. (B) Glomerular filtration rate in severe nocturnal hypoxemia. (C) Renal plasma flow in moderate nocturnal hypoxemia. (D) Renal plasma flow in severe nocturnal hypoxemia. (E) Filtration fraction in moderate nocturnal hypoxemia. (F) Filtration fraction in severe nocturnal hypoxemia. *p < 0.05 vs baseline pre-CPAP.

Responses to angiotensin-II

These data were stratified by NH status (Table 3; Figures 3 and 4). Pre-CPAP, following AngII infusion: (1) GFR decreased in those with moderate NH, but was maintained in those with severe NH (Δ30 min, p = 0.021; Δ60 min, p = 0.017); (2) RPF decreased in both groups but the decrease was greater in those with moderate NH (Δ30 min, p = 0.003; Δ60 min, p = 0.001); (3) FF increased in both groups. As anticipated, both groups demonstrated increased BP and aldosterone, while PRA decreased in response to AngII.

Figure 3.

Changes in glomerular filtration rate (GFR, mL/min) in response to angiotensin II in obstructive sleep apnea participants with moderate and severe nocturnal hypoxemia before and after CPAP therapy. (A) Moderate nocturnal hypoxemia. (B) Severe nocturnal hypoxemia. *p < 0.05 vs pre-CPAP response. ^†p < 0.05 vs moderate nocturnal hypoxemia response. ^‡p < 0.0167 vs baseline. ^§p < 0.0167 vs angiotensin II (3 ng/kg/min).

Figure 4.

Changes in renal plasma flow (RPF, mL/min) in response to angiotensin II in in obstructive sleep apnea participants with moderate and severe nocturnal hypoxemia before and after CPAP therapy. (A) Moderate nocturnal hypoxemia. (B) Severe nocturnal hypoxemia. *p < 0.05 vs pre-CPAP response. ^†p < 0.05 vs moderate nocturnal hypoxemia response. ‡p < 0.0167 vs baseline. ^§p < 0.0167 vs angiotensin II (3 ng/kg/min).

Post-CPAP, both groups maintained GFR, decreased RPF, and increased FF in response to AngII. However, only severe NH participants demonstrated a greater RPF vasoconstrictor response to AngII compared to pre-CPAP (Δ30 min, p = 0.009; Δ60 min, p = 0.001).

As anticipated, all participants demonstrated increased BP, decreased PRA, and increased adrenal secretion in response to AngII compared to baseline both pre- and post-CPAP, with no differences observed between groups. CPAP was associated with a greater increase in DBP (Δ30 min, p = 0.013) and MAP (Δ30 min, p = 0.035) in the moderate, but not severe NH group (MAP, Δ30 min, p = 0.038 vs moderate NH).

Sensitivity analyses

Exclusion of the participant who ingested candesartan, participants without repeat HSATs, participants with mild sleep hypoventilation, and participants with partially corrected NH post-CPAP did not alter our primary findings. Exclusion of controlled hypertensive participants did not appreciably affect the renal hemodynamic results, though post-CPAP, there was increased adrenal sensitivity to AngII among severe NH participants (Δ30 min, 172 ± 36 vs 309 ± 61 pmol/L, p = 0.019; Δ60 min, 264 ± 38 vs 466 ± 73 pmol/L, p = 0.009; moderate vs severe). Exploratory sex-stratified analyses of moderate and severe NH participants are reported in an online supplement (Tables S1−S4). Briefly, no sex differences were observed in baseline renal hemodynamics or responses to AngII pre- or post-CPAP, even after stratification by NH status.

Discussion

We examined the association between correction of NH with CPAP therapy and changes in renal hemodynamics and renal RAAS activity in humans with OSA, and specifically addressed if these relationships are modified by the severity of NH. Our primary findings were: (1) CPAP was associated with improved baseline intraglomerular pressure in those with severe but not moderate NH; and (2) CPAP was associated with down-regulation of renal RAAS activity in those with severe NH, while in moderate NH, CPAP was associated with down-regulation of vascular RAAS activity. Our findings demonstrate that the severity of NH associated with OSA impacts the changes in renal hemodynamics, renal and vascular RAAS activity, and may determine which OSA patients develop hypertension and kidney disease. Further, these findings highlight that the mechanisms by which CPAP reduces intraglomerular pressure and improves renal and vascular RAAS activity are dependent on the underlying NH profile.

