The incidence of treated end-stage renal disease in the United States is 180 cases per 1 million persons and continues to increase at a rate of 7.8% per year. In 1990, more than 45 000 new patients with end-stage renal disease were enrolled in the Medicare program, of which 66% were white, 28% were African-American, 2% were Asian or of Pacific Island descent, and 1% were Native Americans. Forty-three percent of the patients were at least 64 years of age, and fewer than 2% were younger than 20 years of age. On average, African-American and Native American patients are younger at the onset of treated disease and show dramatically higher incidence rates than do white or Asian or Pacific Island patients. Although clinical experience suggests that the incidence in Hispanics is also greater than that in whites, data from the United States Renal Data System are not available to confirm this clinical impression. Hypertension and diabetes accounted for 63% of the new cases in 1990. The incidence of cases of diabetic end-stage renal disease in Native Americans was almost twice that of African-Americans and six times that of whites.

Of the more than 195 000 patients with end-stage renal disease receiving renal replacement therapy during 1990, 70% were treated with either hemodialysis or peritoneal dialysis. Although kidney transplantation is the treatment of choice for many patients with end-stage renal disease, the increase in time spent waiting for cadaveric organs, the presence of disqualifying comorbid conditions, and the low transplantation rates in an aging population will probably ensure that dialysis remains the primary method of renal replacement therapy in the foreseeable future.

According to federal, state, and private funding sources, the cost for care of patients with the disease was approximately $7.26 billion in 1990, an increase of 21% over a similar estimate for the previous year. Not reflected in this Figure are additional expenditures for outpatient drugs and supplies, the cost of disability, and Social Security payments. As the U.S. population continues to grow and a larger proportion of the population at risk reaches age 65 years and older, the cost of kidney disease, including this end-stage component, is projected to increase. According to an analysis done by the Health Care Financing Administration, by the turn of the century it is estimated that more than 300 000 patients will be enrolled in the program. Furthermore, 85 000 new patients will enter the program in the year 2000 alone. Most of the increase will come from the elderly and the diabetic population.

Despite improvements in dialysis technology over the past decade, mortality remains high. For example, at age 49 years, the expected duration of life of a patient with end-stage renal disease is 7 additional years compared with approximately 30 additional years for a person of the same age from the general population. In addition to increased mortality, patients also experience significantly greater morbidity, including a substantial loss in quality of life. In 1986, for example, for all Medicare patients older than 65 years, hospitalization averaged 2.8 days per year, whereas the median time for patients who have received 1 year of dialysis was 15.0 days per year. The relevant information available to prescribe the appropriate dialysis dose is limited and subject to gross errors. Thus, what constitutes an adequate dialysis dose remains a controversial question among professionals caring for patients receiving dialysis.

To resolve questions about delivered dialysis dose, comorbid conditions, and dialysis-related complications, all of which appear to cause increased morbidity and mortality in U. S. patients receiving dialysis compared with those in certain European countries and Japan, the National Institute of Diabetes and Digestive and Kidney Diseases and the Office of Medical Applications of Research of the National Institutes of Health convened a consensus development conference on 1 to 3 November 1993. After one and a half days of testimony by experts in the field, a consensus panel representing the professional fields of general medicine, nephrology, pediatrics, biostatistics, nutrition, and nursing, and a representative of the public considered evidence and agreed on answers to the following questions: 1) How does early medical intervention in predialysis patients influence morbidity/mortality? 2) What is the relationship between delivered dialysis dose and morbidity/mortality? 3) Can comorbid conditions be altered by nondialytic interventions to improve morbidity/mortality in dialysis patients? 4) How can dialysis-related complications be reduced? and 5) What are the future directions for research in dialysis?

How Does Early Medical Intervention in Predialysis Patients Influence Morbidity/Mortality?

It is clear that factors influencing the morbidity and mortality of patients receiving dialysis exist for an extended period before end-stage renal disease develops and the need for dialysis is imminent. Unfortunately, few patients (20% to 25%) are referred to a renal physician before dialysis therapy begins. Managed care programs must recognize the importance of the continued involvement of the renal team in the care of these patients. Several conditions related to renal failure are present before dialysis therapy begins, including anemia, hypertension, malnutrition, renal osteodystrophy, lipid abnormalities, and metabolic acidosis. In addition, smoking and poor glycemic control in diabetic patients will influence subsequent morbidity and mortality. The costs of delayed referral include both emergency dialysis, with its higher morbidity and mortality, and excessive use of health care dollars. Emergency dialysis jeopardizes the choice of the mode of dialysis, endangers the ability to maintain prolonged vascular access, precludes psychological preparation of the patient for care, and necessitates hospitalization for a catastrophic complex illness. The mortality in this crisis situation can be as high as 25%.

In the patient with progressing renal insufficiency, early intervention should be aimed at reversal of hypertension and correction of identified nutritional deficiencies and acidosis. Although data are limited, the use of erythropoietin will prevent severe anemia and may reverse its associated complications. No consensus exists on the ultimate role of dietary protein restriction in slowing the progression of renal failure. However, an intake protein level of 0.7 to 0.8 g/kg per day can maintain nutritional status in noncatabolic patients with end-stage renal disease without placing an undue burden on the capacity to eliminate potentially toxic metabolites, including acid, potassium, sulfate, phosphorus, magnesium, and unidentified uremic toxins. Because of the deleterious effects of parathyroid hormone, therapies aimed at prevention or reversal of secondary hyperparathyroidism should be initiated before dialysis therapy begins.

