Elevated Serum Creatinine and a Normal Urinalysis: A Short Differential Diagnosis in the Etiology of Renal Failure

Abstract

Objective. The diagnosis of renal failure with a normal urinalysis represents a short differential diagnosis that has not been characterized in the literature. This study was designed to identify the specific disease states that encompass this interesting renal condition. Material. The Regenstrief database, which is an electronic medical record containing inpatient laboratory and other clinical data for patients admitted to Wishard Memorial Hospital, was utilized to provide data for this study. Methods. The database was culled to provide data for hospitalized patients admitted between March 1, 1992, and March 1, 2001, with the concurrent findings of a serum creatinine greater than or equal to 2 mg/dL, a normal urinalysis, and diagnoses of obstructive uropathy, multiple myeloma, prerenal azotemia, hypertensive nephrosclerosis, interstitial nephritis, renal vascular disease, hypokalemic nephropathy, and hypercalcemia, as identified by their corresponding ICD-9 codes. Results. A search of the Regenstrief database yielded a total of 190,343 patient admissions. There were 515 patient admissions with renal failure and a concurrent normal urinalysis. The largest specific diagnostic categories within this group were hypertension and prerenal azotemia. Conclusions. An elevated serum creatinine and normal urinalysis present a short differential for the etiologies of renal failure and include such entities as hypertensive nephrosclerosis, prerenal azotemia, obstructive nephropathy, interstitial nephritis, renal vascular disease, and various electrolyte abnormalities. An awareness of these specific disease states may lead to an earlier diagnosis and more effective treatment of renal failure.

Keywords:

• ARF
• CKD
• urinalysis
• serum creatinine

INTRODUCTION

Acute and chronic renal failures are a significant cause of patient morbidity and mortality. An early diagnosis can lead to treatment strategies that forestall the progression of renal failure or, in such cases, of obstructive uropathy.[1],[2] Recent evidence suggests that early treatment and referral to a nephrologist can favorably impact the effective diagnosis and treatment of patients with renal failure.[3–7] The current study proposes a strategy for the early diagnosis of patients with renal failure and a normal urinalysis. The diagnosis of renal failure with a normal urinalysis represents a short differential diagnosis that has not been characterized in the literature. The clinical entities associated with this combination of events include type 2 diabetes,[8] obstructive uropathy,[9],[10] multiple myeloma,[10] prerenal azotemia,[10–14] hypertensive nephrosclerosis,[10] interstitial nephritis,[15] renal vascular disease,[16–19] hypokalemic nephropathy,[20] and hypercalcemia.[21] In addition, renal failure with a normal urinalysis has also been associated with the use of indomethacin[22–24] and lithium.[18],[25–27] An extensive MEDLINE search spanning the years 1966 to the present did not produce any literature that directly studied the phenomenon of renal failure with a normal urinalysis. As a result, evidence supporting the majority of the clinical entities has been gleaned from isolated case reports and a careful review of related studies. The paucity of literature describing the diagnosis of renal failure and normal urinalysis, coupled with the seriousness of the medical conditions found in the differential diagnosis, underlie the importance of an evaluation of this topic.

