Annals of the American Thoracic Society

Rationale: Excessive sodium may have a role in the pathogenesis of obstructive sleep apnea (OSA) for patients with hypervolemic conditions, but it is unclear whether this is valid for all patients with OSA, including those with no significant comorbidities.

Objectives: To test the association of urinary sodium and OSA in a large sample of participants from the ELSA-Brasil (Estudo Longitudinal de Saúde do Adulto-Brasil) Study. In addition, we stratified the analysis participants according to the presence of hypertension.

Methods: In this cross-sectional study, OSA was defined by an apnea–hypopnea index ≥15 events/h. A validated 12-hour urine collection as representative of the 24-hour period was obtained from all participants to measure sodium excretion. We performed a logistic regression analysis to test the association of urinary sodium excretion with OSA (dependent variable) adjusting for age, sex, race and income, glomerular filtration rate, diabetes, physical activity, and antihypertensive classes related to sodium excretion. To address potential residual factors that may influence sodium excretion, we performed additional analysis replacing sodium excretion for salt intake (food frequency questionnaire) using the same models.

Results: We studied 1,946 participants (age 49 ± 8 yr; 43.4% men). A third of them had OSA. Compared with those with no OSA, participants with OSA presented with higher sodium excretion (1.66 [1.19–2.29] vs. 1.99 [1.44–2.69] g/12 h; P < 0.001). After adjustments for confounding factors, we found no overall significant associations of sodium excretion with OSA (odds ratio [OR], 1.09; 95% confidence interval [CI], 0.97–1.23; P = 0.150). Regardless of the OSA status, the sodium excretion was higher in hypertensive than in normotensive participants (1.93 [1.35–2.64] vs. 1.71 [1.22–2.37] g/12 h). An independent association of sodium excretion with OSA was observed in patients with hypertension only (OR, 1.326; 95% CI, 1.067–1.648; P = 0.011), but the interaction of urinary sodium with hypertension was not significant (P = 0.37). The analysis of salt intake revealed consistent findings.

Conclusions: The potential role of sodium in the pathogenesis of OSA seems to be modest and limited for those with higher salt intake and, consequently, higher fluid retention such as observed in patients with hypertension.

Obstructive sleep apnea (OSA) is a chronic clinical condition characterized by recurrent episodes of upper-airway collapse during sleep leading to increased negative intrathoracic pressure, sleep fragmentation, and intermittent hypoxia (1, 2). This sleep-disordered breathing is common worldwide. Using contemporary criteria for defining respiratory events, recent data from two population-based studies showed that roughly one-third of subjects have moderate to severe OSA (2, 3). This prevalence of OSA is much higher in patients with fluid-retaining states such as hypertension (including resistant and refractory forms), heart failure, and end-stage renal disease, reaching rates up to 95% (48). In these scenarios, it is clear that the burden of OSA may not be driven only by traditional risk factors like obesity (5, 7, 8). Recent evidence suggests that the fluid accumulates in the intravascular and interstitial spaces of the legs during the day owing to gravity and upon lying down at night redistributes rostrally (9). Some of this fluid may accumulate in the neck, increasing tissue pressure and causing the upper airway to narrow, predisposing to OSA (9, 10). High sodium intake can contribute to the fluid retention in hypervolemic patients, supporting a potential role in the pathogenesis of OSA (5). Indeed, in patients with resistant hypertension and hyperaldosteronism, increased dietary sodium was related to the severity of OSA (11). Supporting this concept, a recent study found that intensive diuretic therapy in uncontrolled hypertensive patients promoted a modest but significant reduction in the OSA severity in proportion to the degree of the overnight rostral fluid shift attenuation (12).

Despite the evidence above, the potential role of sodium and related fluid retention for the OSA pathogenesis still deserves additional studies. If the excess of sodium may predispose to OSA across the spectrum of patients is not clear. Considering that approximately 95% of the sodium intake is usually excreted in the urine (13), we explored the association between urinary sodium excretion with OSA in a large sample of participants from the ELSA-Brasil (Estudo Longitudinal de Saúde do Adulto-Brasil) Study. We made the hypothesis that sodium excretion is independently associated with OSA, but based on the well-known importance of age, sex, and obesity predisposing to OSA (1), this association would be modest. Because high dietary sodium intake is a major issue in the pathogenesis of hypertension (14), we also speculated that the magnitude of this association would be more evident in patients with hypertension. We performed additional analysis using salt intake from the food frequency questionnaire (FFQ) instead of urinary sodium excretion to address potential residual factors that may influence sodium excretion.

Our local ethical committee approved this study (number 56008616.3.0000.0076), and all participants signed an informed consent term.

The ELSA-Brasil is a study of civil servants from six universities devoted to investigating the associations and the incidence of cardiovascular diseases and their biological, behavioral, environmental, occupational, psychological, and social factors (15). The definitions and standard procedures were detailed before (15).