A patient should be referred to a nephrologist when the serum creatinine level has increased to 1.5 mg/dL in women and to 2.0 mg/dL in men. Later, referral to a renal team consisting of a nephrologist, dietitian, nurse, social worker, and mental health professional allows time to establish a working relationship, to acquaint the patient with the various modes of renal replacement therapy, and to provide information on dialysis access, nutritional modification, avoidance of potentially nephrotoxic drugs, and potential financial support for services. To reduce mortality and morbidity as soon as possible, it is essential to initiate the medical interventions discussed below.

Hypertension

Increasing evidence suggests that aggressive treatment of hypertension in the period before dialysis therapy begins delays progression of renal disease and is the most potent intervention to decrease subsequent cardiovascular mortality in patients receiving dialysis. As in patients without renal disease, hypertension is the most important etiologic factor in the development of left ventricular hypertrophy and diastolic dysfunction. It has been proposed that delay of adequate therapy or failure to lower blood pressure to normal over several years results in changes that become irreversible or only slowly reversible with dialysis. Hypertension is the highest risk factor for coronary artery disease and cerebral vascular disease. The goal of therapy is a normal systolic and diastolic pressure.

Anemia

Studies now suggest that aggressive treatment of anemia in the period before dialysis therapy begins is as important as treatment during dialysis. In fact, to reduce cardiovascular morbidity and mortality, this therapy may be critical because long-standing left ventricular hypertrophy associated with anemia may be poorly reversible or irreversible if therapy is delayed until dialysis therapy begins. In addition, correction of anemia before dialysis therapy starts appears to improve or maintain functional capacity, nutritional adequacy, sexual function, and psychological health. It also reduces the risk for hepatitis and sensitization to transplant antigens associated with transfusion. As in the patient receiving dialysis, the patient not yet receiving it should be evaluated for other causes of anemia besides the renal failure, and any nutritional deficiencies should be corrected. As the anemia worsens, the physician should initiate therapy with subcutaneous erythropoietin. The target hematocrit level has not yet been determined. It is currently recommended that the hematocrit level be maintained above 30%, but studies are now being done to determine if higher hematocrit levels produce better results.

Renal Osteodystrophy

It is known that the factors mediating renal osteodystrophy are present early in the course of progressive renal disease. These factors need to be managed throughout the period before dialysis therapy begins to prevent the ravages of severe, potentially irreversible hyperparathyroidism. Patients should be instructed early about dietary phosphate restriction, probably before the serum phosphate level is elevated. Therapy with calcium-containing phosphate binders should be initiated when phosphate levels are minimally elevated. Metabolic acidosis should be rigorously treated to maintain bicarbonate levels near or at the normal range because the effect of acidosis can increase bone dissolution and inhibit osteoblastic activity, especially in children and women. Treatment of acidosis may also improve protein metabolism.

Nutritional Therapy

At an early meeting with the renal team, a nutritional assessment by a trained dietitian should be done and should include as a minimum the weight, height, recent weight loss, upper arm anthropometry, and serum protein levels (albumin, transferrin, or prealbumin, or all three). In the absence of obvious malnutrition, a modest protein-restricted diet of 0.7 to 0.8 g of protein/kg per day provides good nutrition. When malnutrition is present, adequate caloric intake and intake of greater amounts of dietary protein of up to 1 to 1.2 g/kg should be emphasized to allow nutritional repletion or to counter the catabolic effects of stress. Measurement of urinary urea nitrogen levels to assess the net protein catabolic rate can be useful for monitoring protein intake. In certain patients during the period before dialysis therapy begins, fluid retentive states will make nutritional assessment more difficult. Newer techniques such as multifrequency bioimpedance analysis and dual-emission radiograph absorptiometry offer promise of ease, reproducibility, and accuracy of assessing states of fluid overload and bone mineral status, respectively.

The dietitian should also design dietary prescriptions for energy, fat and carbohydrate, fluid, sodium, and phosphate levels, as well as other micronutrients, recognizing that the adequacy of energy intake will be largely monitored by weight change in outpatients. Although diet modification to minimize lipid abnormalities is reasonable, such modifications should not be so rigid that they limit energy intake below daily requirements. Lipid abnormalities, particularly hypertriglyceridemia, reduced high-density lipoprotein cholesterol levels, and elevations in lipoprotein(a) are common in end-stage renal disease, but limited data exist to support the efficacy of diet or drug therapy; some evidence suggests that the drugs usually used have more serious side effects.

Quality of Life

A patient's quality of life is important in the period before dialysis therapy begins and should be strongly considered in the decision to initiate the therapy. Maintenance of physical strength, appetite, sense of well-being, and optimal physiologic functioning promotes interpersonal relationships with family and friends and, in the working patient, rehabilitation and job retention. As the likely need for dialysis approaches, preparation of the patient by introduction to various aspects of the therapy, members of the renal team, the physical site of the therapy, and other patients receiving dialysis generally facilitates acceptance and compliance. Another potential benefit is the opportunity to involve the patient in the selection of therapy by discussing the characteristics of the various modes of the therapy and to allow early placement of vascular access if hemodialysis is the method chosen.

