Complications during Hemodialysis
Richard A. Sherman
John T. Daugirdas
Todd S. Ing
The most common complications during hemodialysis are, in descending order of frequency, hypotension, cramps, nausea and vomiting, headache, chest pain, back pain, and itching.
I. INTRADIALYTIC HYPOTENSION. Intradialytic hypotension (IDH) is important not only because it can cause distressing symptoms, but because it is associated with poor long-term outcomes. Patients with IDH show increased mortality (Flythe, 2014) and also an increased rate of cardiac wall motion abnormalities during dialysis, the so-called myocardial stunning (McIntyre and Odudu, 2014). There are various definitions for IDH, including a nadir (lowest) systolic BP less than 90 mm Hg, a fall in systolic BP of 20 or 30 mm Hg, or a fall in some percentage of the starting blood pressure. For quality assurance purposes, a definition of nadir systolic BP less than 90 mm Hg might be most useful as this has the strongest association with increased mortality (Flythe, 2014). The incidence of IDH is highest in patients with low predialysis blood pressure. A low predialysis blood pressure may be a marker of cardiac disease, and hearts with functional or structural abnormalities may be less able to compensate hemodynamically for fluid removal. IDH is also associated with an increased risk of access thrombosis (Chang, 2011). Mechanistic causes of IDH are detailed in Table 12.1.
A. IDH related to blood volume changes. Volume-related causes of IDH are most important in that the blood pressure during hemodialysis normally does not decrease (beyond an initial, trivial amount) if fluid is not removed. Thus, any maneuver that slows the ultrafiltration rate, whether it be extending the weekly time on dialysis, reducing the weekly volume of fluid ingestion, or increasing the volume of urine excreted, should reduce the rate of IDH.
1. Avoid large interdialytic weight gains. Emphasizing salt restriction is far more effective in decreasing interdialytic weight gain (IDWG) than focusing on fluid restriction (Tomson, 2001). Observational data find an association between a higher sodium intake and poor outcome (McCausland, 2012).
TABLE 12.1 Causes of Intradialytic Hypotension | ||||||||||||||||||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
2. Increasing weekly treatment time. Increasing weekly treatment time will, by definition, decrease the required ultrafiltration rate (same weight loss, longer time), and this will decrease the frequency of IDH. The weekend long interdialytic interval is associated with a higher IDWG; if the same postdialysis weight is targeted after the weekend, a higher ultrafiltration rate will, by definition, need to be used to achieve it. In-center dialysis patients with fluid removal problems are often treated on a Mon-Wed-Fri-Sat schedule. This cuts out the long-weekend interdialytic interval and also increases the weekly dialysis time.
The KDOQI 2006 adequacy guidelines recommend that treatment time not be reduced below 3 hours (for thrice-weekly dialysis) in patients with little or no residual urine output, regardless of how high their Kt/V may be. The European Best Practice Guidelines recommend that 4 hours of therapy should be provided for all patients dialyzed on a three-per-week schedule, regardless of body size. Increasing dialysis frequency without increasing weekly dialysis time does not always reduce IDH, although in one study the degree of myocardial stunning was reduced with short daily hemodialysis (Jeffries, 2011).
3. Maintaining and increasing urinary volume. In patients with residual kidney function, the amount of urine volume directly subtracts from the amount of fluid that needs to be removed during dialysis. Urine volume can be increased using diuretic therapy (Lemes, 2011).
4. Choose target weight carefully. A patient’s target weight or “dry weight” is usually chosen primarily on a clinical basis taking into account a patient’s blood pressure, presence of edema, and tolerance of ultrafiltration to the chosen weight. The decision can be aided with the results of testing that is slowly making its way into the clinic (e.g., bioimpedance devices, measurement of serum atrial natriuretic factor levels, relative blood volume monitors, and pulmonary ultrasound). The term “target weight” may be more appropriate than “dry weight,” because some level of volume overload is required in many patients to prevent IDH. This is because as the patient’s dry weight is approached, the rate at which the blood compartment refills from surrounding tissue spaces is diminished. Patients who require high ultrafiltration rates may be unable to reach their true dry weight because the progressively slower refill rate as dialysis proceeds provokes transient hypovolemia at the end of treatment, often accompanied by IDH cramps, dizziness, and postdialysis malaise. More ominously, hypoperfusion of the heart, brain, and gut may have cumulative adverse consequences.
