Sepsis, Diastolic Dysfunction & Hypernatremia
Jon-Emile S. Kenny MD [@heart_lung] with illustrations by Carla M Canepa MD
“… And you may ask yourself … well, how did I get here? ... And you may tell yourself … my God! What have I done?"
A 92 year old woman is transferred to the coronary care unit for treatment of pulmonary edema. She was initially admitted to the hospital one week prior from a rehab facility for treatment of pneumonia. On initial presentation she was given 2 litres of saline, vancomycin and piperacillin-tazobactam 4.5 grams every 6 hours. Over the course of 5 days, her inflammatory symptoms improved, however her intake was diminished and she was placed on 75 mL/hour of 0.9% saline. For 48 hours prior to transfer to the CCU she was notably more short of breath and displayed evolving lower extremity edema. 3 months ago, her transthoracic echocardiogram demonstrated a normal ejection fraction with an E/e’ of 16, dilated left atrium and an elevated RVSP. The floor team placed her on fluid restriction and standing IV furosemide. In the CCU, she was noted to have persistent bilateral notable LE edema, bilateral B-lines and pleural effusions on lung ultrasound; as well, she had a serum sodium of 159 mmol/L.
Diastolic dysfunction complicated by clinical heart failure is a complex interplay between the heart and the vascular system. One may be tempted to be more liberal with intravenous fluids in patients with diastolic dysfunction – with fear allayed by the words ‘preserved ejection fraction.’ However, patients with this form of heart failure are just as insulted by excessive salt and water as their counterparts with reduced ejection fraction.
The abnormality undergirding diastolic dysfunction is an impaired compliance, or increased ‘stiffness’ of the heart at end-diastole. This insult may be visualized with the end-diastolic pressure volume relationship [see figure 1]. The curve represents the pressure within the heart over a series of volumes at the very end of diastole. Normally, the heart at end-diastole is quite compliant with minimal increase in filling pressure with fairly large amounts of volume. With stiffening of the left ventricle – for example from years of hypertension, or stiff central arteries – there is a higher left heart filling pressure for the same filling volume.
Figure 1: The end-diastolic pressure volume relationship [EDPVR] between normal [blue curve, high compliance] & those with impaired relaxation or decreased lusitropy [red curve]. For the same filling volume [x-axis] there is a much higher filling pressure [y-axis] for the red curve. LVEDP is left ventricular end-disatolic pressure.Yet, beyond the heart, heart failure with preserved ejection fraction is typified by much stiffer arteries and a lower venous capacitance compared to heart failure patients with reduced ejection fraction. A potential implication of both rigid arteries and veins is a higher intra-capillary pressure and, consequently, greater transudation of fluid into the interstitium. As the right atrial pressure is the back pressure for lymphatic return, the high prevalence of right heart hypertension in diastolic dysfunction favours interstitial edema [Figure 2]. Indeed, in a fascinating study, patients with heart failure and a preserved ejection fraction had a greater volume of excess fluid distributed to the interstitial space than those with a reduced ejection fraction.
Figure 2: Patients with diastolic dysfunction have stiffer arteries and decreased venous capacitance as compared to those with reduced ejection fraction. P is pressure. This physiology may favour increased intra-capillary pressure and therefore interstitial edema - potentially within the lungs as well.
Diastolic Dysfunction & Sepsis
Given that patients with heart failure with preserved ejection fraction have ‘stiffer’ arteries and veins than their reduced-ejection counterparts, it is understandable that those with compromised diastology are more sensitive to vasodilators. Thus, the vaso- and venodilation of sepsis can have significant hemodynamic consequences including precipitous attenuation of mean arterial pressure and venous return. Excessive fluid administration may ‘hide’ within the increased venous capacitance, especially in the splanchnic venous beds. As sepsis abates, venous capacitance is reduced to baseline and blood volume is then redistributed from the splanchnic vasculature to the central compartment raising the risk of acute pulmonary edema.
While underappreciated until relatively recently, the problem of aberrant LV relaxation in sepsis is gaining acceptance. Recent echocardiographic evaluations have demonstrated that impaired LV relaxation in the acute stages of sepsis is actually quite common. Importantly, a stiffer left ventricle will result in a greater intracardiac pressure following volume infusion, increased wall stress and, consequently, elaboration of natriuretic peptides. Curiously, these peptides cleave the glycocalyx – rendering capillaries even more ‘leaky;’ a vicious cycle is thus perpetuated and particularly concerning in the pulmonary vascular tree.