Dialysis Access

The benefits of early establishment of vascular access should be emphasized. Arteriovenous fistula surgery must occur weeks to months before the initiation of dialysis therapy to permit maturation of the fistula. Likewise, a peritoneal dialysis catheter should be placed at least 1 month before its anticipated use. Advantages to newer catheters may exist in which the external segment is initially buried subcutaneously and exteriorized when needed at a later date. Late referral is clearly associated with increased complications, the need for emergency hemodialysis, and possible long-term access problems.

Interventions in Renal Failure in Children

Chronic renal failure is different in children than in adults because of its low incidence (11 cases per 10⁶ children per year) and because in most cases it is caused by obstructive uropathy, renal dysplasia, and congenital or inherited diseases. Morbid conditions associated with childhood chronic renal failure are growth failure, osteodystrophy with bone deformity, salt and water losses due to urologic abnormalities, and neurologic abnormalities, including seizures, deafness, retardation, and learning disabilities. Because of growth requirements, dietary protein intake should be higher than that for adults, perhaps as high as 1.3 to 1.5 g/kg per day or even higher for children receiving peritoneal dialysis. The production of erythropoietin and calcitriol and the functions of the insulin-like growth factor-1 axis may be impaired from birth onward. Because of these features, treatment before dialysis therapy begins should be aimed at correcting malnutrition, hormone deficiencies, salt depletion, and neurologic dysfunction.

What Is the Relationship between Delivered Dialysis Dose and Morbidity/Mortality?

Hemodialysis

Indices of hemodialysis adequacy have historically included measurements of serum creatinine levels and urea, estimates of dialysis delivery (measured by square meter-hour), and assessment of patient well-being.

Recently, an estimate of fractional urea clearance during dialysis has been suggested as a more quantifiable measurement of dialysis efficacy. This estimate uses urea as a marker for uremic toxins cleared during the dialysis procedure. The fractional urea clearance model for hemodialysis is expressed as K_drt/V, where K_d is dialyzer clearance (mL/min), r is residual renal urea clearance (mL/min), t is treatment time (min), and V is total-body urea distribution volume in a single pool (mL). A simpler and more common measurement of fractional urea clearance during a single dialysis treatment is the urea reduction ratio. This ratio is expressed as a percentage and is calculated by the following formula: (blood urea nitrogen level before dialysis minus blood urea nitrogen level after dialysis/blood urea nitrogen) x 100. An approximate relation between these two means of expressing dialysis dose can be made: A K_dr t/V value of 1.2 is approximately equal to a urea reduction ratio of 60%. Although urea may be distributed in multiple body pools, most current measurements use a single-pool model to calculate urea clearance.

Recent reports have shown a direct correlation between dialysis mortality and K_dr t/V (or urea reduction ratio). Several studies have also suggested that the dialysis dose delivered to many patients receiving hemodialysis in the United States was less than that recommended by the National Cooperative Dialysis Study. Although data from controlled, prospective studies are unavailable, retrospective data presented and opinions expressed at the consensus conference favor a recommendation for a minimum delivered hemodialysis K_dr t/V dose of 1.2 (using a conventional dialyzer and single-urea pool analysis) in patients with a protein intake of approximately 1.0 to 1.2 g/kg per day. It is suggested that assessment of dialysis dose by formal K_dr t/V modeling be done on a regular basis. Opinions were expressed that dialysis time may be an independent predictor of mortality, regardless of the dialyzer urea clearance. It is obvious that a randomized controlled study relating the dose of delivered dialysis to morbidity and mortality is of great importance.

In the metabolically stable patient, the net protein catabolic rate reflects protein intake. As changes in K_dr t/V may be paralleled by corresponding changes in the net protein catabolic rate, dietary protein intake may decrease if the dialysis prescription fails to achieve the desired goal and the patient becomes symptomatic.

Morbidity

Attainment of the recommended K_dr t/V is influenced by several modifiable and unmodifiable factors that may alter the delivered dose, including the following: 1) Vascular access: Obstruction to blood flow in the vascular access may occur and result in recirculation of blood through the dialysis circuit, thereby contributing to decreased dialysis dose; 2) equipment: Blood flow rate, dialyzer surface area, and the mass-transfer coefficient must be considered to give optimal delivery to achieve the calculated dialysis dose. Effective dialyzer surface area must be carefully monitored because excessive reuse of dialysis membranes results in loss of dialyzer efficiency and reduction of the delivered dialysis dose; and 3) patient factors: Adherence to salt and water intake limitations must be met to avoid unnecessary fluctuations in blood volume during hemodialysis and the associated loss of effective dialysis. Other patient compliance issues include adherence to appointment schedules and time of dialysis. Patients with certain underlying diseases (for example, diabetes, amyloidosis, and drug dependence) have special problems that may interfere with dialysis.

Dialysis Biocompatibility

The composition of the hemodialyzer membrane may be a factor in establishing urea clearance goals because biocompatible polymer membranes such as polysulfones, polyacrylonitrile, and polymethylmethacrylate have permeability characteristics different from cellulosic membranes. In addition, the composition of the membrane may be a factor in the nature and intensity of the interaction between the membrane and blood. Generally, cellulosic-based membranes, in contrast to the more biocompatible membranes, have a greater capacity to activate complement and to attenuate the granulocyte response. It has also been suggested that the use of biocompatible membranes may result in lower mortality rates.