For this study, we invited consecutive adult participants from the ELSA-Brasil Sao Paulo center with a valid urine sample (see below for details) for performing a sleep study. We excluded shift workers, refusal to perform sleep study, failure to perform sleep assessment tests, presence of sleep apnea of central predominance, current treatment for OSA (including continuous positive upper airway pressure, mandibular advancement, and oxygen use), and use of anxiolytics and sedative drugs. In addition, we excluded participants who did not have urine samples or renal function available.

Sodium Excretion

Following the ELSA-Brasil routine for laboratory exams (16), the urine samples were collected between 7:00 p.m. and 7:00 a.m. The participants were asked to record the exact start and end times of the sample collection, as well as any losses. This 12-hour analysis was chosen to minimize the inaccuracy of the urine volume collected and stored during working periods and to avoid the influence of significant sodium loss in sweat. Urine aliquots were sent to the ELSA-Brasil Central Laboratory for measurements of creatinine (Jaffé method) and sodium and potassium (selective ion electrode method) (17). Previous validation of 12-hour urine collected at night as a reliable tool to estimate 24-hour intake/excretion of sodium was performed observing the following premises: volume ≥250 ml, collecting time between 600 and 840 minutes (10–14 h), and a 12-hour creatinine excretion ≥7.2 and <16.8 mg/kg in men and ≥5.6 and <12.6 mg/kg in women (18, 19).

Salt Intake

As previously described and validated in the ELSA-Brasil (17), we used the FFQ to estimate salt intake. This semiquantitative questionnaire with 114 food items had the aim of assessing regular food consumption over the last 12 months. The nutritional composition of each food item included in the FFQ was estimated using the Nutrition Data System for Research database, Version 2010 (20). The nutrient composition of regional preparations used in ELSA-Brasil was calculated based on recipes provided by technical research publications and teaching institutions (17).

Renal Function

Serum creatinine was analyzed by the Jaffé kinetic method (Advia 1200; Siemen). Subsequently, a derived conversion factor was applied from the standard creatinine sample, calibrated according to the reference method, based on isotopic dilution and mass spectrometry, as recommended by the National Kidney Disease Education Program (21, 22).

The albumin–creatinine ratio (ACR) was calculated from the 12-hour urine samples obtained from concentrations of albumin and creatinine in the urine. The Jaffé kinetics methods (Advia 1200; Siemen) were used to measure urinary creatinine and the immunochemical assay (BN II Nephelometer; Siemens Dade Behring) to measure urinary albumin (23). Abnormal ACR value was defined by >30 mg/g of creatinine.

The estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration formula (24, 25). This study adopted the three variables of the Chronic Kidney Disease Epidemiology Collaboration equation: serum creatinine, age, and sex. Race was not used to calculate the eGFR because the Brazilian population represents a significant admixed sample. Moreover, a previous study showed that adjustments for race did not contribute to a better performance of the chronic kidney disease (CKD) equation in our population (26, 27). CKD was defined by an eGFR <60 ml/min/1.73 m2 (28).

Sleep Study

A detailed description of the sleep analysis was previously reported (3). Briefly, we performed portable sleep monitoring (Embletta Gold; Natus Medical Inc.). We recently performed a pragmatic validation of this portable monitor in 300 participants of the ELSA-Brasil (29). We classified sleep events according to the American Academy of Sleep Medicine (2012) using a 3% oxygen desaturation criterion for hypopnea (30). As previously described (3), OSA was defined by an apnea–hypopnea index (AHI) ≥15 events per hour of recording (limited to moderate to severe cases of OSA). This option was based on recent evidence that mild OSA appears to have no significant impact as suggested by a systematic review (31).

Statistical Analysis

The variables were expressed as means and standard deviations or as medians and interquartile ranges. The visual inspection of the normal curve and the Kruskal-Wallis test were used to test the normality of continuous variables. The categorical variables were presented as percentage. Continuous variables were evaluated by analysis of variance with Bonferroni correction for comparison between groups. The frequencies were compared using the chi-square test.

We initially performed a logistic regression model to explore the associations of urinary sodium excretion, microalbuminuria, eGFR, and ACR with OSA. An unadjusted model was initially done, followed by a second model adjusted for age, sex, race, and income. A final model included model 2 plus traditional variables related to renal sodium excretion such as eGFR, diabetes, physical activity, and antihypertensive classes related to sodium excretion such as diuretics, angiotensin-converting enzyme inhibitors, and angiotensin-receptor blockers. We further performed a subanalysis stratifying the participants into normotensives and hypertensives. Hypertension was defined as the use of medications to treat hypertension, or systolic blood pressure (BP) ≥140 mm Hg or diastolic BP ≥90 mm Hg; for this, three measurements were taken at one-minute intervals. The mean of the two latest BP measurements was used (32). We presented additional analysis replacing sodium excretion for salt intake using the same models. We estimated the odds ratio and the 95% confidence interval (CI) in the different regression models. Effect modification of relationships by hypertension was also assessed using a urinary sodium by hypertension interaction term.