METHODS

For the purposes of this study, renal failure was defined as a level of serum creatinine ≥2.0 mg/dL. Initially, no attempt was made to distinguish between acute and chronic renal failure, but subsequent analysis did include acute and chronic renal failure diagnosis as identified by International Classification of Disease (ICD-9) codes. A urinalysis was defined as normal if red and white blood cell counts were fewer than five per high power field and if urinary protein, hemoglobin, and bacteria were all negative. The methodology for urinalysis at our hospital utilized the Clinitek 200 Plus^TM device (Bayer Corporation, Elkhart, Indiana) and microscopic screening of all urine specimens. The Regenstrief Medical Record System was used to extract laboratory data from March 1, 1992, to March 1, 2001.[28] This electronic medical record contains inpatient laboratory and other clinical data for patients admitted to Wishard Memorial Hospital, Indianapolis, Indiana. For the purposes of this project, data were extracted for hospitalized adult patients with concurrent findings of a serum creatinine ≥2.0 mg/dL, a normal urinalysis, and the aforementioned clinical entities as identified by their corresponding ICD-9 codes. The ICD-9 codes were based on those codes entered at the date of discharge. For patients with multiple diagnoses, clinical judgment was utilized on an individual patient basis to decide which diagnostic category to assign. For example, a patient admitted with the dual diagnosis of hypertension and obstructive uropathy would be assigned to the latter diagnostic category because it was felt that this would be the more likely renal pathology that caused their renal insufficiency. The patients had an average of three diagnostic categories. A retrospective chart review was not performed. The time interval for the collection of these data elements spanned one to three days post-admission. The serum creatinine level was ≥2.0 mg/dL, and normal urinalysis on the date of admission was paired with the same patient discharge ICD-9 code. It was assumed that a certain percentage of patients with a discharge diagnosis of any type of renal failure would enter the hospital and have a normal initial serum creatinine and urinalysis. No attempt to quantitate 24-hour urine protein or urinary-protein-to-creatinine ratios or identify microalbuminuria was made. Creatinine clearance was neither queried nor estimated by Cockcroft-Gault or MDRD formulae. The study was approved by the Indiana University, Purdue University, and Clarian Institutional Review Boards (Indianapolis, Indiana). The data obtained from the search of the Regenstrief database were then tabulated, and the diagnostic categories were expressed as a percentage of total diagnostic categories. The hypothesis was considered valid if the cases of renal failure with a normal serum creatinine were found to conform to the hypothesized differential diagnosis. A pilot search of the Regenstrief database was undertaken prior to this study to establish that an adequate patient population existed among those patients who received concurrent laboratory tests for urinalysis and serum creatinine.

RESULTS

Over the nine-year time interval from March 1, 1992, to March 1, 2001, there were a total of 190,343 admissions within the Regenstrief database. A subset of this group, 38,257 patients, comprised those with the target discharge diagnoses. Within this group, 28,259 patients had a serum creatinine and urinalysis; patients with a serum creatinine greater than 2.0 mg/dL and a urinalysis numbered 12,197. Finally, within the patient population that had a serum creatinine greater than or equal to 2.0 mg/dL, 515 (4.2%) had a concurrent normal urinalysis. Figure 1 describes the distribution of the patients within the various subsets, and Figure 2 illustrates the percentage composition of the diagnostic categories. As shown in Figure 2, the majority of the patients within this study who have a serum creatinine greater than or equal to 2.0 mg/dL and a normal urinalysis have hypertensive nephropathy (HN) as their primary diagnosis. Specifically, HN was found in 38% of cases. The next closest specific diagnostic categories were prerenal azotemia and obstructive uropathy, with 20.8% and 3.3%, respectively. However, miscellaneous categories that comprise a number of electrolyte abnormalities, renal neoplasms, ARF and CRF subcategories, fluid overload, and coding error represented the largest proportion of patients (39%). Figure 3 illustrates the distribution of the patients within the miscellaneous diagnostic categories.

Figure 1 Numerical distribution of the patients in the various study subsets.

Figure 2 Percentage distribution of the patients in the selected diagnostic categories. HN=hypertensive nephrosclerosis; HKN=hypokalemic nephropathy; OU=obstructive uropathy; MISC=miscellaneous; AIN=acute interstitial nephritis; HC= hypercalcemia; PRA=prerenal azotemia; MM=multiple myeloma; RVD=renal vascular disease.

Figure 3 Miscellaneous diagnoses associated with a serum creatinine >2 mg/dL and a normal urinalysis. ARF ATN = acute renal failure with acute tubular necrosis; CRF non-HD = chronic renal failure without hemodialysis; ARF unspec = acute renal failure unspecified; Mg++ metabolism = magnesium metabolism; Phos metabolism = phosphorus metabolism.