Peritoneal Dialysis

Peritoneal dialysis uses a natural membrane to remove nitrogenous products from the body fluids of persons with impaired renal function. The use of relatively long dwell-time peritoneal exchanges (continuous ambulatory peritoneal dialysis) has enabled persons to carry on normal daily activities without using machines or other appliances. The dose of peritoneal dialysis has been established empirically and depends to some extent on patient acceptance of frequent interruptions for the exchange of peritoneal fluid. Recently, an effort has been made to prescribe for each individual patient the dose of peritoneal dialysis needed to attain target levels of urea clearance. In general, four exchanges of 2 L each may generate as much as 10 L of dialysate (allowing for the removal of ultrafiltrate). Assuming nearly complete equilibration of urea between plasma and peritoneal fluid, this equates to a weekly urea clearance of approximately 70 L. For a 70-kg man with a urea "space" of 42 L, the calculated K_pr t/V is 1.7. Current evidence indicates that this K_pr t/V value is a reasonable minimal delivered dose for most functionally anephric patients receiving continuous ambulatory peritoneal dialysis who daily eat approximately 0.9 to 1.0 g/kg of protein. The dose of nighttime peritoneal dialysis is usually increased above that of continuous ambulatory peritoneal dialysis.

The prescription of dialysis depends on the volume of urea distribution, the efficiency of peritoneal exchange, and the residual renal urea clearance. Peritoneal dialysis is a demanding and time-consuming therapy. Omission of exchanges or shortening exchange times by the patient reduces urea clearance and leads to increased morbidity and mortality. The use of urea as an index of peritoneal dialysis efficiency is complicated because the peritoneal membrane is more permeable to large molecules than are dialyzer membranes.

Peritoneal dialysis efficiency can be increased by more frequent exchange (5 per day), increased volume per exchange (2.5 to 3.0 L), and the coupling of continuous ambulatory peritoneal dialysis with nighttime cycler dialysis in large persons or in those with relatively low peritoneal clearances.

Children

Children receiving long-term dialysis therapy are more likely to receive peritoneal dialysis than adults. This preference is based on technical factors, including problems maintaining long-term hemodialysis access. Because of the serious problems of growth failure and neurologic dysfunction, children require appropriate hormone therapy (erythropoietin, calcitriol, and growth hormone), nutrition support services, and neurologic evaluation. A qualified pediatric nephrologist is an essential member of the renal team. Data exist that intervention with specified nutrition, growth hormone, erythropoietin, and calcitriol therapy, and avoidance of aluminum can clearly improve growth velocity. Because of the serious problems of growth failure and neurologic dysfunction, children with renal insufficiency should be referred to centers with specialized pediatric nephrologic care. Children also require educational and play facilities at the dialysis center.

Children of all ages with end-stage renal disease benefit from treatment with peritoneal dialysis and hemodialysis. The principles of dialysis outlined for adults generally also apply to children, although no retrospective or prospective studies have been done that indicate reasonable targets of K_pr t/V or K_dr t/V values to maximally allay morbidity and mortality.

Children with chronic renal failure experience a cycle of depression, anxiety, and loss of self-esteem. The difficulties encountered often result in family stress with a high divorce rate among the parents of a child receiving dialysis. For these reasons, a mental health professional is an essential component of the pediatric renal disease center.

Finally, dialysis should be a temporary therapy because renal transplantation is considered the treatment of choice for children.

Can Comorbid Conditions Be Altered by Nondialytic Interventions To Improve Morbidity/Mortality in Dialysis Patients?

Cardiovascular Abnormalities

Cardiovascular events (primarily systolic and diastolic dysfunction, myocardial infarction, and stroke) account for 50% of the mortality in patients receiving dialysis and also substantially contribute to mortality after renal transplantation.

Studies of patients entering dialysis treatment show a high prevalence of established cardiovascular abnormalities, including hypertension, left ventricular hypertrophy, coronary artery disease, and cardiac failure. For example, two-dimensional echocardiograms are abnormal in 70% of patients with such conditions. The increasing mean age of patients receiving dialysis will probably further increase these cardiovascular pathologic conditions.

We believe that optimum reduction of dialysis morbidity and mortality should start with intervention before dialysis therapy begins. The patient with chronic renal failure is at high risk for cardiovascular events. It is likely but not yet proven that prevention of severe anemia by erythropoietin also prevents, diminishes, or partially reverses left ventricular overload.

Smoking cessation, correction of obesity, and regular aerobic exercise may also contribute to reducing mortality from cardiovascular disease. Normotension and nonsmoking have been two characteristics of patients who have survived 20 years or more while receiving long-term dialysis.

It is not yet known whether modifications of the common lipid abnormalities in chronic renal failure and patients with end-stage renal disease can be safely achieved in the long term by currently available lipid-lowering agents or whether this would be beneficial.

Because myocardial calcification and fibrosis may especially contribute to diastolic dysfunction (which accounts for 50% of cardiac failure in patients receiving dialysis), control of calcium, phosphorus, and parathyroid hormone levels may help to prevent cardiovascular disease as well as bone disease.