Finally, linear regression analysis was performed to evaluate whether urinary sodium excretion, microalbuminuria, eGFR, and ACR are associated with markers of OSA severity (AHI, total time [%] spent with an oxygen saturation <90% and lowest oxygen saturation levels). The values of the β correlation coefficient (β) and the 95% CI in the different regression models were estimated. All statistical analyzes were performed using the Statistical Package for Social Sciences, V.25.0 for Windows (SPSS Inc.). For all tests, values of P < 0.05 were considered statistically significant.

During a 2-year period, 2,561 participants were invited, and 1,946 participants were included in the final analysis. Detailed reasons for exclusions are depicted in Figure 1.

A third of the participants presented with OSA. Table 1 describes the characteristics of the studied sample. Overall, the participants were middle-aged adults, white, predominantly women, and overweight. Approximately one-fourth presented with the diagnosis of hypertension. Table E1 in the online supplement reports the characteristics of patients with hypertension with and without OSA. Compared with participants without OSA (Table 1), those with OSA were older, heavier, and predominantly male. In addition, they had more comorbidities such as hypertension, diabetes mellitus, and dyslipidemia, but no differences in the frequency of CKD or heart failure were observed. Consistently, the OSA group took more medications than participants without OSA.

Table 1. Demographic, anthropometric, and clinical characteristics of the total sample stratified by the presence of OSA

CharacteristicsTotal (N = 1,946)No OSA [n = 1,303 (67%)]OSA [n = 643 (33%)]P Value
Sex, M, n (%)43.436.856.8<0.001
Age, yr49 ± 848 ± 851 ± 8<0.001
Race, %   0.011
Physical activity intensity, %   <0.001
BMI, kg/m226.9 ± 4.725.8 ± 4.129.3 ± 4.9<0.001
Waist circumference, cm88.8 ± 12.185.4 ± 10.895.6 ± 11.7<0.001
Waist/hip ratio0.88 ± 0.080.86 ± 0.080.92 ± 0.09<0.001
Neck circumference, cm35.8 ± 3.634.9 ± 3.337.6 ± 3.6<0.001
Systolic BP, mm Hg117.8 ± 15.0115.5 ± 14.1122.4 ± 15.7<0.001
Diastolic BP, mm Hg74.5 ± 10.372.9 ± 9.777.9 ± 10.7<0.001
Comorbidities/habits, %    
 Chronic kidney disease3.
 Heart failure1.
 Diabetes mellitus15.711.324.6<0.001
Medications, %    
 ACE inhibitor9.77.414.3<0.001
 Angiotensin receptor blocker4.53.27.0<0.001
 Calcium channel blocker2.51.44.6<0.001
Sleep monitoring    
 AHI, events/h9.9 (4.5–19.5)6.0 (3.2–9.9)26.1 (19.6–38.4)<0.001
 Baseline SpO2, %94 (93–95)95 (94–96)93 (92–94)<0.001
 Lowest SpO2, %86 (82–89)88 (85–90)81 (76–84)<0.001
 Total time SpO2 < 90%, %0.4 (0.0–2.6)0.1 (0.0–0.6)3.2 (1.1–9.4)<0.001

Definition of abbreviations: ACE = angiotensin-converting enzyme; AHI = apnea–hypopnea index; BMI = body mass index; BP = blood pressure; OSA = obstructive sleep apnea; SpO2 = oxygen saturation as measured by pulse oximetry.

Data are presented as mean ± standard deviation or median (interquartile range) when appropriate.

Figure 2 reports the 12-hour urinary sodium excretion. Overall, the mean sodium excretion was 1.76 (1.24–2.43) g/12 hours. Compared with participants without OSA, the OSA group presented with higher sodium excretion. We observed a modest but significant correlation of urinary sodium excretion with the AHI (Figure E1). Regardless of OSA status, the sodium excretion was higher in hypertensive than in normotensive participants (Table E2).

Table 2 reports data on eGFR, microalbuminuria, and ACR. The OSA group presented with higher values of serum and urinary creatinine and expected lower eGFR when compared with the participants without OSA, but the mean values were in the normal range.

Table 2. Urinary sodium excretion, microalbuminuria, and renal function of the total sample of participants studied and stratified by the presence of OSA

 Total (N = 1,946)No OSA [n = 1,303 (67%)]OSA [n = 643 (33%)]P Value
Urinary sodium excretion, g/12 h1.76 (1.24–2.43)1.66 (1.19–2.29)1.99 (1.44–2.69)<0.001
Microalbuminuria, mg/dl0.54 (0.36–0.86)0.55 (0.37–0.86)0.53 (0.33–0.90)0.222
eGFR, ml/min94.3 (83.3–107.7)96.1(84.5–109.5)91.3 (81.0–104.6)<0.001
ACR, mg/g6.3 (4.9–8.1)6.5 (5.1–8.1)5.8 (4.6–8.2)<0.001
ACR >30, mg/mmol, %

Definition of abbreviations: ACR = albumin–creatinine ratio; eGFR = estimated glomerular filtration rate; OSA = obstructive sleep apnea.