DISCUSSION

The data indicate that a small subset of patients (4.2%) have an elevated serum creatinine and concurrent normal urinalysis. This emphasizes the importance of obtaining a urinalysis in all patients who present with an elevated serum creatinine and the utility of this test in screening patients with serum creatinine levels greater than 2.0 mg/dL. Our study also showed that the majority of patients with an elevated serum creatinine and a normal urinalysis have hypertension and presumably hypertensive nephrosclerosis (HN). Hypertension was the most common etiology of renal failure in our study population—an important finding, as it is widely known that adequate treatment of hypertension in patients with kidney disease can forestall the progression of renal failure.[1] In addition, this study demonstrated that urinalysis and serum creatinine determinations are being underutilized. Of the 38,257 patients that fit one of the diagnostic categories, 26% did not have both a serum creatinine determination and urinalysis on admission. This occurred despite the long-established utility of both of these laboratory tests in providing diagnostic and therapeutic information.

There were also some unexpected results from this study; specifically, the patients found in the chronic renal failure and renal cell carcinoma diagnostic categories. A normal urinalysis is not an anticipated laboratory finding in patients with chronic renal failure. In a study of CRF patients, urinalyses were universally replete with white blood cells and white blood cell casts that were presumably the interstitial accumulations of leukocytes in their atrophic kidneys.[29] An elevated serum creatinine is also an atypical finding in patients diagnosed with renal cell carcinoma.[9],[30–32] However, hypertension has been observed in 20–40% of patients with renal cell carcinoma.[10],[30] This has been attributed to concomitant polycythemia (erythropoietin-producing renal cell carcinoma), renin-secreting adenocarcinoma, arteriovenous shunting, or the compression or invasion of the renal artery by the mass.[9],[31] Perhaps the patient in this study had uncontrolled hypertension secondary to an underlying renal cell carcinoma that led to hypertensive nephropathy.

Another unexpected finding in this study was the absence of a normal urinalysis and an elevated serum creatinine in the 210 cases of multiple myeloma. This was not anticipated because it had been identified in the literature.[9] The absence of multiple myeloma cases with a normal urinalysis and an elevated serum creatinine may be explained by the sample size, patient population, and the reliance on ICD-9 codes as the primary means to cull the diagnoses from the database, or a combination of these three variables. Additionally, our data suggest more advanced disease in our population of patients with multiple myeloma and thus a higher likelihood of an abnormal urinalysis. Finally, there remains the possibility that a urinalysis was not obtained with a serum creatinine level in the specified time period of this study.

There are some important limitations in this study that may have affected the accuracy of the data obtained. These limitations are associated with a subset of the diagnostic categories and the threshold serum creatinine utilized to identify patients with renal insufficiency. First, the hypertensive nephrosclerosis (HN) diagnostic category may have a frequency that is spuriously elevated. Through coding error, this diagnostic category may have erroneously included patients whose primary diagnosis is actually renal vascular disease or renal artery stenosis, as hypertension is a frequent finding in these clinical entities. Consequently, the renal vascular disease and renal artery stenosis diagnostic categories may be under-represented. Our study identified two patients out of 181 in these categories with a normal urinalysis and an elevated serum creatinine. The possibility of coding error was well illustrated by one patient in the study who was found to have a normal urinalysis and yet was given the diagnostic category of “hematuria” (see Figure 3).

Figure 3 also illustrates another example of the pitfalls associated with culling data by ICD-9 codes—the large percentage of patients that are found in the “ARF unspecified” diagnostic category. This category may actually contain patients with a variety of the etiologies of ARF illustrated in Figure 2. Unfortunately, a lack of initial coding precision has relegated a large percentage of these patients to a mostly undefined diagnostic category.

Additionally, the electrolyte abnormalities identified in Figure 3 are problematic. The patients within the diagnostic categories of hypokalemia, hyperkalemia, and disorders of magnesium and phosphorous metabolism likely did not develop renal failure from these abnormalities. These were instead probably associated findings and again represent coding imprecision. The hypokalemic nephropathy diagnostic category also has an important limitation: its failure to delineate other possible etiologies of this condition. Hypokalemic nephropathy can occur secondary to medications (e.g., diuretics and laxatives), anorexia nervosa, primary aldosteronism, or diabetes mellitus.[20] Hypokalemia may complicate other causes of renal failure. Given the study design, though, these data are lacking. The prerenal azotemia category also fails to elucidate its possible etiologies, such as hypotension, CHF, nephrotic syndrome, and hepatic disease.[9],[10] Likewise, the ATN category fails to describe its likely causes. The ATN category may contain patients with PRA related to volume depletion or drugs, such as ACE inhibitors, ARBs, or NSAIDs. No attempt was made in this study to retrospectively review individual charts.