Two thirds of cases of end-stage renal disease are caused by two primary diseases, diabetes mellitus and essential hypertension, that themselves contribute substantially to cardiovascular disease. Not infrequently, patients with these diseases have had erratic treatment and follow-up programs before the onset of chronic renal disease. The identification of a diabetic patient has not routinely led to inclusion of that patient in a program of strict glycemic control and follow-up of potential microvascular and renal complications, such as microalbuminuria or gross albuminuria. We also now understand that careful control of blood pressure upon diagnosis of diabetes mellitus is crucial.

Current studies suggest that blood pressure is not being adequately controlled in many patients receiving dialysis. Blood pressure at the initiation of each dialysis treatment should be in the normal range or as near as possible to it. Adequate ultrafiltration and restriction of interdialytic intake of sodium chloride should establish normotension in up to 80% of patients receiving dialysis. Mechanisms of hypertension in the remaining patients include an inappropriately hyperactive renin-angiotensin system, nephrogenic activation of the sympathetic nervous system, and, possibly, an altered balance of endothelial factors (such as nitric oxide and endothelin) influencing arteriolar smooth muscle tone.

Nutritional Deficiency

The nutritional status of the patient is a major factor in the outcome of hemodialysis treatment and may be maintained in the period before dialysis is begun by the use of low-protein diets (0.7 to 0.8 g/kg per day with an adequate calorie intake of 35 kcal/kg per day). It is essential that during this period, malnutrition, as evidenced by a decrease in albumin level and body weight, is not allowed to develop in patients with renal disease. Serum albumin levels greater than 3.5 g/dL are associated with little mortality, whereas mortality increases dramatically with lower serum albumin values.

Once the patient is receiving hemodialysis, dietary protein levels should be liberalized to equal 1.0 g/kg per day, with appropriate calorie supplementation, to sustain nutrition at a normal level. The complexity of nutritional intervention for the patient with renal disease is of such degree and importance to require the expert guidance of a well-trained renal dietitian. High cholesterol levels indicate an increased risk for morbidity and mortality, but values lower than 100 mg/dL are also associated with increased mortality. The reasons hypoalbuminemia and hypocholesterolemia are indices of high mortality are unknown.

Educational programs instituted by the renal center and by organizations concerned with the welfare of all patients with renal disease should explain the need for adequate dialysis time and correction of malnutrition because these factors contribute to a longer life of higher quality and correction of many comorbid conditions. Patient participation with the other members of the renal team is essential if quality of life is to be improved.

Current concerns about morbidity and mortality raise issues regarding the present uniform reimbursement system for dialysis, especially in the area of nutritional and psychosocial support systems. Linking direct reimbursement for such care to important outcomes such as serum albumin levels, mean blood pressure, and measurements of fractional urea clearance during dialysis should be explored.

How Can Dialysis-related Complications Be Reduced?

Although dialysis allows effective and productive lives for many patients with end-stage renal disease, various complications can develop. Problems with dialysis access, infections, atherosclerosis and cardiovascular disease, malnutrition, and metabolic abnormalities, as well as persisting uremic symptoms and acute symptoms related to the dialysis procedure itself, may limit a patient's health and quality of life. Disorders of calcium, phosphorus, vitamin D, and parathyroid hormone are common and may be disabling.

Hemodialysis

Perhaps the major complication limiting continued effective hemodialysis involves vascular access. The most effective, durable access is the arteriovenous fistula. Unfortunately, a satisfactory fistula cannot be established in many patients because of inadequate vessels (especially in diabetic patients). The chances of a successful fistula are enhanced by early planning and placement well before dialysis becomes necessary. When early planning is not possible, the use of a tunnelled subcutaneous catheter may make dialysis possible while an arteriovenous fistula is maturing, but repeated use of temporary subclavian catheters is often accompanied by infection or thrombosis with ultimate impairment of subclavian flow and loss of the whole arm for dialysis access purposes. Use of temporary catheters should be avoided when possible.

When a fistula is unsuccessful or not feasible, a synthetic graft is usually placed. Current experience indicates that 60% of these grafts fail each year because of thrombosis. Anatomic stenosis is responsible for four fifths of these clots (almost all are on the venous side of the anastomosis), whereas the rest result from causes such as excessive post-venipuncture pressure by manual compression or clamp or by sleeping on the graft. Medical thrombolysis may remove the clot and restore flow, but often surgical thrombectomy is required. The stenosis, usually formed by endothelial proliferation, sometimes responds to percutaneous angioplasty but may require surgical intervention. The present life of a synthetic graft is about 2 years, with loss caused by thrombosis in 80% of patients and infection in 20% of patients.

Consistently elevated venous dialysis pressure may provide a warning of developing stenosis and hence of impending thrombosis and may indicate the need for a fistulogram. An increase in recirculation may also indicate an incipient problem. Attention to these signs may allow intervention before clotting of the graft and may prevent its loss.

The need for meticulous, experienced surgical skill in establishing satisfactory fistulas and shunts must be emphasized. Although the procedure may not be dramatic, the life of a patient receiving dialysis often depends on the presence of a reliable access. Nursing skill in access use has a major influence on dialysis success.

