The Issue of Timing

The natural history of progressive renal insufficiency is now clearly defined. Signs and symptoms of developing uremia can be approximately correlated with the level of residual renal function. Mitch et al. plotted the slope of serial measurements of serum creatinine concentration against time and demonstrated that the need for treatment of end stage renal disease (ESRD) can be anticipated.(1) In a cohort of 34 patients with chronic renal insufficiency caused by several diseases, the course in 31 patients was that of a linear decline in the reciprocal serum-creatinine concentration as creatinine concentration rose from a mean of 2-6 mg/dl to 14-8 mg/dl over a mean of 71 months. In other words, with worsening renal disease, the value of 1/Cr decreases, ultimately intersecting the X axis (time). It is this projected crossing of the 1/Cr line with the time axis that marks the date when GFR will be 0 ml/min.

If we extrapolate from this observation to the large population of more than 15 million Americans with chronic kidney disease, we can reasonably project the national need for renal services.(2) Derivative studies have shown that the slope of a "Mitch Plot" varies by patient, disease and therapy. The start date of ESRD therapy in a specific patient is, therefore, less a science than a clinical "guess."

But kidney patients, like those afflicted with any chronic disease, desire to know what to expect from their affliction and their physician. An individualized long-range plan that incorporates major life style alterations forced by ESRD may minimize medical and socioeconomic risks while optimizing rehabilitation. In practice, either a Mitch Plot of reciprocal serum creatinine values or instantaneous computation of residual renal function is determined. Alternatively, glomerular filtration rate (GFR) can be estimated from a single measurement of serum creatinine to "approximate" the creatinine clearance (CCr) using the Cockcroft Gault formula.(3)

The Cockcroft Gault formula provides an "estimated" creatinine clearance expressed in ml/min based on the serum creatinine level in mg/dl. It has been validated for men aged 18-110, weighing 35-120 kg, with a serum creatinine concentration of 0.6-7 mg/dl). To apply the Cockcroft Gault formula to women, multiply the above result by 0.85. First step: Determine ideal body weight (IBW in kg) by calculating (2.3 kg x [height (in) - 60] + (45.5 kg if female, 50.0 kg if male). Second step: Multiply [140 - age (yrs)] x [lesser of IBW (kg) or actual weight (kg)] / [72 x serum creatinine (mg/dl)]. Actual measurement of CCr necessitates the "inconvenience" of, and thus suboptimal compliance with, collecting a timed urine collection (overnight or 24 hours) plus a blood sample, a technique that is too complicated for large-scale studies.

In a previous Cyberounds^®, we advocated the positive value of reducing the amount of proteinuria as a means of renoprotection. Normalizing a hypertensive blood pressure is another empirically validated component of nephroprotection. We now explore whether or not patient care benefits from early referral by the primary care physician to a nephrologist. The focused question is whether referral to a nephrologist, prior to the actual onset of ESRD, should, like repeated sampling of serum creatinine concentration, be advocated as a standard of care.

Can we document any benefit from "early" collaboration with (or total management by a nephrologist? Listed in Table 1 are variables that might be suitable to compare outcome in cohorts of ESRD patients assigned to early versus late referral to a nephrologist.

An underlying premise of preventive medicine is that recognition of risk leads to implementation of a clinical strategy to minimize and/or delay the consequences of that risk. Common sense suggests that early intervention in any disease process improves outcome. Can we trust our common sense in medicine or must we restrict our actions to those based on objective trials? For decades, it was thought that early referral to a nephrologist improves subsequent morbidity and mortality in patients with ESRD. Is this a myth or a substantiated reality?