Hypernatremia in Heart Failure
The hypernatremia on presentation to the CCU is almost certainly due to a combination of free water restriction – instituted by the primary team – as well as hypotonic urinary losses in response to loop diuretics. The reasoning behind free water restriction in normo-natremic patients with volume overload continues to elude me. As we are all taught in medical school, free water enters and leaves the body compartments in proportion to the relative sizes of said compartments. Normally, the intra-cellular compartment is the largest; thus, free water restriction mostly dehydrates the intra-cellular space with much less benefit to the interstitial space and minimal effect on the intra-vascular space [see figure 3].
Figure 3: Illustration demonstrating volume of distribution of free water in an elderly woman. A 4 litre free water deficit affects each compartment proportionally. Accordingly, a free water deficit has the greatest absolute impact on the intra-cellular space. This diagram assumes the intra-cellular space is roughly 2/3 of the body's water sink. In the overloaded patient, the extra-cellular space is expanded, but the principle is the same.
Even if the interstitial and intravascular space are expanded in heart failure and increase their proportional sizes, free water restriction is still – physiologically – ‘robbing Peter to pay Paul.’ The free water debt from the intra-cellular space must be recompensed because hypernatremia has adverse effects on cellular function as briefly described below.
In an old, esoteric canine study, the cardiac effects of independently-altered ambient sodium and osmolality were studied. The authors found that hypernatremia itself decreased cardiac contractility and that this was thought secondary to depletion of intracellular calcium currents – most likely from sodium-calcium antiports. Additionally, high osmolality also reduced contractility, but this did not occur until very high serum osmolality; the mechanism behind this effect was likely true intra-cellular dehydration, increased intracellular viscosity and embarrassed contractile elements. Of great interest is a letter-to-the editor describing a patient whose left heart filling pressure fell following many litres of free water replacement – in theory due to the enhanced contractility afforded by repletion of the free water deficit.
Figure 4: The top cartoon represents the normal patient. The bottom cartoon represents hypernatremia in diastolic heart failure. While the vascular and interstitial spaces [the extra-cellular compartment] are relatively expanded, the intra-cellular space is dehydrated.There are many cellular effects of hypernatremia throughout the body, but within the central nervous system, hypernatremia will lead to profound thirst; hypernatremia alters neuronal metabolism by similar mechanisms at play in cardiac myocytes; delirium and agitation may be corollaries.
Having had a family member in the ICU, the importance of unrecognized thirst is never lost on me. In my opinion it is cruel and unusual to leave patients in states of hypernatremia while simultaneously free water restricting them. Recall that in the 8th Circle of Dante's The Inferno, that counterfeiters of money were condemned to an eternity of anasarca, immobility and profound thirst. Do not treat your patients as if they are in Dante’s Inferno.
Return to the Case
The patient is edematous because she received extraordinary amounts of salt. Take, for instance, her maintenance fluids. 0.9 per cent saline means that there are 0.9 grams of sodium chloride per hundred mL. In other words, 900 milligrams of NaCl per 100 mL [360 mg of sodium alone per 100 mL]. She received 75 mL per hour or 270 mg of sodium per hour – 6.5 grams per day! Additionally, 4.5 grams of piperacillin-tazobactam contains 260 mg of sodium per dose, or roughly 1 gram of sodium per day. In totality, her maintenance fluids and antibiotics alone approached 8 grams of sodium per 24 hours!
For reference, there is about 1 gram of sodium in 1 Big Mac and heart failure patients should be restricted to no more than 2 grams of sodium per day.
To correct for her free water deficit, free water is given back. It is not a sin to give diuretics and free water simultaneously – if there is a free water deficit. While she is likely losing free water from hyperventilation, she is also having excessive free water losses from her loop diuretic, especially if there is distal tubular sodium reabsorption – as can occur in response to chronic loop diuretics. The goal in heart failure is to rid the body of isotonic fluid, not free water. To increase renal natriuresis and enhance isotonic fluid loss, metolazone was added.
With the additional assistance of high-flow oxygen delivery and salt-restriction, she lost many litres of urine; her edema and work-of-breathing improved markedly and her serum sodium fell to 142 mmol/L.