Infection

Infections remain the major cause of death in 15% to 30% of all patients receiving dialysis, a Figure that has not changed significantly over the years. Infections are usually caused by common organisms and often appear to be related to access. Approximately 60% of bacteremic infections are gram-positive, especially Staphylococcus aureus. Perhaps 50% to 60% of patients receiving dialysis are carriers of this organism (compared with 10% to 30% of the general population), and the carrier rate among diabetic patients is still higher. It is possible to reduce the carrier rate with prophylactic antibiotic treatment, but this may encourage the emergence of resistant organisms.

Uremia itself causes an impairment in cell-mediated immunity that is not totally corrected by dialysis. In addition, granulocyte phagocytosis and killing functions appear to be impaired by cellulosic dialysis membranes. Biocompatible membranes may have less deleterious effects on leukocyte function and other defense mechanisms. Some studies suggest a 50% decrease in incidence of infection accompanying a switch to more biocompatible dialyzers.

Peritoneal Dialysis

The most frequent cause of unsuccessful peritoneal dialysis is peritonitis. Although improvement has followed recent changes in tubing and connection systems, recurrent peritonitis is a continuing problem for many patients. Catheter tunnel infection often underlies this peritonitis, and changes in catheter design (for example, a U shape), placement (with both peritoneal and skin ends directed caudad and a cuff placed in the rectus muscle), and the use of prophylactic antibiotics at the time of placement or thereafter have been proposed as deterrents to infection. The use of a vaccine against Staphylococcus organisms and of bacteriostatics such as silver-coated catheters are under investigation.

Calcium, Phosphorus, and Parathyroid Hormone

The disturbances in body calcium, phosphorus, vitamin D, and parathyroid hormone and bone disease that usually start before the initiation of dialysis therapy continue to demand consistent attention as long as dialysis is required. Mainstays in therapy include control of dietary phosphorus, minimization of its absorption by using phosphate-binders, and the use of calcitriol. Control of dietary intake of phosphorus requires patient education by the renal team and the patient's adherence to the recommended diet. Most physicians no longer rely on aluminum hydroxide to prevent absorption of phosphorus because it leads to accumulation of aluminum in the brain and bone, which leads to severe neurologic disorders and osteomalacia. To prevent absorption of phosphorous, ingestion of calcium carbonate or calcium acetate with meals is currently recommended for most patients. Use of these calcium salts may require adjustments in the concentration of calcium in the dialysate fluid to prevent hypercalcemia and consequent formation of calcium phosphate salt deposits with damage to the heart, blood vessels, and other tissues. Careful titration of the calcitriol dosage is required to obtain its benefits without causing hyperphosphatemia or hypercalcemia. Careful attention to dietary phosphorus, calcium salts, and calcitriol often enable parathyroid hormone concentrations to be maintained at or near normal. Of serious concern is the emergence of "adynamic bone disease," a condition diagnosed by bone biopsy in which the normal correction of bone wear and tear by "remodeling" fails to occur. The exact cause or causes and results of adynamic bone disease are not yet known.

Amyloidosis

In patients receiving dialysis, amyloidosis is associated with long-term dialysis (lasting longer than 6 years) and is increased in frequency in older patients. The formation of ß-2-µglobulin protein deposits as amyloid causes the carpal tunnel syndrome, destructive arthropathy in medium- and large-sized joints, and cystic bone disease. The disorder may be caused both by increased release of ß-2-µglobulin from macrophages and, significantly, by reduction in the destruction of ß-2-µglobulin that normally occurs in functioning kidneys. Some evidence indicates that amyloidosis is a lesser problem in patients receiving dialysis with high-flux membranes than in those receiving dialysis with cellulosic membranes, perhaps because of both decreased release of the protein from macrophages and from partial removal of the protein during dialysis by filtration or binding with some synthetic polymer membranes. Consideration should be given to the use of these membranes for dialysis of patients in whom amyloidosis is a problem or may become a clinical concern.

Anemia

Attention to the management of anemia, begun in the phase of care before dialysis therapy begins, must be continued throughout dialysis therapy.

Intradialytic Complications

Acute complications related to the dialysis procedure itself may severely compromise the quality of life in patients receiving long-term dialysis. A mild degree of hypotension is "normal" while receiving dialysis, but severe degrees may be disabling. Muscle cramps, chest or back pain, hypoxemia, fever, nausea, seizures, or cardiac arrhythmias may occur. In addition, mechanical problems related to dialysis machines, cartridges, and water purifiers may occur.

Some of these problems have been lessened by the use of bicarbonate rather than acetate dialysis solutions, by longer dialysis periods with lower rates of ultrafiltration, with the use of synthetic polymer dialysis membranes that are biocompatible, and perhaps by the reuse of these membranes. Reuse brings the potential for problems as well as benefits; however, additional research will be necessary to define the optimum mix of membranes, reuse, solutions, and time and intensity of dialysis to ensure maximum safety and minimum complications of dialysis.

Psychosocial Concerns

Early assessment before dialysis therapy begins and continuous, active intervention by the renal team, including mental health professionals, in the care of a patient beginning dialysis therapy are more likely to be effective than efforts initiated later in treatment. This assessment should include measures of quality of life and social role function in addition to measures of lack of mental acuity and depression. Ensuring patients' understanding and positive participation in their care is a primary goal of this intervention, in addition to optimizing the relationships between patient and physician and patient and staff. The earlier this assessment is done, the greater will be the potential for a positive effect on physical and social rehabilitation. Exercise and physical training can add to physical well-being and should also be initiated at the start of dialysis therapy or before it begins.