The role of hypoxia on renal and systemic hemodynamics and the RAAS has been evaluated in several animal models, although there remains a paucity of human studies. In rats, CIH causes progressive increases in BP, mediated in part through renal sympathetic nerve activity that acts to increase RAAS activity through up-regulation of AngII type-1 receptors (AT₁R) [12, 42, 43]. Renal artery denervation [42], adrenal demedullation [42], and AT₁R blockade [13] ablate the increase in BP in response to CIH. Further, RAAS suppression with a high-sodium diet blocks the BP increase in rats exposed to CIH by suppressing kidney renin mRNA levels [12]. However, kidney AT₁R expression was not affected, suggesting that the local tissue RAAS may have a more prominent role in mediating the increased BP in response to CIH [12]. Angiotensin-I and -II are not only produced in the systemic circulation, but also at local tissue sites throughout the body [29]. The tissue response to exogenous AngII infusion has previously been shown to be inversely proportional to local tissue-RAS activity [34, 44, 45]. All components of the RAAS have been demonstrated in numerous organ tissues including heart, kidney, adrenal, brain, vessel wall, ovary, and testis [29]. Tissues may regulate their local angiotensin concentrations by varying the number of renin receptors, renin-binding proteins, the ACE level, metabolizing enzymes, and the angiotensin/mineralocorticoid receptor density [29]. In rats, renal hypoxia precedes the development of structural pathology, is promoted by both exogenous and endogenous AngII, and can be acutely reversed by AT₁R-blockade [46]. In humans, treatment with the AT₁R-blocker losartan ablates the rise in BP in healthy men exposed to CIH [13].

In a rabbit model of hypoxia, the renal response to similar levels of hypoxia was due to the reflex activation of sympathetic nerves, as renal denervation abolished these responses [47, 48]. Subsequent experiments evaluating two different levels of hypoxia (14% and 10% O₂) produced graded increases in total renal sympathetic nerve activity (RSNA) and decreases in renal blood flow with no change in MAP [10]. Reflex activation of sympathetic nerves produced different pre-glomerular and post-glomerular effects, which were dependent on the severity of hypoxia and extent of increase in RSNA [10]. Moderate hypoxia predominantly increased post-glomerular resistance, resulting in decreased renal blood flow, maintenance of GFR, and increased FF [10]. Conversely, severe hypoxia increased both pre-glomerular and post-glomerular resistance, resulting in decreased GFR, but FF was unchanged [10]. The lower levels of increased RSNA in response to moderate hypoxia produced a greater increase in post-glomerular than pre-glomerular resistance, which would tend to conserve GFR, despite renal vasoconstriction, while higher levels of RSNA, evoked by severe hypoxia, resulted in greater pre-glomerular renal vasoconstriction, with subsequent reductions in GFR [10]. Consequently, in severe hypoxia, there is both marked renal vasoconstriction and GFR loss [10]. The findings of this rabbit model are partially in contrast to our current study, though this may be related to intrinsic differences between species. Our study did not evaluate renal denervation of the sympathetic nerve, but rather focused on the RAAS. We also evaluated renal plasma flow as opposed to renal blood flow (RPF/[1 − Hematocrit]), which makes direct comparison of the two study models more challenging.

The “chronic hypoxia hypothesis” states that primary glomerular disease leads to chronic ischemic damage in the kidney, ultimately resulting in tubulointerstitial fibrosis, which is well established as the best predictive indicator of progression to ESKD [8]. We have previously reported that both moderate and severe NH are associated with increased intraglomerular pressure, as reflected by FF, compared to obese controls, and that severe hypoxemia was associated with a blunted RPF response to AngII indicative of increased renal RAAS activity [22]. Further, the RPF response to AngII positively correlated with NH severity parameters [22], highlighting the importance of NH as a potential therapeutic target.

The discordance observed between GFR and RPF responses in our study has been reported previously [26, 27, 49, 50]. Differences in the renal microvasculature and local tissue RAAS may account for this observation, as local AngII generation may occur independently in the respective renal vascular beds, thus impacting GFR and RPF differently [29]. AngII is a powerful endogenous vasoconstrictor with selective action on the renal blood supply [15]; even small changes in concentration lead to renal vasoconstriction and decreased RPF, although the impact of AngII on GFR is less predictable [50]. The mechanism may involve intravascular oncotic pressure changes along the glomerular capillary which result in greater surface area available for filtration [50].