Intradialytic hematocrit monitors may help recognize a dry weight that is too high. A “flat-line” hematocrit response (e.g., lack of an increase during dialysis) despite fluid removal indicates rapid blood compartment refilling and suggests fluid overload. However, a randomized trial in which these data were utilized clinically resulted (paradoxically) in an increased, rather than a decreased, hospitalization rate (Reddan, 2005). Identification of a specific level of hemoconcentration (“crash-crit”) does not appear to be useful in avoiding IDH.
The use of multifrequency bioimpedance devices to adjust target postdialysis weight is growing in popularity. Reduction of fluid overload results in a lower prevalence of left ventricular hypertrophy, a finding that is strongly associated with poor outcome. Trying to aggressively reduce blood pressure without technological guidance has been associated with increased IDH (Davenport, 2008), and with an increased rate of access failure and cardiovascular hospitalization (Curatola, 2011). Use of a multifrequency impedance monitor was associated with reduced blood pressure and left ventricular mass (Hur, 2013) without apparent side effects, although the rate of loss of urine volume was accelerated in the group using bioimpedance to lower target weight.
5. Use an appropriate dialysis solution sodium level. When the dialysis solution sodium level is less than that of plasma, the blood returning from the dialyzer is hypotonic with respect to the fluid in the surrounding tissue spaces. To maintain osmotic equilibrium, water leaves the blood compartment, causing an acute reduction in the blood volume. Higher dialysis solution sodium levels limit the reduction in
blood volume accompanying ultrafiltration, but they also increase IDWG, blood pressure, and postdialysis thirst.
blood volume accompanying ultrafiltration, but they also increase IDWG, blood pressure, and postdialysis thirst.
So-called sodium modeling (or sodium gradient dialysis) is widely practiced. It generally involves use of a high dialysis solution sodium early in treatment (145-155 mM) with a progressive fall (linear, step, or logarithmic) to lower levels (135-140 mM) at the end of treatment. The objective is to obtain the benefits of high-sodium dialysis solution without its complications. Review of the large literature on this subject shows that sodium modeling is of uncertain benefit (Stiller, 2001). It should also be noted that a patient’s postdialysis serum sodium is a function of a treatment’s time-averaged concentration of dialysis solution sodium, not the terminal level of dialysis solution sodium.
Instead of a “one size fits all” level of dialysis solution sodium, using a fixed level set close to the patient’s predialysis serum value—an “individualized” dialysis solution sodium— may reduce symptoms as well as interdialytic thirst (Santos, 2010). Recent data indicate that using a relatively high dialysis solution sodium (>142 mmol/L) may benefit frail patients at high risk for IDH, likely because the consequences of recurrent IDH are more dire than those from using a high-sodium dialysis solution (Marshall and Dunlop, 2012). On the other hand, use of a relatively low dialysate sodium level can reduce IDH because it tends to lower IDWG and the need for ultrafiltration (Shah and Davenport, 2012).
6. Blood volume control devices with feedback loop. For a number of years now, software has been allowing improved feedback control of ultrafiltration rate based on monitoring of blood volume during dialysis. Some randomized trials suggest that such feedback devices can reduce the incidence of dialysis-induced hypotension while avoiding a positive sodium balance (Davenport, 2011).
