Edema is defined as an abnormal buildup of fluid anywhere in the body. When utilized with no qualifier, the term “edema” usually refers to the accumulation of plasma transudate in the interstitial space, as in CHF, nephrotic syndrome, and hepatic cirrhosis (24–26).

For fluid to accumulate at the interstitial space, a positive sodium balance must establish. Since the kidneys are ultimately responsible for maintaining sodium balance, it follows that edema formation always demands some degree of renal sodium retention. Nevertheless, impaired sodium excretion is insufficient to ensure fluid accumulation. For instance, in primary hyperaldosteronism, excess sodium reabsorption by the distal nephron translates into hypertension, rather than edema formation. To reach the interstitial space, fluid retained by the kidneys must be driven by an imbalance of Starling forces at the complex interface between the intravascular and interstitial compartments (27).

Under normal conditions, small amounts of fluid do reach the interstitial compartment due to a slight predominance of hydrostatic over oncotic forces. Actually, the normal interstitium contains about 10 L of fluid, an amount kept within narrow limits by three mechanisms (27): the action of lymphatic capillaries, carrying extravasated fluid back to the circulation; the dilution of interstitial protein that results from transcapillary fluid passage; and the tight disposition of the protein molecules that constitute the interstitial matrix—due to this arrangement, substantial elevation of local hydraulic pressure is required to accommodate even small amounts of extra fluid (low interstitial compliance).

An important consequence of these physical characteristics of the normal interstitial matrix is that fluid cannot move freely across the interstitium following gravity and, therefore, will not accumulate in the lower limbs while standing, or in the cervical region after several hours in the recumbent position.

Although the classical view centered on Starling forces still predominates, recent evidence suggests that this theory should be revised taking into account the segmentation of the capillary wall and the adsorption of ions by interstitial macromolecules (28, 29). An overview of the mechanisms evolved in edema formation is summarized in Figure 1.

When sodium excretion is hindered by intrinsic renal disease, a positive sodium balance establishes, leading to expansion of the extracellular (EC) and intravascular volumes. If Starling equilibrium is not disrupted, sodium retention will cause hypertension, according to Guyton’s theory (27), but not edema. However, if tissue autoregulation fails, capillary hydraulic pressure will rise, and the resulting imbalance of Starling forces will lead to fluid extravasation. This mechanism of fluid retention, known as overflow (also overfill), operates in primary renal disease, such as glomerulonephritis (30) and advanced chronic kidney disease (CKD). As EC volume is expanded, mechanisms that increase sodium excretion are triggered, counteracting the renal limitation and allowing a new sodium balance to be reached. For this reason, edemas accumulated by overflow are relatively modest and confined to the lower limbs and eyelids.

In patients on chronic dialysis, maintenance of fluid balance is entirely dependent on an artificial procedure. Therefore, development of edema in this context usually results from insufficient fluid removal (31) or poor adherence to treatment. Thus, the mechanism of edema formation in these patients can be considered as analogous to overflow.

In a number of situations, effective arterial blood volume (EABV), hence renal perfusion, cannot be maintained despite normal or even increased total blood volume. In this context, the kidneys (assumed to be normal) react to the reduction of EABV by retaining sodium and water, which nevertheless escape the intravascular space because of a disequilibrium of Starling forces, promoting further sodium retention. In this manner, the retained fluid tends to accumulate at the interstitial compartment, instead of recomposing the EABV (32).

This mechanism of edema formation resulting from chronic reduction of the EABV is known as underfilling. Unlike what happens with overflow, here renal dysfunction is not the primary cause of sodium retention. Rather, the kidneys act as expected, responding to hypoperfusion by reabsorbing as much sodium as possible. Underfilling is central to the pathogenesis of edema in CHF, hepatic cirrhosis and some cases of nephrotic syndrome.

Under normal conditions, the heart easily meets the needs of all tissues, keeping cardiac output at physiological levels. In CHF, the weakened myocardium can no longer maintain adequate perfusion of the peripheral territories, including the renal circulation. The consequent fall of EABV stimulates the kidneys to retain sodium. On the other hand, the malfunctioning pump leads to venous damming of blood. Retrograde transmission of the resulting venous hypertension to the capillaries promotes the passage of fluid to the interstitial space. This process is continuously fueled by the renal retention of sodium, which nevertheless fails to restore the EABV. Therefore, the basic mechanism of edema formation in CHF is underfillling (33).