The Case for Early Referral of Patients with Chronic Kidney Disease to a Nephrologist

Dozens of uncontrolled, mainly anecdotal, studies recounted a policy of early referral to a nephrologist associated with early initiation of dialysis. Van Biesen, Wiedemann and Lameire, for example, in 1998, summarized "two important non-medical factors that affect the outcome of ESRD patients in Europe: late referral and the structure of the health care system."(4) Relevant to the issue under examination, 30% of patients were referred to a nephrologist less than one month before starting ESRD therapy, leading to a one-year death rate of 28.6% versus a 15.5% death rate for individuals with "early" referral (P=0.07). Maintenance hemodialysis (rather than peritoneal dialysis) was utilized in 77.5% of late referrals versus only 51.1% of those with early referral. Conclusion: early nephrologic intervention saved lives. Further to that point, in 1984, Ratcliffe, Phillips and Oliver reviewed outcomes in 55 patients begun on maintenance dialysis at Oxford in 1981, and noted that 23 patients were referred late, "very shortly before the need for dialysis." Late referral correlated with higher morbidity at the start of dialysis, major complications (70%) and a higher mortality. The authors concluded that early referral "benefits the patient and allows Health Service resources to be used more economically."(5)

Two decades later, studies published between 2001 and 2003 affirmed that early participation by a nephrologist reduced mortality and morbidity during subsequent ESRD treatment. Table 2 lists nine studies in which investigators concurred in concluding that delayed referral to a nephrologist exacted a toll in terms of mortality and morbidity once ESRD therapy began in U.S., European and Australian hemodialysis programs.

Jungers et al.,(6) or example, reviewed outcomes in 1,057 consecutive ESRD patients who started dialysis treatment at the Necker Hospital (Paris) from 1989 to 1998. Survival at five years was significantly lower (59%) in those who had a nephrologist's participation starting <6 months before initiation of hemodialysis compared with 77% in those patients who were treated by a nephrologist starting 6-35 months before onset of ESRD (P<0.001). Similarly, Stoves, Bartlett, and Newstead reviewed all 1,260 ESRD patients who started dialysis between 1980 and 1999 comparing Kaplan-Meier survival curves for those who had nephrologic care predialysis for<90 days versus those followed for >90 days.(7) [Kaplan-Meier calculations plot proportion surviving on the vertical (Y) axis versus time on the horizontal (X) axis, making it easy to compare cohorts with declining survival visually.]

At four months, one year and five years, late referral restricted survival proportions to 87%, 74% and 31% versus 94%, 87% and 55% in those with more extensive involvement by a nephrologist (P<0.001). The authors interpreted their finding as an "important message relevant to all potential referring physicians" that late referral of ESRD patients promotes poor survival on dialysis. The remaining cited investigators assessed different nuances in outcome after early or late referral to a nephrologist, but all reached the conclusion that early referral confers substantive benefit to the ESRD patient.(8),(9),(10),(11),(12)

Referral Timing and Renal Transplant

Does early referral increase the patient's chance of receiving a kidney transplant as has been claimed? Cass et al. quantified success competition for limited donor kidneys and found that late referral decreased the probability of ultimately receiving a renal transplant (adjusted rate ratio 0.78).8 Roderick et al. recognized that late referral to a nephrologist might have been caused by avoidable (misdiagnosis) and unavoidable (precipitous onset of renal failure, late presentation of the patient) factors.(9)

Kinchen et al.'s broad analysis of 828 patients with new-onset ESRD, in 81 dialysis facilities throughout the U.S., found late referral more common among African American men than white men (44.8% vs. 24.5%; P<0.05), uninsured patients than insured patients (56.7% vs. 29.0%; P<0.05) and patients with severe comorbid disease, such as diabetic macrovasculopathy, than those with mild comorbid disease (35.0% vs. 23.0%; P<0.05).(10) Compared with patients who had early evaluation, the risk for death was 1.8 times greater for those with late evaluation.

Both Avron et al.(11) and Stack(12) attempted to stratify ESRD mortality according to the number of pre-ESRD visits to a nephrologist. They found higher mortality in patients who had less than five nephrologist visits (15% in the first year of dialysis) and lower mortality risk in patients who saw a nephrologist at least twice in a year preceding dialysis (RR=0.8).