We also observed similar baseline FF levels between moderate and severe NH participants despite evidence of RAAS activation suggesting the presence of increased vasodilatory activity, opposing the effect of RAAS activation in severe NH. This phenomenon has been reported previously among oral contraceptive (OC) users who exhibit elevated AngII levels and AT₁R expression, indicative of RAAS activation, yet the renal and systemic consequences are minimal, suggesting that there is increased vasodilatory activity, counteracting the effect of RAAS activation [51]. The nitric oxide (NO) system has been demonstrated to be upregulated in OC users and increased activity of the NO pathway may modulate the hemodynamic effects RAAS activation [51]. We speculate that a similar mechanism may occur among severe NH participants whereby the NO pathway or another vasodilatory pathway is upregulated counteracting RAAS activation. Assessment of counter-regulatory vasodilatory pathways in OSA and severe NH will be an important area of future study.

The impact of NH severity and CPAP on norepinephrine and aldosterone secretion deserve mention. Norepinephrine, the main vasoconstrictor of the sympathetic nervous system, and AngII, the main vasoconstrictor of the RAAS, have been demonstrated to have different effects on control of the renal blood supply in healthy humans [52]. Multiple studies have demonstrated abnormalities in sympathetic and parasympathetic activity in OSA [53] and improvement with CPAP [54]. The pathogenesis has been attributed to upper-airway obstruction, with initial stimulation of the parasympathetic nervous system, followed by sympathetic activation due to associated intermittent hypoxia [55], intrathoracic pressure changes [56], and recurrent arousals [57]. There is growing recognition of the consequences of increased aldosterone secretion and its contributions to both hypertension and heart failure with reduced ejection fraction [14, 19–21, 58]. Treatment with aldosterone antagonists reduces morbidity, mortality, and hospitalizations in patients with heart failure with reduced ejection fraction [19–21] and there is an inverse and linear relationship between renin and blood pressure response to spironolactone in patients with resistant hypertension [59]. Aldosterone excess is hypothesized to contribute to OSA through increased sodium and water retention, which, coupled with nocturnal rostral fluid shift results in tissue edema around the neck thereby promoting upper airway narrowing and collapse [60]. Treatment with spironolactone attenuated OSA and NH severity by ~50% in participants with resistant hypertension [14]. Our findings suggest that both norepinephrine and aldosterone secretion may have important modulating roles in the pathogenesis of OSA and NH.

What are the clinical implications of the study? Higher FF values reflect greater glomerular pressure, which cannot be measured directly in humans and is a precursor to glomerulosclerosis and ultimately, ESKD [61, 62]. An increase in GFR without glomerular hypertension due to a parallel increase in both GFR and RPF with normal FF, such as seen in pregnancy, does not lead to glomerular damage [61]. In contrast, glomerular hypertension leads to progressive kidney disease even in the absence of whole-kidney hyperfiltration, such as in patients with type-2 diabetes and secondary focal segmental glomerulosclerosis (FSGS) [61]. The development of glomerulomegaly and secondary FSGS has also been previously described in patients with OSA [63]. Importantly and underscoring the clinical importance of FF, a study of kidney transplant recipients found a higher FF was an independent predictor for graft loss [64]. Further, in non-transplant populations increased FF and glomerular hyperfiltration have been associated with impaired cardiac function [65] and mortality [66]. RAAS activation predisposes to the development of both kidney [15, 16] and cardiovascular disease [16–21] and RAAS-blockade is associated with improved cardiorenal outcomes [16–21]. CPAP therapy is an effective treatment for OSA, associated with down-regulation of RAAS activity [9] and improvement of glomerular hyperfiltration [9, 11]. Our findings suggest that severe NH participants receive more renal benefit from CPAP therapy than moderate NH. In the era of personalized medicine, if we are to use treatment of OSA as a potential renal protective strategy, then our study findings suggest that there should be a greater focus on patients with severe NH rather than moderate NH. More resources should be deployed to patients with this OSA profile to ensure good adherence to CPAP therapy or even consideration of nocturnal oxygen in those who cannot tolerate CPAP, which may be useful in a time of limited healthcare resources.