What Are the Future Directions for Research in Dialysis?

1. Studies should be done to evaluate the effect of aggressive nutritional support in malnourished patients who have not yet received dialysis, to determine the mechanisms by which malnutrition increases mortality and morbidity rates, and to develop sensitive and specific methods to detect the early stages of malnutrition.

2. Studies should be done to determine the benefits and risks of early control of renal osteodystrophy on morbidity and to explore the causes and treatment of disturbances in calcium, phosphorus, and vitamin D, both at the basic level on regulation of bone metabolism and at the clinical level on the importance of the formation of soft-tissue calcium deposits. Studies should include development of new phosphate-binding agents and noncalcemic analogs of vitamin D and determination of the optimal degree of suppression of parathyroid hormone.

3. Basic and clinical studies should be done to evaluate the effect of chronic uremia on neurologic function.

4. Basic and clinical studies should be done to evaluate the effect of uremia on growth in children.

5. Studies should be done to determine the effect of early treatment of anemia on mortality, morbidity, and rehabilitation. Studies to determine when to initiate treatment of anemia and what should be the target hematocrit level are needed both in patients not yet receiving dialysis and in those currently receiving it.

6. A randomized controlled clinical trial should be initiated to examine the differences in patient morbidity and mortality at K_dr t/V levels of 1.2 (single pool) and 1.6 for patients receiving hemodialysis.

7. A randomized controlled clinical trial should be initiated to examine the differences in morbidity and mortality at delivered weekly K_pr t/V levels of 1.47 and 2.10 in patients receiving peritoneal dialysis.

8. A randomized controlled clinical trial at a specified level of delivered dialysis dose should be initiated to determine the differences in the effects of biocompatible, high-flux membranes compared with cellulosic membranes in studies that include, but are not limited to, patient survival, incidence of infection, and incidence and course of ß-2-µglobulin amyloidosis.

9. Additional studies to establish the effect of reuse of dialysis membranes on hemodialysis effectiveness and morbidity and mortality are recommended.

10. A prospective study of the feasibility and effectiveness of modification of cardiovascular risk factors in patients with chronic renal failure both before and after initiation of dialysis therapy should be done. Risk factors to be evaluated would include hypertension (mechanism of development and regression of left ventricular hypertrophy and characterization of the best pharmacologic approaches to antihypertensive treatment), smoking, obesity, and uremic dyslipidemia. The role of metabolic factors such as hyperinsulinemia and parathyroid hormone and calcium phosphorous relations including tissue calcium burden in the myocardium and methods of its detection should be examined. Finally, development of noninvasive testing for coronary artery disease in these patients should be explored.

11. Studies to determine the mechanisms of interdialytic hypertension should be done and should include the respective roles of abnormal renin-angiotensin responses, abnormal thirst and salt craving, vascular endothelial factors (such as endothelin, nitric oxide production, and inhibitors), the renal-sympathetic axis, the relation to erythropoietin administration, and the role of continuous blood pressure monitoring.

12. Studies of the mechanisms by which malnutrition increases mortality and morbidity rates due to infections, anorexia, hypogeusia, and related problems in the patient receiving dialysis should be done.

13. Improved methods for detecting stenosis and thrombosis of access grafts and understanding the mechanism of endothelial proliferation leading to vascular graft stenosis are needed. Improved material and techniques should be developed to diminish access clotting and infection and new methods should be identified for cost-effective thrombolysis in clotted grafts.

14. Studies of the immunodeficiency of uremia and evaluation of antibacterial vaccines, antibiotic prophylaxis, and dialyzer membrane characteristics in the prevention of infection in patients receiving dialysis should be done.

15. Evaluating and standardizing methods for measurement of psychological well-being and quality of life in patients receiving dialysis and applying these instruments in studies on the effectiveness of interventions should be done.

Conclusions

1. Before dialysis therapy begins, patients, including children, should be referred to a renal team consisting of a nephrologist, dietitian, nurse, social worker, and mental health professional to reduce the morbidity and mortality during this period and during subsequent dialysis therapy.

2. The social and psychological welfare and the quality of life of the patient receiving dialysis are favorably influenced by early and continued involvement of a multidisciplinary renal team.

3. To avoid a catastrophic onset of dialysis, attempts should be made to institute intervention before dialysis therapy begins and to appropriately initiate dialysis access.

4. Quantitative methods to measure the delivered dose of hemodialysis and peritoneal dialysis have now been developed. These methods permit an objective evaluation of the relation between the delivered dose of dialysis and morbidity and mortality. These methods suggest that the dose of hemodialysis and peritoneal dialysis has been suboptimal for many patients in the United States.

5. Factors contributing to underdialysis of some patients include problems with vascular and peritoneal access, nonadherence to the dialysis prescription, and underprescription of the dialysis dose.

6. Until randomized controlled trials have been completed, a delivered hemodialysis dose at least equal to a measured K_dr t/V value of 1.2 (single pool) and a delivered peritoneal dialysis dose at least equal to a measured K_pr t/V value of 1.7 (weekly) are recommended.