A Negative View of Early Referral for ESRD Therapy

The contrary view, that late referral is not injurious to either morbidity or survival in ESRD, has also been expounded after review of clinical outcomes in Europe. Roubicek et al.(13) retrospectively studied 309 ESRD patients in Marseille, France, who began dialysis between 1989 and 1996, of whom 177 patients (58%) had an early referral 16+ weeks before the start of dialysis and 93 patients (31%) had late referral <16 weeks before dialysis. Survival analysis showed no difference between the two groups and was excellent (96% in both groups at three months and 89-90% at one year). Five-year survival was 52% in early referral versus 56% in late referral. However, an accompanying editorial by Friedman criticized the study's methodology and selection bias as practiced in French dialysis programs, and noted a lack of definable criteria which might distinguish early from late referral.(14)

More recently, other studies, particularly the prospective multicenter study in the Netherlands by Korevaar et al., have failed to discover the claimed benefits of early referral to a nephrologist.(15) Korevaar et al. commented that implementation of the U.S. National Kidney Foundation Dialysis Outcomes Quality Initiative (DOQI) Guidelines would not substantially benefit the patients' daily lives while increasing costs. The authors compared 94 (37%) of 253 patients who started dialysis reatment later than recommended by the DOQI guidelines (between January 1997 and May 1999) with those who started dialysis on time. While late referral patients did have increased mortality, the authors attributed the survival advantage to their better initial renal function when commencing ESRD care. The authors "question the benefit of putting this guideline into daily practice, given the current clinical evidence and the effects it would have on patients and dialysis resources."

Finally, we have the study by Traynor et al. who argue that DOQI guidelines are largely "opinion-based, because evidence is still lacking."(16) Electronic records at the Glasgow Royal Infirmary had sufficient data to calculate the time at which 275 patients who were started on dialysis for chronic renal failure between January 1997 and May 1999 deteriorated to reach an estimated creatinine clearance (CCr) of 20 ml/min. Patients were divided at initiation of dialytic therapy, by a median CCr of 8.3 ml/min, into early and late start groups. Survival, measured from this date forward, indicated no significant benefit in patient survival in the early start cohort compared with the late start cohort. Traynor et al. conclude that their study did not show any survival advantage from earlier initiation of dialysis for ESRD, "a practice that has enormous personal, social, and economic implications." As did Korevaar et al., Traynor et al. advise that "Until evidence becomes available from a prospective randomized trial that eliminates the effect of lead-time bias, early initiation of dialysis cannot be supported." In fact, when a Cox Proportional Hazards statistical model was applied to outcome data, there was a significant inverse relationship between CCr at the start of dialysis and survival (hazard ratio, 1.1; P=0.02); patients who started dialysis with a lower CCr actually survived longer! No significant benefit to an early start of dialysis was evident 10 years after the date of the first dialysis treatment.

Lessons Learned and Conclusions

Like so many issues in nephrology, the jury must remain out if judged by Journal Club Criteria of evidence based medicine. Insufficient data is the key limitation which prevents more certain conclusions. Every aspect of the case for early referral, though consonant with modern professional organizational thinking (National Kidney Foundation, American and International Societies of Nephrology), deserves the British jurisprudence verdict of "unproven." Because the criteria for assessing benefit of early referral to a nephrologist are not standardized (Table 3), "unproven" is the most appropriate assessment.

Given the pragmatic reality that patients suffering progressive renal failure today cannot wait for tomorrow's answers before receiving treatment, what should the nephrologist of 2003 do with the evidence available?

While we support and wait for the results from more prospective, carefully controlled studies of every controversial aspect of nephrology practice, we nevertheless urge early and repeated referral to a nephrologist for every patient with deteriorating kidney function. Based on our experience with kidney patients in the three primary hospitals served by New York's Downstate Medical Center (federal, state, city), the best kidney care cannot be obtained without qualified nephrologists. True, this view is clearly self-serving but like our colleagues -- radiologists who advise more CT scans, gastroenterologists who advocate more endoscopies and cardiologists who assert the value of repeated echocardiograms for young people -- we believe in proactive care.

