Few diagnostic categories draw more denial activity right now than kidney diseases. Acute kidney injury (AKI), acute tubular necrosis (ATN), and chronic kidney disease (CKD) each carry their own documentation traps.
Payers have become fluent in exploiting the gaps between them to support issuing a denial. What makes this particularly frustrating is that there is clinical agreement on how to identify these conditions and clear coding guidelines, but AKI and ATN remain auditing targets.
The Kidney Disease: Improving Global Outcomes (KDIGO) 2012 Clinical Practice Guideline remains the operative definition of AKI, and it is deliberately simple: a serum creatinine rise of at least 0.3 mg/dL within 48 hours, a rise to at least 1.5 times baseline within 7 days, or urine output below 0.5 mL/kg/h for 6 hours. Any one of the three is sufficient. However, I continue to see payers argue that AKI criteria “do not apply” until a patient has been adequately rehydrated, effectively demanding that the diagnosis prove itself twice.
This conflates two separate clinical questions. Whether AKI was present at the moment the criteria are met is not the same question as which subtype of AKI the patient has: prerenal, which codes to AKI, or intrinsic injury such as ATN. KDIGO criteria do not require these values to persist after fluid resuscitation before the diagnosis of AKI is valid. Fluid resuscitation is the primary treatment for AKI, which requires supportive measures and the avoidance of nephrotoxins.
A rapid, complete response to fluids does not retroactively erase kidney injury; it identifies the injury as prerenal rather than intrinsic (i.e., ATN). That distinction matters clinically, and it belongs in the documentation, but it is a question of etiology, not of whether the diagnosis of AKI or ATN is valid.
ATN, also referred to as acute tubular injury (ATI), is a form of intrinsic (renal) AKI that is characteristically persistent (AHA Coding Clinic 2nd Quarter 2026 p. 5). True ATN takes 72 hours or longer to resolve even with adequate fluid resuscitation because the sloughed tubular epithelium must regenerate. That is a real, evidence-based way to differentiate AKI from ATN. Because kidney biopsy is rarely pursued clinically, there is no true non-invasive gold standard for diagnosing ATN.
The most effective non-invasive strategy for identifying ATN is the Perazella-Coca urine sediment scoring system (Perazella et al., Clin J Am Soc Nephrol, 2008;3(6):1615-1619), in which a score of 2 or higher (based on granular casts and renal tubular epithelial cells) is described as an extremely strong predictor of ATN, carrying a positive predictive value of 100 percent in patients with a high pretest probability of the diagnosis. However, this is a manual process that is rarely performed because most hospitals rely on automated urine analyzers, which are poorly suited to identifying granular casts and renal tubular epithelial cells.
A nephrologist’s manual read of the urinalysis can reach the correct diagnosis of ATN over 90 percent of the time compared with just 19 percent when relying on the standard automated lab report. In practice applying this scoring system reliably requires a nephrology consultation and a clinician willing to personally examine the sediment, which is not something a treating hospitalist can expect from a routine urinalysis.
All of this to say that when appealing a denial for ATN, CDI professionals should push back if the payer cites a lack of cellular evidence, including muddy brown casts. Research has also shown that these casts may be absent in up to 30% of patients with ATN. The absence of cellular evidence should never be used to exclude the diagnosis of ATN.
CKD presents a different kind of documentation discipline. Per the FY 2027 ICD-10-CM Official Guidelines, Section I.C.14.a.1, CKD must be staged 1 through 5 or documented as end-stage renal disease (ESRD), and that stage must come from the provider, not from a coder or CDI specialist reading a GFR off a lab report. Coding CKD without a stage understates how sick the patient is, but documentation of “chronic renal insufficiency” is worse. If the provider were to drop the word “chronic,” the term maps to N28.9, which is not even a CKD code.
Where this gets genuinely interesting is the boundary between Stage 5 CKD and ESRD. Both describe a GFR under 15 mL/min. The difference on paper is whether the patient requires ongoing dialysis. N18.5 is reported for Stage 5 without dialysis, but N18.6 plus Z99.2 (Dependence on renal dialysis) for ESRD. But this is not merely a coding nuance; it is a Medicare eligibility trigger. Under Section 226A of the Social Security Act, a documented diagnosis of ESRD certified on the CMS-2728 Medical Evidence Report allows an individual of any age to become entitled to premium-free Medicare Part A.
If a patient has started maintenance dialysis, the documentation should reflect ESRD, not “Stage 5 CKD,” and CDI professionals should understand why that distinction carries weight well beyond the inpatient claim.
It is also possible for a patient to have both CKD and AKI superimposed on it. However, there is no “acute on chronic renal failure” combination code in ICD-10-CM. Per Guideline I.C.9.a.2 and I.C.9.a.3, when hypertensive CKD and acute renal failure are both present, the acute renal failure must also be coded, sequenced according to the reason for the encounter.
Another common argument is that a patient cannot have AKI in the setting of ESRD because these patients are unable to produce urine. ESRD is defined by GFR/dialysis-dependence, not by anuria. Per the United States Renal Data System (USRDS) only 15% of patients starting dialysis in 2020 had an eGFR below 5 mL/min/1.73m². The substantial majority of those with ESRD retain meaningful residual kidney function that can range from oliguria to normal urine output levels. However, elevated lab values in the setting of ESRD do not equate to AKI if the patient has missed or was unable to complete their regular dialysis schedule. This scenario is associated with fluid overload (and hyperkalemia) not AKI.
AKI on ESRD requires a new insult to the kidneys (e.g., infection, hypotension, nephrotoxic exposure, obstruction) that worsens the patient’s residual function or urine output from their individual baseline. This is why KDIGO’s definition requires a change from the patient’s baseline, whatever that baseline is. CDI and coding professionals can use past serum creatinine levels to try to establish a baseline if the patient has frequent visits to the facility when querying the provider for clarification.
What ties all this together is not clinical complexity; it is discipline in going back to the primary source rather than accepting a payer’s paraphrase of it. KDIGO says what it says about fluid resuscitation. The Official Guidelines say what they say about staging and dual coding. CMS’s own ESRD entitlement provisions say what they say about who qualifies for Medicare and why.
When a denial letter reinterprets any of these, the correct response is not a clinical argument about the patient’s course; it is the guideline’s actual language, quoted back. Renal disease coding will keep drawing scrutiny as kidney diseases continue climbing in the Medicare inpatient volume.
The programs that hold their ground will be the ones that treat the primary sources as non-negotiable, not the ones that quietly adjust their documentation and coding to satisfy the payer.