7. Cardiovascular mortality accounts for approximately 50% of deaths in patients receiving dialysis. Relevant risk factors should be treated as soon as possible after diagnosis of chronic renal failure. These factors include hypertension, smoking, and chronic anemia.

8. Patients with diabetes mellitus face especially severe cardiovascular risk, which contributes to reduced survival while receiving dialysis.

9. Malnutrition is another important comorbid condition contributing to mortality. A serum albumin level lower than 3.5 g/dL is clearly associated with increased relative risk. Early detection and treatment of malnutrition should substantially improve survival.

10. Control of renal osteodystrophy requires patient adherence to the prescribed regimen and careful attention by the renal team to calcium and phosphorus intake and to the use of phosphate binders and calcitriol.

11. Early creation of an arteriovenous fistula is preferable to placement of a synthetic graft for vascular access. Both require an experienced, meticulous surgeon.

12. Skilled management by nursing and other clinical personnel will help prolong the life of the vascular access.

13. Attention to catheter design, placement, and care and to exchange procedures can minimize infection in patients receiving peritoneal dialysis.

14. Biocompatible dialysis membranes may reduce infection and formation of amyloid deposits in patients receiving hemodialysis, but evidence is currently inconclusive.

15. Financial support for clinical investigation, including outcomes and health services delivery research, should be incorporated into the budgets of the Medicare End-Stage Renal Disease Program, Health Care Financing Administration, Agency for Health Care Policy Research, and the Food and Drug Administration. This support will enable studies to be done that promise to improve morbidity and mortality, enhance cost-effective care, and create long-term financial savings in the Medicare program.

Appendix

Consensus Development Panel Members

C. Craig Tisher, MD, Chair, Division of Nephrology, Hypertension, and Transplantation, University of Florida College of Medicine, Gainesville, Florida.

Christine P. Bastl, MD, Division of Nephrology, Temple University Health Sciences Center, Philadelphia, Pennsylvania.

Bruce R. Bistrian, MD, New England Deaconess Hospital, Harvard Medical School, Boston, Massachusetts.

Russell Chesney, MD, Department of Pediatrics, University of Tennessee College of Medicine, Memphis, Tennessee.

Cecil Coggins, MD, Renal Unit, Massachusetts General Hospital, Boston, Massachusetts.

Marie Diener-West, PhD, Department of Biostatistics, Johns Hopkins University School of Hygiene and Public Health, Baltimore, Maryland.

Darrell D. Fanestil, MD, Department of Medicine, University of California, San Diego, La Jolla, California.

Jared Grantham, MD, Division of Nephrology and Hypertension, Kansas University Medical Center, Kansas City, Kansas.

Robert Kunau, MD, Dallas Nephrology Associates, Dallas, Texas.

Robert G. Luke, MD, Department of Medicine, University of Cincinnati Medical Center, Cincinnati, Ohio.

Sandra L. Madison, RN, Department of Nursing, Medical/Surgical Division, George Washington University Medical Center, Washington, DC.

Manuel Martinez-Maldonado, MD, Department of Medicine, Emory University School of Medicine, Decatur, Georgia.

Richard Salick, National Kidney Foundation, Orlando, Florida.

Planning Committee Members

M. James Scherbenske, PhD, Chair, Renal Physiology/Cell Biology Program, Division of Kidney, Urologic, and Hematologic Diseases, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, Maryland.

Lawrence Y. Agodoa, MD, End-Stage Renal Disease Program, Division of Kidney, Urologic, and Hematologic Diseases, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, Maryland.

Juan P. Bosch, MD, Division of Renal Diseases and Hypertension, George Washington University Medical Center, Washington, DC.

Elsa Bray, BGS, Office of Medical Applications of Research, National Institutes of Health, Bethesda, Maryland.

Benjamin T. Burton, PhD, Disease Prevention and Technology Transfer, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, Maryland.

Jose A. Diaz-Buxo, MD, Metrolina Nephrology Associates, PA, Charlotte, North Carolina.

John H. Ferguson, MD, Office of Medical Applications of Research, National Institutes of Health, Bethesda, Maryland.

Willis R. Foster, MD, Office of Disease Prevention and Technology Transfer, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, Maryland.

Raymond M. Hakim, MD, PhD, Division of Nephrology, Vanderbilt University Medical Center, Nashville, Tennessee.

William Hall, BS, Office of Medical Applications of Research, National Institutes of Health, Bethesda, Maryland.

William Harmon, MD, Pediatric Nephrology, Children's Hospital, Boston, Massachusetts.

Mary M. Harris, Office of Health Research Reports, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, Maryland.

Gladys H. Hirschman, MD, Chronic Renal Diseases and Pediatric Nephrology Programs, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, Maryland.

Nathan Levin, MD, Division of Nephrology, Beth Israel Medical Center, New York, New York.

Thomas F. Parker III, MD, Dallas Nephrology Associates, Dallas, Texas.

Gary E. Striker, MD, Division of Kidney, Urologic and Hematologic Diseases, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, Maryland.

C. Craig Tisher, MD, Division of Nephrology, Hypertension, and Transplantation, University of Florida College of Medicine, Gainesville, Florida.