About one-third of patients with nephrotic syndrome exhibit clear signs of hypovolemia despite massive EC fluid expansion. In these patients, edema formation is believed to result from hypoalbuminemia, hence decreased systemic oncotic pressure, leading to an imbalance of starling forces, fluid displacement to the interstitium, EABV reduction, and incessant renal sodium retention. This sequence is fully compatible with the concept of underfilling (25, 34). However, two-thirds of nephrotic patients exhibit clear clinical evidence of fluid overload. It is believed that, in these patients, the basic event is primary sodium retention by the kidneys, with hypoalbuminemia facilitating ultrafiltration through the capillary walls, so that the magnitude of swelling is much higher than in the nephritic syndrome. Therefore, the basic mechanism of edema formation in most cases of nephrotic syndrome is overflow, facilitated by the simultaneous decrease in plasma oncotic pressure (2, 12).

If fluid escape into the interstitium persists, the initially slow accumulation of edema raises gradually the local hydraulic pressure, until it becomes positive. When this happens, the normally tight architecture of the interstitium is disrupted, leading to an abrupt increase of compliance, enabling the interstitium to accommodate increasing amounts of fluid with a small rise of hydraulic pressure. The shift of fluid throughout the interstitium is no longer restricted, being now governed by gravity: during daytime, edema accumulates in the lower limbs; at night, interstitial fluid tends to be redistributed rostrally, reaching the cervical region. These movements largely explain the occurrence of airway obstruction and SA in edematous states.

Sleep apnea can be classified as obstructive sleep apnea (OSA), associated with airway obstruction and, therefore, respiratory effort, or central sleep apnea (CSA), in which the main pathogenic factor is respiratory center instability. The most common sleep disorder is OSA. CSA is far less common although equally as dangerous as OSA.

In CHF, fluids displaced from the lower body during the night can accumulate at cervical and head areas, thus promoting upper airway obstruction and OSA. It has been postulated that systemic fluid retention, with consequent venous engorgement and mucosal fluid accumulation, can increase tongue volume, facilitating airway obstruction (39). Of note, fluid accumulation in the neck, causing mucosal edema and OSA, was seen in healthy men after IV saline infusion during sleep (40).

In men with CHF and OSA, Kasai and colleagues showed that application of lower body positive pressure (LBPP) in the awake state, thus forcing rostral fluid shift, was accompanied by a significant increase in neck circumference and an increase in upper airway resistance in proportion to the volume of fluid displaced from the legs (41). Interestingly, the relationship between rostral fluid shift and OSA in CHF is less pronounced in women (13, 42).

In CHF patients, fluid retention and fluid shift from the legs can also lead to pulmonary congestion. In this case, however, SA is unrelated to obstruction. Rather, it seems to result from a central respiratory mechanism, thus conforming to the CSA type.

The mechanism underlying the establishment of CSA in these patients has not been fully elucidated, although pulmonary congestion, increased central and peripheral chemosensitivity, and frequent arousals may play a role (43, 44). Pulmonary congestion, a common finding in CHF, can stimulate so-called pulmonary vagal irritant “J” receptors (45), causing reflex inhibition of the respiratory drive through afferent C fibers. The consequent apnea causes PaCO₂ to increase, now leading to hyperventilation and generating a Cheynes–Stokes-like pattern (35, 37, 46). In consistency with this concept, PaCO₂ in CHF is inversely proportional to pulmonary capillary wedge pressure (47), which is an index of pulmonary congestion (48).

Dietary sodium intake can be associated with the severity of both OSA and CSA in CHF patients, with increased sodium intake presumably resulting in worsening of edema around the upper airway (OSA) and/or pulmonary congestion and CSA, through the mechanisms discussed earlier. Increased leg fluid retention, and consequently nocturnal overnight rostral fluid shift, can also be favored by excessive sodium intake (49).