Data are presented as median (interquartile range) or percentage when appropriate.

In the logistic regression analysis using the presence of OSA as the dependent variable, we found no significant associations of OSA with sodium excretion (Figure 3). However, when we stratified the sample in normotensives and hypertensives, we found significant and independent associations of sodium excretion with OSA in patients with hypertension only (Figures 4A and 4B). The interaction of urinary sodium with hypertension was not significant (P = 0.37). No significant associations with OSA were observed regarding the other parameters (microalbuminuria, eGFR, and ACR) in the whole sample as well as when stratified by normotensives and hypertensives (Tables E3 and E4). Consistent findings were obtained by the analysis of salt intake (Tables E5–E7).

Table 3 reports the linear regression of AHI, total time with oxygen saturation as measured by pulse oximetry <90%, and lowest O2 saturation with urinary sodium excretion, albuminuria, and renal function. Despite significant associations in the unadjusted model, these OSA components were not associated with any variable.

Table 3. Linear regression of urinary sodium excretion, microalbuminuria, eGFR, and ACR with apnea–hypopnea index, total time SpO2 <90% (%), and lowest SpO2 (%) in total sample

 Model 1Model 2Model 3
β95% CIP Valueβ95% CIP Valueβ95% CIP Value
 Urinary sodium excretion, g/12 h0.1752.042 to 3.409<0.0000.036−0.110 to 1.2200.1020.035−0.146 to 1.2210.123
 Microalbuminuria, mg/dl0.004−0.125 to 0.1500.860−0.019−0.181 to 0.0630.344−0.021−0.186 to 0.0610.320
 eGFR, ml/min−0.116−0.138 to −0.062<0.0000.011−0.029 to 0.0480.6320.015−0.027 to 0.0520.529
 ACR, mg/g0.008−0.008 to 0.0120.721−0.012−0.012 to 0.0060.553−0.012−0.012 to 0.0060.562
Total time SpO2 <90%, %         
 Urinary sodium excretion, g/12 h0.1090.740 to 1.798<0.0000.036−0.121 to 0.9670.1280.035−0.136 to 0.9550.141
 Microalbuminuria, mg/dl0.017−0.065 to 0.1430.4650.004−0.090 to 0.1070.8640.005−0.088 to 0.1100.830
 eGFR, ml/min−0.054−0.065 to −0.0060.0200.033−0.010 to 0.0520.1880.027−0.014 to 0.0490.276
 ACR, mg/g0.029−0.003 to 0.0120.2140.014−0.005 to 0.0090.5200.017−0.004 to 0.0100.436
Lowest SpO2, %         
 Urinary sodium excretion, g/12 h−0.099−0.415 to −0.152<0.000−0.033−0.232 to 0.0440.180−0.032−0.230 to 0.0460.192
 Microalbuminuria, mg/dl−0.006−0.029 to 0.0220.7850.002−0.024 to 0.0260.933−0.003−0.027 to 0.0230.901
 eGFR, ml/min0.036−0.002 to 0.0130.125−0.038−0.014 to 0.0020.135−0.037−0.014 to 0.0020.147
 ACR, mg/g−0.014−0.002 to 0.0010.550−0.006−0.002 to 0.0020.799−0.011−0.002 to 0.0010.619

Definition of abbreviations: ACR = albumin–creatinine ratio; AHI = apnea–hypopnea index; BMI = body mass index; CI = confidence interval; eGFR = estimated glomerular filtration rate; SpO2 = oxygen saturation as measured by pulse oximetry.

Model 1: no adjustment. Model 2: age, sex, race, BMI, and income. Model 3: model 2 plus diabetes, eGFR, physical activity, diuretics, use of angiotensin-converting enzyme inhibitor, and use of angiotensin receptor blocker.

In a subanalysis using the traditional AHI cutoffs for mild OSA (AHI 5–14.9 events/h), moderate OSA (AHI 15–29.9 events/h), and severe OSA (AHI ≥30 events/h), we observed a progressive increment in the sodium excretion (Figure E2). Further stratifications by normotensive and hypertensive as well as in hypertensive patients using no antihypertensive medications or no diuretics only revealed the following results (Figure E3): Overall, sodium excretion was only associated with severe OSA, but normotensive participants with severe OSA were not associated with sodium excretion; this association was observed for patients with moderate and severe OSA with hypertension only. When restricting to the sample with hypertension who did not report antihypertension medication use (n = 121), we found no association, but a progressive trend was observed for those hypertensives using no diuretics. Change in AHI with body position was unassociated with sodium excretion (Tables E8 and E9).