Whether to modulate hypertension, arrange for evaluation of potential kidney donors or just to provide reassurance, our experience strongly favors early and continuing participation by nephrologists for every patient at near-term risk of ESRD.

Illustrative Cases

Case 1

A 57-year-old Buddhist monk, with known autosomal dominant polycystic kidney disease (ADPKD), returns from four years of service in Bangkok and complains of a "rotten taste," progressive weight loss and generalized itching. Physical examination is remarkable for obvious lethargy and weight loss in a urochrome pigmented hypertensive (187/112 mm Hg) man with a loud precordial friction rub and 4+ lower extremity edema. Both fundi have silver wire arterial narrowing and diffuse flame hemorrhages. Abdominal examination discloses bilateral large, nontender, ballotable flank masses consistent with nephromegaly.

Key laboratory data include: hematocrit 22%, serum albumin 2.5 g/dl, serum creatinine 21.7 mg/dl, serum phosphorous 11.6 mg/dl. While undergoing femoral cannulation for emergency hemodialysis, the patient screams, has a grand mal seizure, manifests right hemiparesis and dies within 90 minutes, despite immediate and appropriateresuscitation attempts. An autopsy performed by the Medical Examiner showed huge, bilateral, cystic renal masses, polycystic liver and lung disease, and a massive subarrachnoid hemorrhage traced to a ruptured aneurysm of the posterior communicating artery.

Q. After review of the chart, the most reasonable inference is:

1. Death would have been prevented by early referral to a nephrologist.
2. Once the diagnosis of ADPKD was known, the patient should have been referred for an immediate cerebral CT scan from which life saving discovery of the aneurysm was probable.
3. Before the patient left the country, routine supervision by a nephrologist would have detected the asymptomatic aneurysm leading to life sustaining neurosurgery.
4. While earlier initiation of uremia therapy would have been desirable, the patient's fatal course would not have been changed.
5. ADPKD afflicting the liver and lungs is an ominous finding in which attempts at life prolongation under any circumstance are futile.

A. Optimized planning of uremia therapy requires time for patient reaction to the grim news that "normal" life has ended. Shock, anger, denial and withdrawal are common responses that vary from patient to patient. The physician and social worker will face strong challenges as they seek to marshal family, clergy and friends into a defense perimeter. When clinical deterioration and systemic complications are advanced, demanding immediate medical intervention, there is little time for discussion and weighing choices other than for the patient to follow the nephrologist's suggestion (bias).

In most U.S. renal programs, Case 1 would be started on acute hemodialysis, an alert for possible intrafamilial kidney donors would be sounded and vascular access surgery scheduled. This patient's early death was preventable had a monitoring program (serum creatinine measurements and quarterly evaluations by a nephrologists) been in place. Sadly, such advanced presentation of near-terminal kidney failure is seen at least once a month in the authors'  municipal hospital.

Case 2

A 34-year-old African American electrical engineer is referred to a nephrologist for evaluation of proteinuria and anasarca. With the exception of discomfort when putting his shoes on caused by ankle swelling, the patient admits to neither systemic nor local symptoms. Physical examination notes: BP 141/76 mm Hg, normal fundi, flat neck veins, normal chest and abdomen, and pitting edema of both legs. Laboratory findings include: hematocrit 35.1%, leukocyte count 8,300, serum albumin 1.9 g/dl, serum creatinine 3.1 mg/dl, serum phosphorous 5.7 mg/dl, HIV culture and antibody negative, hepatitis C antigen positive. 24-hour urinary protein excretion is 15.7 g with a creatinine clearance of 34 ml/min. A percutaneous right kidney biopsy discloses that three of 14 glomeruli were globally sclerotic, while all other glomeruli evinced mild to moderate membranoproliferative glomerulonephritis with diffuse deposition of IgG and Complement.