B. Hypotension related to lack of vasoconstriction. In the hypovolemic state, cardiac output is limited by cardiac filling; a reduction in either peripheral vascular resistance or cardiac filling in this setting can precipitate hypotension. Under conditions of decreased cardiac filling, increases in heart rate have little effect on cardiac output. Because more than 80% of the total blood volume circulates in veins, changes in venous capacity can have important effects on a patient’s effective circulating blood volume and cardiac output. When arteriolar resistance decreases, more arterial pressure is transmitted to veins, causing passive stretching and distension, and an increased sequestration of blood. While not important in euvolemic patients given a vasodilator (because cardiac filling is more than adequate), this mechanism can result in hypotension when hypovolemia is present (Daugirdas, 1991). The degree of arteriolar constriction, or total peripheral resistance (TPR), is also important because TPR will determine the blood pressure for any level of cardiac output.
1. Lower dialysis solution temperature. Ideally, the dialysis solution temperature should be one that maintains a patient’s arterial blood temperature at its initial level throughout dialysis. When the dialysis solution temperature is higher than this ideal level, cutaneous vasodilation occurs to allow heat to be dissipated. This vasodilation reduces vascular resistance and predisposes the patient to hypotension. Blood temperature modules are available for dialysis machines, which can provide patients with a euthermic treatment. Without such a device, the choice of dialysis solution temperature is problematic, with even small (1.1°C) differences in temperature having a notable impact on blood pressure (Sherman, 1984). The widely used dialysis solution temperature of 37°C is almost always in excess of euthermic values. Levels of 35.5°C-36.0°C are better initial choices, with adjustment made up or down depending on tolerance (chills) and effectiveness (blood pressure). Cool dialysis solutions cause patient discomfort only when the dialysate temperature is below the optimal (usually unknown) level; euthermic dialysis is not associated with shivering and only rarely with chills (Maggiore, 2002). One group has favored individualizing dialysis solution temperature at the patient level. The tympanic membrane temperature is measured, and the dialysis solution temperature is set 0.5°C below this level. This system of individualized cooling has been shown to avoid the sensation of cold and chills commonly found with simply lowering dialysate temperature to a given level for all patients (Odudu, 2012). Individualized, cooled dialysate is associated with a shorter postdialysis recovery time, better maintenance of blood pressure, reduced myocardial stunning, and less evidence of progressive ischemia-related brain white matter damage (McIntyre, 2014).
A number of studies have found that hemodiafiltration is associated with a better tolerance to ultrafiltration and less IDH than hemodialysis. However, it appears that the beneficial effect of hemodiafiltration may have been due primarily to a lower extracorporeal circuit temperature due to the cooling effect of the replacement solution. When heat transfer from the extracorporeal circuit was kept constant, the hemodiafiltration advantage over hemodialysis with regard to blood pressure was no longer found (Kumar, 2013).
2. Avoid intradialytic food ingestion in hypotension-prone patients. Eating during hemodialysis can precipitate or accentuate a fall in blood pressure (Sherman, 1988; Strong, 2001). The effect is probably a result of dilation of resistance vessels in the splanchnic bed, which reduces TPR and increases splanchnic venous capacity (Barakat, 1993). The “food effect” on blood pressure probably lasts at least 2 hours. Patients who are prone to hypotension during dialysis should avoid eating just before or during a dialysis session.
3. Minimize tissue ischemia during dialysis. During any type of hypotensive stress, the resulting tissue ischemia causes release of adenosine. Adenosine blocks release of norepinephrine from sympathetic nerve terminals and also has intrinsic vasodilator properties. Severe hypotension can therefore amplify itself: Hypotension → ischemia → adenosine release → impaired norepinephrine release → vasodilation → hypotension.
This may be one reason for the clinical observation that patients with low hematocrit levels (e.g., <20%-25%) are very susceptible to dialysis hypotension (Sherman, 1986). Few patients currently have levels of anemia severe enough to cause hypotension. Such patients may benefit from transfusion, although the current trend is to strongly discourage transfusion of acutely ill patients in an intensive care setting. Use of nasal oxygen in hypotension-prone patients may be another way of limiting tissue ischemia and IDH (Jhawar, 2011).