Using a large sample of adult participants, we did not find significant associations between sodium excretion and OSA in the whole population. Interestingly, when we stratified the sample according to the hypertension status, we found 1) that the sodium excretion was higher in hypertensive versus normotensive participants, regardless of the OSA status, and 2) a significant and independent association of sodium excretion with OSA in individuals with hypertension only. Of note, this association may be relevant (each 1 g of sodium excretion was associated with a 33% higher chance of having moderate to severe OSA in patients with hypertension). However, the effect modification of relationships by hypertension was not significant, requiring confirmation in further studies. Taken together, our results underscore that the role of dietary sodium on OSA pathogenesis may be limited to participants with higher salt intake and related high-volume status, such as observed in those with hypertension.

The role of sodium in promoting fluid retention, vascular remodeling, cardiac hypertrophy, and hypertension is well established (3336). More recently, the role of high dietary salt and fluid retention has gained increasing interest by its potential role in the pathogenesis of OSA. Indeed, in hypervolemic conditions, such as CKD (37, 38), heart failure (5, 39), resistant hypertension (5), and hyperaldosteronism (11), high salt intake was independently related to the AHI and presence of OSA. Importantly, it was not clear whether we could extrapolate the role of dietary sodium and related fluid retention on OSA pathogenesis to the myriad of patients with this sleep-disordered breathing. A previous investigation showed no associations between fluid retention and OSA severity in patients referred to sleep studies (40). In our study, we measured objective sodium excretion in a large population containing an appropriate control group without OSA. Overall, we found no association between sodium excretion and the presence of OSA. Our main results are consistent with a previous publication comprising 65 patients referred to sleep studies. In this study, Jafari and Mohsenin did not find an independent association between OSA severity and the displacement of fluids during sleep (40). It is important to mention that the control group (patients without OSA) of this study were patients referred to the sleep laboratory with complaints of snoring or suspicion of OSA (40). Using a large sample size (n = 1,946, 510 of whom were hypertensive), our findings showed an independent association of sodium excretion with OSA only in the hypertension subgroup.

Our subanalysis revealed interesting results. When we stratified our cohort into normotensive and hypertensive participants, normotensive participants had lower sodium excretion than hypertensive subjects regardless of OSA status. In both groups, however, participants with OSA had higher sodium excretion than those without OSA. It is not clear why the relation between OSA and sodium excretion is restricted to hypertensive patients. The higher salt intake, lower eGFR, older age, and higher body mass index in hypertensive patients may partially explain these results. However, adjusting for these and other confounding factors did not change our main findings (data not shown). How to explain the higher sodium excretion in participants with OSA versus those without OSA? And the higher sodium excretion in hypertensive patients regardless of OSA status? This epidemiological study was not designed to explore potential mechanisms. However, we speculated that because obesity is a risk factor for OSA, it is reasonable to expect higher food intake in participants with OSA (41). Regarding the higher sodium excretion observed in hypertensives, a previous study showed higher correlations between taste sensitivity and salt intake in hypertensive than in normotensive individuals (42, 43). Another potential explanation relates to the fact that the renin–angiotensin system is more active in hypertensive than in normotensive controls and may contribute to higher sodium and water intake and fluid retention (44). Experimental data found that the increased cerebral production of angiotensin II or increased sensitivity to its action may be responsible for the higher preference for solutions with higher salt concentration seen in hypertensive rats (45). Based on this evidence, future studies in this area are warranted.

Strengths and Limitations

Our study has strengths to be addressed. We included a large sample of participants. Additionally, we performed our analysis using objective sodium excretion (46). Even considering the limitations of the subjective measurements of salt intake, we performed additional analysis to address some potential residual factors that may influence sodium excretion. Similar results reinforce the main study message. However, the following limitations deserve mentioning: 1) We used portable sleep monitoring for diagnosing OSA. However, this portable sleep monitor was previously validated and presented with an excellent agreement as compared with a simultaneous actigraphy recording to defined sleep time (29). 2) Our study has a cross-sectional design. Therefore, we cannot claim any causal relationship. Whether urinary sodium excretion is a compensatory mechanism to lower the blood pressure, mainly by activation of atrial natriuretic peptide during intrathoracic pressure swings, remains to be determined. However, data on sodium intake revealed consistent results. 3) The reverse causality between sodium and OSA may have biological plausibility; previous evidence suggested that OSA seems to increase nocturnal sodium excretion (47). Therefore, it is reasonable to speculate that the current study may overestimate the impact of dietary sodium in the OSA group. However, our validated 12-hour urine sample certainly surpasses the sleep time. Indeed, most of the participants from the ELSA-Brasil study slept 6–7 hours (3). Based on the aforementioned findings using salt intake estimation, we strongly believe that this option had no major impact on our results. 4) The lack of markers such as atrial natriuretic peptide and renin–angiotensin–aldosterone as well as bioimpedance data to confirm the hypervolemic status in patients with hypertension, although intuitively, prevented us from performing correlations between sodium excretion/salt intake and fluids. 5) Multiple urine samples would be ideal for minimizing the effects of day-to-day variation of sodium excretion. Our study had almost 2,000 participants; multiple urine collections in this scenario is certainly costly and we would probably have low adherence and significant urine losses. Unfortunately, it is usually not feasible for population-based studies. 6) We did not perform additional analysis comparing uncontrolled hypertension versus controlled or resistant hypertension. In this large cohort, no formal monitoring of drug adherence was performed and ambulatory blood pressure monitoring was not available, preventing any precise definition of resistant hypertension diagnosis.