The other important limiting area of this study concerns the threshold serum creatinine value selected to identify patients with renal insufficiency. A value of 2.0 mg/dL (177 μmol/L) was chosen to reduce the potential “false-positives”, those patients with normal renal function and a normal urinalysis. However, patients can manifest renal insufficiency with serum creatinine values much less than 2.0 mg/dL, and this selection criterion may have missed those patients. This study also treated the genders equally, which could have led to an incomplete representation of the patients with renal insufficiency in our study population. Gender differences, especially in muscle mass, have important implications in the diagnosis of renal insufficiency. Generally, women have less muscle mass and thus a lower serum creatinine than men; as a result, women may manifest renal insufficiency at lower serum creatinine levels. Additionally, the study's hospital population has a significant representation of African American patients (46%), another factor that can affect serum creatinine, creatinine clearance, and estimated GFR.[33] A study addressing the optimization of care for patients approaching end stage renal disease suggests that early screening for chronic kidney disease be considered at serum creatinine levels of 2.0 mg/dL (177 μmol/L) in men and 1.5 mg/dL (133 μmol/L) in women.[34] The present study's authors recommend in concert with K-DOQI that in clinical practice and for the purposes of the diagnostic algorithm proposed by this study, the serum creatinine should not be completely relied upon to rapidly assess a patient's renal function. A GFR estimation tool like the Cockcroft-Gault equation or the MDRD formula offers greater accuracy in determining the level of renal function.[35–37]

Finally, the algorithm presented here did not identify patients with diabetes, even though this is a common morbidity in this patient population. Approximately 20% of all patients admitted to Wishard Hospital in 2005 have a diagnosis of diabetes.[38] This may have added an additional category to the differential diagnosis, especially given the recent study by Kramer et al.[8] This study surveyed adults with type 2 DM and found that 30% of individuals with a GFR less than 60 mL/min per 1.73 m² BSA lacked albuminuria and retinopathy.[8]

Even with these notable limitations, this study does provide evidence to support our primary hypothesis: the findings of an elevated serum creatinine and a normal urinalysis present a short differential for the etiologies of renal failure. As a result, this study provides a targeted differential diagnostic list for those patients who present with an elevated serum creatinine and a normal urinalysis.

The majority of patients with a serum creatinine greater than 2.0 mg/dL have an abnormal urinalysis and hypertension, emphasizing the importance of blood pressure measurement, serum creatinine analysis, and urinalysis in the early identification of renal disease. The finding of an elevated serum creatinine with a normal urinalysis has important diagnostic and therapeutic implications as well. The authors advocate confining screening laboratory studies, including serum creatinine determinations and urinalysis, to older patients and those with morbidities that can cause renal damage (e.g., diabetes and hypertension). Indeed, screening urinalysis studies to determine albuminuria in an attempt to forestall chronic kidney disease and decrease mortality has been shown to be cost-effective only in these high-risk groups.[39]

Based on the results of this study, the authors recommend that prerenal and postrenal etiologies such as obstructive uropathy be considered first in those patients with normal urinalysis and abnormal renal function. However, other diagnostic entities to consider in this setting include hypertensive nephrosclerosis, renal vascular disease, multiple myeloma, hypercalcemia, hypokalemic nephropathy, renal cell carcinoma, and interstitial nephritis. If prerenal and postrenal etiologies can be excluded, then additional studies should be considered based on the pertinent clinical presentation and physical exam findings. These additional studies may include a serum calcium level, SPEP/UPEP, urine electrolytes and urine creatinine, urine eosinophils, and imaging such as renal ultrasound or MRA of the renal arteries. A renal biopsy may also prove necessary if the diagnosis remains an enigma. The ramifications of this diagnostic strategy are that it may lead to the earlier detection and treatment of renal failure, as well as delineate a serious primary condition.

Notes

38. Regenstrief Medical Record System, Wishard Memorial Hospital, Indianapolis, Indiana. Database inquiry of 2005 performed January, 2006.