Rostral fluid shift may exert a similar influence in CKD as in CHF. In a study of 26 patients on conventional hemodialysis, Elias and coworkers (14) showed that SA, present in 46.1% of subjects, was associated with age, male gender, and time spent in the sitting position during the day. Rostral fluid shift correlated significantly with the severity of SA and with the overnight increase of neck circumference. In a related study, fluid shift was shown to correlate with the increase of internal jugular vein volume, mucosal water content, and AHI (60). Likewise, Lyons et al. (61) showed a correlation between the magnitude of rostral fluid shift and the severity of both OSA and left atrial size in 40 patients on conventional hemodialysis, reinforcing the view that fluid shift may have an impact on both OSA and cardiac dysfunction in ESRD. The proof of concept that fluid overload can impact in the severity of OSA in patient on dialysis was demonstrated by Lyons and coworkers, who showed that AHI fell by 36% after removal of an average of 2.2 L by ultrafiltration alone in ESRD patients (62).

The importance of fluid retention in the pathogenesis of SA is not restricted to CKD and ESRD. In patients with nephrotic syndrome, even with normal renal function, the treatment of hypervolemia, with contraction of EC volume and disappearance of lower limb edema, was shown to alleviate SA (15).

In summary, patients with kidney disease, particularly those on dialysis, and patients with nephrotic syndrome are more prone to have SA. The role of fluid overload and overnight fluid shift as risk factors for SA were well demonstrated in these settings (14, 15, 54, 60, 62–66). The data presented in these cited studies suggest that kidney disease and nephrotic syndrome might cause SA independently of confounding factors. Fluid overload per se contributed to the presence of SA in patients on hemodialysis that can be partly reversible through fluid removal by ultrafiltration (67).

Basically, the therapeutic action of CPAP is to mechanically impede the collapse of the upper airways, thus preventing OSA. In addition, CPAP prevents CSA because it maintains a continuous airflow. CPAP seems to have no effect on overnight fluid shift in patients on hemodialysis (68), although is considered the mainstream treatment for OSA regardless of volume overload given its mechanism of action.

Targeted therapy for fluid retention and/or rostral fluid shift has been tested in several studies. In CHF patients with left ventricular diastolic dysfunction and severe OSA, intensive diuretic therapy increased upper airway cross-sectional area and lowered AHI by 24% (69). Increased physical activity during cardiac rehabilitation has also been associated with attenuation of both OSA and CSA (70, 71), possibly by preventing lower body fluid accumulation.

Head-elevated patient positioning can ameliorate OSA in CHF by preventing cervical fluid accumulation (72). Interestingly, this maneuver can also prevent CSA in CHF patients. Similar results were obtained in CHF patients with predominant OSA (72), with no effect on thoracic fluid content or left ventricular hemodynamics (73). This effect was attributed to increased venous return and dilation of the left heart while in the supine position, although lung congestion may also play a role.

As remarked earlier in this review, overflow is the mechanism of fluid retention in patients on hemodialysis. Accordingly, amelioration of SA in this population is expected to be proportional to the efficiency of fluid removal. In patients on conventional hemodialysis (three 4-h sessions/week), the SA severity tends to increase during interdialytic periods, reaching a maximum immediately before each session (12, 16, 51). Therefore, increasing the duration and/or frequency of sessions, thus mimicking more faithfully the operation of normal kidneys, may be a sound strategy to prevent SA in ESRD. In a study of 14 patients transferred from conventional to intensive hemodialysis (five 6-h sessions/week), AHI decreased by 68%, with marked improvement of oxygen saturation (63). A similar trend was observed in patients treated with peritoneal dialysis (PD) (74). Tang et al. (65) showed that in patients undergoing automated nocturnal PD, in which fluid removal was more efficient than with the manual procedure, a greater reduction of AHI was achieved, in association with less airway obstruction.

The importance of fluid retention is highlighted by the behavior of SA in patients with nephrotic syndrome, even when renal function is normal. In patients with steroid-responsive nephropathy, Tang et al. (15) showed that the severity of SA was reduced after kidney disease remission, in association with disappearance of lower limb edema and reductions in body water content.

Reducing leg swelling by wearing compression stockings during the day attenuated SA in patients with venous insufficiency (75, 76) and ESRD (68). This beneficial effect was observed in a general OSA population (77), highlighting the impact that even small amounts of fluid retained in the legs during the day might have on SA.

Table 2 summarizes several studies in which the efficacy of the aforementioned therapies was tested.