Q. Optimized subsequent management would:

1. Assign the patient to the nephrologist as coordinating "Team Captain" with the expectation that coordination of care and all needed referrals will be handled promptly.
2. Transfer management to an infectious disease specialist with return to the nephrologist for semiannual reassessment.
3. Enroll the patient in a combined liver-kidney transplant program with the aim of preemptive dual organ transplant before the complications of uremia develop.
4. Return the patient to the referring general practitioner for antiviral therapy with return to the nephrologist for semiannual reassessment.
5. Advise the patient of his imminent need for ESRD therapy and the need to seek any available intrafamilial kidney donors.

A. Unlike the preceding patient, Case 2 permits thought and physician-patient discourse to assess treatment options. Once the probable progression of membranoproliferative glomerulonephritis is grasped, it is highly likely that the patient will consent to initiation of immunosuppressive drug therapy. Accumulating evidence supports trial of a regimen based on mycophenylate, though cyclophosphamide plus corticosteroids is still "standard." Interpretation of serial measurements of daily urine protein excretion and serum creatinine level is best handled by a nephrologist.

About one-third of patients will stabilize with no evident progression over the next five years. Conversely, one-third will deteriorate to ESRD within five years, while one-third will slowly lose renal function and become increasingly symptomatic. Early referral to a nephrologist will help ensure that the course of Case 2 is free of surprises in a better informed patient able to cope with the uncertainties of an inconstant disease.

Case 3

An 81-year-old widowed nursing home resident with type 2 diabetes survived two myocardial infarctions and left below-the-knee amputation attributed to diabetic macrovasculopathy. Vision was severely restricted due to bilateral cataracts and macular edema plus proliferative diabetic retinopathy. Review of the patient's chart showed that the serum creatinine two years earlier was 2.3 mg/dl, one year ago was 3.6 mg/dl and three weeks ago was 7.2 mg/dl. The patient is minimally communicative and unable to state the date or location. BP is 194/93 mm Hg with an irregularly irregular pulse. There is a small area of gangrene in the right great toe. Current lab data includes: hematocrit 21%, serum creatinine 9.9 mg/dl, serum phosphorous 8.8 mg/dl, and serum albumin 2.4 g/dl.

Q. How would you manage this patient?

1. Judgment about medical futility and the patient's mental capacity should be withheld until at least three weeks of vigorous peritoneal or hemodialysis.
2. Earlier referral for initiation of uremia therapy would have had minimal impact in this case.
3. Creation of an arteriovenous fistula in patients with severe diabetic macrovasculopathy risks compromising limb integrity and should not have been performed.
4. Patients who are unable to communicate with the responsible nephrologist should not be subjected to ESRD therapy.
5. Urgent kidney transplantation is not indicated in patients older than age 70.

A. Few issues in clinical nephrology are more vexing and likely to induce stress in the medical team than the decision concerning when "enough is enough." Both dialytic therapy (peritoneal dialysis and hemodialysis) and kidney transplantation are wondrous life sustaining regimens. But, is the availability of a treatment sufficient justification for its use? Should every uremic patient be treated irrespective of potential outcome?

What about a 95-year-old individual with extensive metastatic cancer who develops agonal obstructive uropathy or acute tubular necrosis? Or, as in Case 3 above, a demented patient with no hope of restoration to a life status that includes functional awareness?

A reasonable approach is to "test" the value of dialysis by a limited course of peritoneal or hemodialyses. The flaw in this approach is the ethical challenge that results from the discontinuance of dialysis in a patient whose life is sustained by dialysis. Much greater upset occurs in the family and the medical team of a patient who starts a limited course of dialysis compared to a patient who did not begin dialysis because their cases were deemed futile. Thus, there is no simple, correct answer to the issue raised by Case 3.