Why is serum creatinine not used alone as an indication to initiate haemodialysis?

Why Serum Creatinine Alone Should Not Guide Hemodialysis Initiation

Serum creatinine is an unreliable marker for determining when to start hemodialysis because it is profoundly influenced by muscle mass rather than kidney function alone, and initiating dialysis based solely on creatinine or eGFR levels—rather than uremic symptoms—does not improve mortality or quality of life. 1

Primary Limitations of Serum Creatinine in Advanced CKD

Muscle Mass Confounding

• Serum creatinine-based eGFR is substantially influenced by muscle mass in advanced CKD, making it both a marker of sarcopenia AND kidney function—not just renal function alone. 1

• Low muscle mass in elderly patients, those with sarcopenia, chronic illness, or malnutrition falsely lowers serum creatinine, masking significant renal impairment and causing overestimation of GFR. 1, 2

• A creatinine level of 1.2 mg/dL can correspond to a creatinine clearance of 110 mL/min in a young, muscular male but only 40 mL/min in an elderly woman with low muscle mass. 2

• Among elderly patients with calculated GFR ≤50 mL/min, 40% had serum creatinine levels within the normal laboratory range, demonstrating how creatinine fails to detect severe renal impairment. 2, 3

Paradoxical Mortality Associations

• Most observational cohort studies showing higher mortality with higher eGFR at dialysis initiation likely reflect confounding by indication—sicker, frailer patients with more comorbidities and lower muscle mass start dialysis earlier, not that early dialysis causes harm. 4

• While cohort studies assessing the association between eGFR at initiation of dialysis and mortality have shown higher risk for death with higher eGFR, the same association is not demonstrable with measured clearances, confirming that creatinine-based estimates are misleading. 1

• Creatinine values from incident ESRD populations have a weak relationship with the timing of dialysis initiation but represent a strong measure of health status—lower creatinine predicts higher mortality because it reflects sarcopenia and frailty. 5

Evidence Against Using GFR/Creatinine Thresholds Alone

The IDEAL Study Findings

• The landmark IDEAL Study randomized 828 patients to early (eGFR 10-14 mL/min) versus late (eGFR 5-7 mL/min) dialysis initiation and found no significant difference in time to death, cardiovascular events, infectious events, or dialysis complications. 1

• The median difference in time to dialysis initiation was only 5.6 months because 76% of patients assigned to late start required early initiation due to uremic symptoms, demonstrating that symptoms—not GFR—ultimately dictate timing. 1

• There was no difference in cardiac structure, function, or total health care costs between early and late start groups. 1

Guideline Consensus

• There is no compelling evidence that initiation of dialysis based solely on measurement of kidney function leads to improvement in clinical outcomes, including overall mortality. 1, 4

• In otherwise asymptomatic individuals, there is no reason to begin maintenance dialysis solely based on a serum creatinine or eGFR value. 1, 4

• The KDOQI 2015 guidelines explicitly state that sufficient data exist to discourage reliance on a specific eGFR level for dialysis initiation. 1

What Should Guide Dialysis Initiation Instead

Absolute Indications (Life-Threatening Complications)

• Initiate RRT emergently when life-threatening changes in fluid, electrolyte, and acid-base balance exist, including: 4
  • Severe hyperkalemia unresponsive to medical management 4
  • Volume overload with pulmonary edema refractory to diuretics 4
  • Severe metabolic acidosis unresponsive to conservative therapy 4
  • Uremic complications including pericarditis, encephalopathy, or bleeding 4

Symptomatic Uremia

• Symptomatic uremia (regardless of GFR level) is an indication for RRT, with uremic symptoms including nausea, anorexia, pruritus, altered mental status, and protein-energy malnutrition despite optimized intake. 4

• The decision to initiate dialysis should not be based solely on estimated GFR because creatinine-based formulae are inaccurate in patients with ESKD. 6

Practical Clinical Algorithm

Step 1: Assess for Absolute Indications

• Check for diuretic-unresponsive pulmonary edema, severe hyperkalemia (>6.5 mEq/L unresponsive to treatment), uremic pericarditis/encephalopathy/bleeding, or severe metabolic acidosis (pH <7.1). 4
• If any present, initiate dialysis immediately regardless of creatinine or eGFR. 4

Step 2: Systematically Evaluate for Uremic Symptoms

• Assess for protein-energy malnutrition, nausea, vomiting, anorexia, pruritus, altered mental status, and declining functional status. 4
• If symptomatic uremia is present, initiate dialysis even if eGFR >10 mL/min. 4

Step 3: In Asymptomatic Patients with Stage 5 CKD

• Dialysis may be safely delayed until eGFR is at least as low as 5-7 mL/min/1.73 m² if there is careful clinical follow-up and adequate patient education. 6
• Do not initiate dialysis based on eGFR or creatinine alone in asymptomatic patients, even at eGFR <10 mL/min/1.73 m². 4

Step 4: Consider Alternative GFR Measurements When Creatinine is Unreliable

• In patients with altered muscle mass (elderly, sarcopenic, malnourished, or highly muscular), measure cystatin C-based eGFR, which is less biased by muscle mass. 2
• In selected cases where symptoms appear discordant with eGFR, consider direct measurement of GFR or 24-hour urine collection for creatinine clearance. 1

Critical Pitfalls to Avoid

• Never delay dialysis until severe complications develop in critically ill patients—earlier intervention based on clinical status is safer than waiting for arbitrary creatinine thresholds. 4

• Serum creatinine alone should not be used to guide timing decisions; always calculate GFR and interpret in context of muscle mass, age, and nutritional status. 4, 2

• In elderly patients with multiple comorbidities, serum creatinine does not reflect age-related GFR decline due to concomitant decline in muscle mass, leading to underrecognition of severe renal failure. 2, 3

• GFR must decline to approximately half the normal level before serum creatinine rises above the upper limit of normal, making it an insensitive marker for detecting renal impairment. 2

• Creatinine concentration is affected by creatinine secretion, generation, and extrarenal excretion—not just GFR—further limiting its utility as a standalone marker. 2, 7