Our findings have potential clinical implications. Despite the association of OSA with hypertension, a significant proportion of individuals with OSA are normotensives (3). Salt restriction as an adjunctive therapy for treating OSA in normotensive patients might not be useful. For patients with hypertension, this is a well-established nonpharmacological recommendation for lowering BP (32), but its real utility in patients with OSA and hypertension is still uncertain. Supporting this concept, a recent randomized clinical trial in which men with severe OSA underwent 1 week of intervention and had their AHI reassessed subsequently showed that two interventions, namely, diuretic use or a restricted sodium diet, reduced body salt and water content in parallel to a significant decrease in the AHI among male patients with severe OSA (48). However, a small proportion of these patients had a standard hypertension diagnosis. Moreover, the authors found low correlations of OSA severity outcomes (AHI and lowest oxygen saturation) with diet and/or diuretics, indicating that only a minor fraction of the change in OSA severity is explained by a change in total body water content (48).


Our results pointed to an independent association between sodium excretion and OSA only in hypertensive patients and underscore that the role of sodium in the pathogenesis of OSA may be limited to specific clinical conditions.

1 . Drager LF, Togeiro SM, Polotsky VY, Lorenzi-Filho G. Obstructive sleep apnea: a cardiometabolic risk in obesity and the metabolic syndrome. J Am Coll Cardiol 2013;62:569576.
2 . Heinzer R, Vat S, Marques-Vidal P, Marti-Soler H, Andries D, Tobback N, et al. Prevalence of sleep-disordered breathing in the general population: the HypnoLaus study. Lancet Respir Med 2015;3:310318.
3 . Drager LF, Santos RB, Silva WA, Parise BK, Giatti S, Aielo AN, et al. OSA, short sleep duration, and their interactions with sleepiness and cardiometabolic risk factors in adults: the ELSA-Brasil Study. Chest 2019;155:11901198.
4 . Pedrosa RP, Drager LF, Gonzaga CC, Sousa MG, de Paula LK, Amaro AC, et al. Obstructive sleep apnea: the most common secondary cause of hypertension associated with resistant hypertension. Hypertension 2011;58:811817.
5 . Friedman O, Bradley TD, Chan CT, Parkes R, Logan AG. Relationship between overnight rostral fluid shift and obstructive sleep apnea in drug-resistant hypertension. Hypertension 2010;56:10771082.
6 . Martínez-García MA, Navarro-Soriano C, Torres G, Barbé F, Caballero-Eraso C, Lloberes P, et al.; on behalf the Spanish Sleep Network. Beyond resistant hypertension. Hypertension 2018;72:618624.
7 . Kasai T, Arcand J, Allard JP, Mak S, Azevedo ER, Newton GE, et al. Relationship between sodium intake and sleep apnea in patients with heart failure. J Am Coll Cardiol 2011;58:19701974.
8 . Harmon RR, De Lima JJG, Drager LF, Portilho NP, Costa-Hong V, Bortolotto LA, et al. Obstructive sleep apnea is associated with interdialytic weight gain and increased long-term cardiovascular events in hemodialysis patients. Sleep Breath 2018;22:721728.
9 . Kasai T, Motwani SS, Elias RM, Gabriel JM, Taranto Montemurro L, Yanagisawa N, et al. Influence of rostral fluid shift on upper airway size and mucosal water content. J Clin Sleep Med 2014;10:10691074.
10 . White LH, Lyons OD, Yadollahi A, Ryan CM, Bradley TD. Night-to-night variability in obstructive sleep apnea severity: relationship to overnight rostral fluid shift. J Clin Sleep Med 2015;11:149156.
11 . Pimenta E, Stowasser M, Gordon RD, Harding SM, Batlouni M, Zhang B, et al. Increased dietary sodium is related to severity of obstructive sleep apnea in patients with resistant hypertension and hyperaldosteronism. Chest 2013;143:978983.
12 . Kasai T, Bradley TD, Friedman O, Logan AG. Effect of intensified diuretic therapy on overnight rostral fluid shift and obstructive sleep apnoea in patients with uncontrolled hypertension. J Hypertens 2014;32:673680.
13 . Frost CD, Law MR, Wald NJ. By how much does dietary salt reduction lower blood pressure? II--Analysis of observational data within populations. BMJ 1991;302:815818.
14 . Kotchen TA, Cowley AW Jr, Frohlich ED. Salt in health and disease--a delicate balance. N Engl J Med 2013;368:12291237.
15 . Aquino EML, Barreto SM, Bensenor IM, Carvalho MS, Chor D, Duncan BB, et al. Brazilian Longitudinal Study of Adult Health (ELSA-Brasil): objectives and design. Am J Epidemiol 2012;175:315324.
16 . Mill JG, Pinto K, Griep RH. Medical assessments and measurements in ELSA-Brasil [in Portuguese]. Rev Saude Publica 2012;47(Suppl 2):5462.
17 . Pereira TS, Benseñor IJ, Meléndez JG, Faria CP, Cade NV, Mill JG, et al. Sodium and potassium intake estimated using two methods in the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil). Sao Paulo Med J 2015;133:510516.
18 . Mill JG, Silva AB, Baldo MP, Molina MCB, Rodrigues SL. Correlation between sodium and potassium excretion in 24- and 12-h urine samples. Braz J Med Biol Res 2012;45:799805.
19 . Mill JG, Baldo MP, Molina MDCB, Schmidt MI, Barreto SM, Chor D, et al. Sex-specific patterns in the association between salt intake and blood pressure: the ELSA-Brasil study. J Clin Hypertens (Greenwich) 2019;21:502509.
20 . Nutrition Data System for Research software version. Minneapolis, MN: Nutrition Coordinating Center, University of Minnesota; 2010 [accessed 2021 Jan 25]. Available from:
21 . Lustgarten JA, Wenk RE. Simple, rapid, kinetic method for serum creatinine measurement. Clin Chem 1972;18:14191422.