Our study has strengths and limitations. First, it was restricted to OSA participants without co-morbidities (apart from obesity and controlled hypertension), limiting the generalizability of our results. However, by studying a healthier population, we were able to examine the impact of NH severity on the effects of CPAP on renal hemodynamics and RAAS activity without confounding factors. The participants in our study were obese, which is common in OSA [67]. Increased BMI impacts renal hemodynamics and renal RAAS activity [68, 69]. By design, we included only participants with both OSA and moderate-severe NH. In selecting for OSA participants with more severe hypoxemia, it is not surprising that there were differences in baseline BMI. While we did not include a control group of obese participants without OSA, we have previously reported that renal RAAS activation is independent of obesity in OSA participants compared to otherwise healthy obese controls [22]. Importantly, each subject served as their own control in our pre-post study design and BMI was stable within groups between study days. Second, our sample size was limited and did not include control groups of non-OSA, hypertensive OSA, CPAP non-adherent OSA, or OSA participants who received either no CPAP or sham-CPAP. We attempted to minimize the effect of sample size and intra-individual variability by utilizing a homogenous study group with careful pre-post study design. While no control group was included, conditions during the renal assessment were standardized to minimize the impact of potential confounders. Specifically, participants were studied supine in the morning to account for RAAS circadian variations [35], sodium-replete to ensure maximal RAAS suppression [34], during the same stage of the menstrual cycle to eliminate estrogen-mediated RAAS differences [37], had stable BMI [68], and all participants had BP <140/90 with no other comorbidities and were free of RAAS-inhibiting medications and exogenous sex hormones. Nonetheless, while unlikely, it remains possible that observed changes in renal hemodynamics and RAAS activity were due to factors other than CPAP and correction of NH. Third, by ensuring that all participants were on effective CPAP for the same duration (4 weeks) before renal reassessment, we standardized the exposure of efficacious CPAP, a duration that has been associated with improved cardiovascular outcomes [70]. However, by only studying RAAS activity after a relatively short exposure to CPAP, we are unable to comment on the potential effects of longer CPAP exposure. Fourth, we used portable monitoring instead of polysomnography both to diagnose OSA and evaluate patients’ responses to CPAP and as such, did not include an objective measure of sleep or account for arousals. This raises the possibility that the severity of NH in our study population may have been underestimated, which would have biased against finding a difference between NH severity groups. However, the use of portable monitoring in place of polysomnography was appropriate for our study population according to previously published guidelines [23] and the measurement of NH with oximetry is robust regardless of the sleep parameters. A greater duration of sleep would be anticipated to result in a greater amount of hypoxemia; thus, it is possible that the relationship observed in our study between severe NH and renal outcomes may be an underestimate. Fifth, we did not exclude primary hyperaldosteronism in our study participants with controlled arterial hypertension. Next, given that OSA and associated NH is a systemic illness affecting multiple organ systems and that CPAP therapy itself is a systemic treatment, the beneficial effects of CPAP for OSA on renal and systemic hemodynamics and RAAS activity are likely multifactorial and may involve multiple direct and indirect neurohormonal pathways. Finally, our study was not designed to differentiate between fixed- and auto-CPAP modalities. Fixed-CPAP but not auto-CPAP was associated with an attenuation in the decline in GFR in OSA participants.[71] CPAP modality will be an important area of future study.

In conclusion, in this community-based OSA population, the severity of NH influenced the impact of CPAP therapy on renal hemodynamics and down-regulation of renal and vascular RAAS activity. The underlying mechanisms by which CPAP improves kidney function and RAAS activity appears to differ between those with moderate and severe NH. This study highlights the importance of NH in OSA-mediated kidney and vascular disease, and the potential greater benefit of CPAP therapy in those with more severe NH. In the era of precision-based medicine, the impact of identifying and targeting this OSA phenotype deserves further study.

Acknowledgments

We thank the FMC Sleep Centre for recruitment and diagnostic testing. This study was performed at the University of Calgary.

Funding

This study was funded by an Establishment Grant from Alberta Innovates – Health Solutions. SB Ahmed is supported by Alberta Innovates – Health Solutions, the Canadian Institute of Health Research, and a joint initiative between Alberta Health and Wellness and the Universities of Alberta and Calgary. Funding sources had no role in study design, conduct, or reporting.

Disclosure Statement

Financial disclosure: none.

Non-financial disclosure: none.

Author Contributions

Study design: D.D.M.N., P.J.H., A.A.Z., S.B.A. Acquisition, analysis, data interpretation, and manuscript review: D.D.M.N., P.J.H., A.A.Z., G.B.H., D.Y.S., S.B.A. Manuscript drafting: D.D.M.N., P.J.H., S.B.A. All authors have seen and approved the manuscript.

Data Availability

The data underlying this article are available in the article and in its online supplementary material.