4. Midodrine. Midodrine, an orally acting α-adrenergic agonist, reduces the frequency of IDH. A dose of 10 mg orally 1.5-2 hours before a dialysis session is typical, though use of as much as 40 mg has been reported. Supine hypertension is the major dose-limiting factor. Active cardiac ischemia (but not simply coronary artery disease) is a contraindication. Concomitant use of α-adrenergic blockers renders midodrine ineffective. No data exist as to whether the drug is especially useful in patients with autonomic insufficiency (half of the dialysis population) as might theoretically be the case. One problem with midodrine is that its effect does not seem additive to that of using cool dialysate (Cruz, 1999).
5. Sertraline. At least three reports have indicated that 4 to 6 weeks of therapy with the selective serotonin reuptake inhibitor sertraline reduces the frequency of IDH. Some evidence suggests that the drug improves autonomic function (Yalcin, 2003). Like midodrine, sertraline was not shown to give added protection against IDH when cool dialysate was used (Brewster, 2003).
6. Antihypertensive medication. Antihypertensive medications administered prior to dialysis adversely impact the ability of the cardiovascular system to adjust to volume removal. Whether those with vasodilatory properties are more problematic than those with other mechanisms of action has not been well studied.
7. Dialysis fluid potassium level. A low concentration (1 mEq/L) of dialysis fluid potassium is associated with more frequent IDH, perhaps via autonomic effects. If otherwise possible, using higher potassium levels is advisable for hemodynamic benefits as well as reduced arrhythmogenic effect.
8. Fludrocortisone. One preliminary report (Landry, 2011) found low random aldosterone levels in a group of five dialysis patients with low predialysis blood pressures and refractory
IDH. All had a normal cosyntropin test. Their blood pressures improved with fludrocortisone treatment, ultrafiltration volumes increased, and rate of IDH was reduced. There was no improvement with fludrocortisone in hypotensive patients with normal levels of adrenal hormones.
IDH. All had a normal cosyntropin test. Their blood pressures improved with fludrocortisone treatment, ultrafiltration volumes increased, and rate of IDH was reduced. There was no improvement with fludrocortisone in hypotensive patients with normal levels of adrenal hormones.
9. Vasopressin. Vasopressin levels normally increase with hypotension, but in dialysis patients, the increase is often suboptimal. Vasopressin preferentially constricts splanchnic vessels, and such constriction may help to redistribute blood volume centrally during fluid removal. In one study, vasopressin infusion reduced the incidence of IDH (van der Zee, 2007).
C. Hypotension related to cardiac factors
1. Diastolic dysfunction. A stiff, hypertrophied heart is especially prone to a reduction in output when there is even a minor reduction in filling pressure. Diastolic dysfunction is common in dialysis patients owing to the effects of hypertension, coronary artery disease, and probably uremia itself. Some limited data suggest that using verapamil as an antihypertensive agent may reduce the frequency of IDH in such patients.
2. Heart rate and contractility. Most, but not all, dialysis hypotension is associated with decreased cardiac filling, a setting in which cardiac compensatory mechanisms can do little to increase output. In some patients, TPR may fall (owing to temperature effects, food ingestion, or tissue ischemia) without a fall in cardiac filling. In this setting, impairment of cardiac compensatory mechanisms can play a direct role in the development of hypotension.
3. Dialysis solution calcium. A dialysis solution calcium concentration of 1.75 mM increases cardiac contractility and helps maintain intradialytic blood pressure better than a level of 1.25 mM, especially in patients with cardiac disease (van der Sande, 1998). However, in the chronic outpatient setting (as opposed to an intensive care unit), symptomatic IDH is not less frequent with a higher-calcium dialysis solution (Sherman, 1986); use of high dialysis solution calcium levels may contribute to vascular calcification, and the trend is to not use them for prolonged periods. Dialysis solution magnesium levels may impact dialysis hypotension, but whether a higher or a lower level should be used is controversial (Chapter 10).
D. Unusual causes of hypotension during dialysis. Rarely, hypotension during dialysis may be a sign of an underlying, serious event. Causes are listed in Table 12.1.
E. Detection of hypotension