22 . Coresh J, Astor BC, McQuillan G, Kusek J, Greene T, Van Lente F, et al. Calibration and random variation of the serum creatinine assay as critical elements of using equations to estimate glomerular filtration rate. Am J Kidney Dis 2002;39:920929.
23 . Barreto SM, Ladeira RM, Duncan BB, Schmidt MI, Lopes AA, Benseñor IM, et al. Chronic kidney disease among adult participants of the ELSA-Brasil cohort: association with race and socioeconomic position. J Epidemiol Community Health 2016;70:380389.
24 . Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF III, Feldman HI, et al.; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150:604612.
25 . Earley A, Miskulin D, Lamb EJ, Levey AS, Uhlig K. Estimating equations for glomerular filtration rate in the era of creatinine standardization: a systematic review. Ann Intern Med 2012;156:785795, W-270, W-271, W-272, W-273, W-274, W-275, W-276, W-277, W-278.
26 . Zanocco JA, Nishida SK, Passos MT, Pereira AR, Silva MS, Pereira AB, et al. Race adjustment for estimating glomerular filtration rate is not always necessary. Nephron Extra 2012;2:293302.
27 . Lotufo PA. Renal disease screening: a potential tool for reducing health inequity. Sao Paulo Med J 2016;134:12.
28 . Levey AS, de Jong PE, Coresh J, El Nahas M, Astor BC, Matsushita K, et al. The definition, classification, and prognosis of chronic kidney disease: a KDIGO Controversies Conference report. Kidney Int 2011;80:1728.
29 . Aielo AN, Santos RB, Silva WA, Parise BK, Souza SP, Cunha LF, et al. Pragmatic validation of home portable sleep monitor for diagnosing obstructive sleep apnea in a non-referred population: the ELSA-Brasil study. Sleep Sci 2019;12:6571.
30 . Berry RB, Budhiraja R, Gottlieb DJ, Gozal D, Iber C, Kapur VK, et al.; American Academy of Sleep Medicine; Deliberations of the Sleep Apnea Definitions Task Force of the American Academy of Sleep Medicine. Rules for scoring respiratory events in sleep: update of the 2007 AASM manual for the scoring of sleep and associated events. J Clin Sleep Med 2012;8:597619.
31 . Chowdhuri S, Quan SF, Almeida F, Ayappa I, Batool-Anwar S, Budhiraja R, et al.; ATS Ad Hoc Committee on Mild Obstructive Sleep Apnea. An official American Thoracic Society research statement: impact of mild obstructive sleep apnea in adults. Am J Respir Crit Care Med 2016;193:e37e54.
32 . Malachias MVB, Plavnik FL, Machado CA, Malta D, Scala LCN, Fuchs S. 7th Brazilian Guideline of Arterial Hypertension: chapter 1 – concept, epidemiology and primary prevention. Arq Bras Cardiol 2016;107(Suppl 3):16.
33 . Brewster UC, Setaro JF, Perazella MA. The renin-angiotensin-aldosterone system: cardiorenal effects and implications for renal and cardiovascular disease states. Am J Med Sci 2003;326:1524.
34 . Fields NG, Yuan BX, Leenen FHH. Sodium-induced cardiac hypertrophy: cardiac sympathetic activity versus volume load. Circ Res 1991;68:745755.
35 . Strazzullo P, D’Elia L, Kandala NB, Cappuccio FP. Salt intake, stroke, and cardiovascular disease: meta-analysis of prospective studies. BMJ 2009;339:b4567.
36 . Meneton P, Jeunemaitre X, de Wardener HE, MacGregor GA. Links between dietary salt intake, renal salt handling, blood pressure, and cardiovascular diseases. Physiol Rev 2005;85:679715.
37 . Elias RM, Chan CT, Paul N, Motwani SS, Kasai T, Gabriel JM, et al. Relationship of pharyngeal water content and jugular volume with severity of obstructive sleep apnea in renal failure. Nephrol Dial Transplant 2013;28:937944.
38 . Elias RM, Bradley TD, Kasai T, Motwani SS, Chan CT. Rostral overnight fluid shift in end-stage renal disease: relationship with obstructive sleep apnea. Nephrol Dial Transplant 2012;27:15691573.
39 . Yumino D, Redolfi S, Ruttanaumpawan P, Su MC, Smith S, Newton GE, et al. Nocturnal rostral fluid shift: a unifying concept for the pathogenesis of obstructive and central sleep apnea in men with heart failure. Circulation 2010;121:15981605.
40 . Jafari B, Mohsenin V. Overnight rostral fluid shift in obstructive sleep apnea: does it affect the severity of sleep-disordered breathing? Chest 2011;140:991997.
41 . Shechter A. Obstructive sleep apnea and energy balance regulation: a systematic review. Sleep Med Rev 2017;34:5969.
42 . Piovesana Pde M, Sampaio Kde L, Gallani MCBJ. Association between taste sensitivity and self-reported and objective measures of salt intake among hypertensive and normotensive individuals. ISRN Nutr 2012;2013:301213.
43 . Villela PTM, de-Oliveira EB, Villela PTM, Bonardi JMT, Bertani RF, Moriguti JC, et al. Salt preferences of normotensive and hypertensive older individuals. J Clin Hypertens (Greenwich) 2014;16:587590.
44 . Fitzsimons JT. Angiotensin, thirst, and sodium appetite. Physiol Rev 1998;78:583686.
45 . Kraly FS, Moore AF, Miller LA, Drexler A. Nocturnal food-related hyperdipsia in the adult spontaneously hypertensive rat. Physiol Behav 1982;28:885891.
46 . Cogswell ME, Maalouf J, Elliott P, Loria CM, Patel S, Bowman BA. Use of urine biomarkers to assess sodium intake: challenges and opportunities. Annu Rev Nutr 2015;35:349387.
47 . Gjørup PH, Sadauskiene L, Wessels J, Nyvad O, Strunge B, Pedersen EB. Increased nocturnal sodium excretion in obstructive sleep apnoea. Relation to nocturnal change in diastolic blood pressure. Scand J Clin Lab Invest 2008;68:1121.
48 . Fiori CZ, Martinez D, Montanari CC, Lopez P, Camargo R, Sezerá L, et al. Diuretic or sodium-restricted diet for obstructive sleep apnea-a randomized trial. Sleep 2018;41:DOI: 10.1093/sleep/zsy016.
Correspondence and requests for reprints should be addressed to Luciano F. Drager, M.D., Ph.D., Center of Clinical and Epidemiologic Research (CPCE), Av. Prof. Lineu Prestes, 2 565 – 4° andar, Cidade Universitária, São Paulo - SP, 05508-000, Brazil. E-mail: .

Supported by grants from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) (2012/02953-2 and 2019/23496-8) (L.F.D.).

Author Contributions: S.G., R.B.S., P.A.L., I.M.B., and L.F.D. contributed to the study design, analysis, interpretation of data, and writing manuscripts. L.F.D. contributed to the study design and study supervision. S.G., R.B.S., A.N.A., W.A.S., B.K.P., S.P.S., A.P.-A., L.A.B., P.A.L., I.M.B., and L.F.D. contibuted to the collection of data. A.P.-A. and L.A.B. contributed to interpretation of data and critical review of the manuscript.

Data sharing statement: The data from the present study are available on request, following the submission of a proper research protocol. All data are anonymized to respect the privacy of patients who have participated in the cohort according to applicable laws and regulations.

This article has a related editorial.

This article has an online supplement, which is accessible from this issue’s table of contents at

Author disclosures are available with the text of this article at

Comments Post a Comment

New User Registration

Not Yet Registered?
Benefits of Registration Include:
 •  A Unique User Profile that will allow you to manage your current subscriptions (including online access)
 •  The ability to create favorites lists down to the article level
 •  The ability to customize email alerts to receive specific notifications about the topics you care most about and special offers
Annals of the American Thoracic Society

Click to see any corrections or updates and to confirm this is the authentic version of record