Discordance between serum creatinine and

creatinine clearance for identification of ED patients with abdominal pain at risk for Contrast-induced nephropathy

Roger A. Band MD*, David F. Gaieski MD, Angela M. Mills MD, Keara L. Sease MAEd, Frances S. Shofer PhD,

Jennifer L. Robey BSN, Judd E. Hollander MD

Department of Emergency Medicine, University of Pennsylvania School of Medicine, Hospital of the University of Pennsylvania, Philadelphia, PA 19104-4283, USA

Received 2 May 2006; revised 26 July 2006; accepted 31 July 2006

Abstract

Study Objective: Despite creatinine clearance (CrCl) being a better estimate of renal function, serum creatinine (Cr) is more commonly used to screen for renal insufficiency in patients scheduled for an enhanced Abdominal computed tomography (CT) in an attempt to reduce the likelihood of contrast- induced nephropathy (CIN). Our objective was to determine the incidence of renal insufficiency (a CrCl b60 mL/min) among patients who have serum Cr below 1.5 mg/dL (the most commonly accepted Cr cutoff for the administration of intravenous contrast). This study was conducted in a population of emergency department patients with acute abdominal pain being considered for CT scan.

Methods: We performed post hoc analysis of a prospective cross-sectional study that enrolled nongravid adults with acute nontraumatic abdominal pain. Patients on dialysis were excluded. The data that we collected included demographics, history, duration/description of pain, patient reported weight, laboratory data, imaging studies, and final diagnosis. Creatinine clearance values (b or N60 mL/min) were compared to Cr values of 1.0, 1.2, 1.5, and 1.8 mg/dL to determine the percentage of patients at risk for nephropathy after contrast injection at each Cr cutoff. Descriptive statistics were used with 95% confidence intervals (CIs).

Results: Seven hundred sixty-five patients were enrolled; 59% (451/765) had an abdominal CT scan. Of 108 patients with CrCl less than 60 mL/min, 59 patients had a Cr less than 1.8 mg/dL (55%; 95% CI, 45%-64%); 43 had a Cr less than 1.5 mg/dL, the most commonly accepted Cr cutoff for contrast administration (40%; 95% CI, 31%-50%); 21 patients had a Cr less than 1.2 mg/dL (19%; 95% CI,

12%-28%); and 10 had a Cr less than 1.0 mg/dL (9%; 95% CI, 5%-16%).

Conclusion: The most commonly used Cr cutoff (1.5 mg/dL) for contrast administration fails to identify 40% of the patients at risk for CIN. Future studies should address whether using CrCl rather than serum Cr decreases the incidence of contrast-induced nephropathy.

D 2007

* Corresponding author. Tel.: +1 215 662 7260; fax: +1 215 662 3953.

E-mail address: [email protected] (R.A. Band).

0735-6757/$ – see front matter D 2007 doi:10.1016/j.ajem.2006.07.011

Introduction

Annually, nearly 9 million patients undergo diagnostic imaging of the abdomen. Many of these studies involve the administration of intravenous (IV) Contrast material and lead to the development of contrast-induced nephropathy (CIN) in up to 14% of patients [1-4]. In an effort to reduce the likelihood of CIN [3,5] and the increased morbidity and mortality associated with it [2,3,6], serum creatinine (Cr), a rough estimate of renal function [7-9], is used to screen for underlying renal insufficiency in patients receiving an abdominal computed tomography scan.

Development of acute renal failure (ARF) is associated with increased in-hospital mortality [2,6], and contrast- induced nephropathy (defined as an increase of serum Cr of 25% or higher within 2 days of receiving contrast or as a rise in plasma Cr of 0.5 mg/dL above baseline) is a leading cause of acute renal failure in this patient population [3]. Although the clinical significance is variable, several studies have demonstrated markedly increased mortality. In one study by Levy et al [6], the development of CIN increased mortality by 27%; therefore, it is imperative that we identify patients at risk for this disease.

The calculated creatinine clearance (CrCl) has been demonstrated by many authors to be a more accurate measure of renal function than serum Cr, taking into account patient age, weight, and sex. This has been discussed by Stevens et al [10] and shown in work by Duncan et al [7], who noted a bsubstantial prevalence of significantly abnormal renal function among patients identified by laboratories as having normal range serum creatinine,Q when calculating glomerular filtration rate (GFR) using

Setting

This study was conducted at the Hospital of The University of Pennsylvania ED, an urban tertiary care center with an annual ED census of over 52,000 patients.

Participants

Participants were nongravid adult ED patients with acute, nontraumatic abdominal pain, who consented to be enrolled in a serum marker study between April and November 2004. All participants provided written informed consent, and the study was approved by the University of Pennsylvania’s Institutional Review Board. Patients had to be 18 years or older and present to the ED with abdominal pain of less than 3 days duration to be included in the study. Patients were excluded if they were pregnant or had sustained blunt or penetrating trauma during the 7 days before ED presentation.

Measurements

Patients were prospectively enrolled in the ED. Struc- tured data collection for this study was performed by trained research assistants, 16 hours per day and 7 days per week during the study period. Recorded data included patient demographics, medical history, duration and description of pain, basic laboratory data, self-reported patient weights, results of imaging studies, and final diagnoses.

Creatinine values were obtained and compared with the calculated values for the CrCl. These calculations were made using the Cockcroft-Gault equation:

(140 — age) x lean body weight [kg]

the Cockcroft-Gault equation.

However, most emergency medicine physicians do not calculate CrCl as part of their routine practice [11,12].

CCr, in mL/min =

PCr [mg/dL]x 72

( x 0.85 if female)

Most institutions use a serum Cr of 1.5 mg/dL to

approximate a CrCl of 60 mL/min [13], the minimum estimated GFR considered to be safe for administration of IV contrast material [9,14]. The process of substituting serum Cr for CrCl has the potential to allow many patients at risk for CIN to be unwittingly exposed to IV contrast.

We hypothesized that the use of Cr rather than CrCl to screen for patient at risk of CIN would fail to identify many at risk patients. We evaluated this hypothesis in a population of emergency department (ED) patients with abdominal pain, who are likely to receive IV contrast for an abdominal CT.

Methods

Study design

We conducted a cross-sectional study of ED patients with a chief complaint of abdominal pain to determine the relationship between various cutoffs for serum Cr and a CrCl of 60 mg/dL.

Lean body weight was self-reported by patients in the ED.

Outcome measures

We prospectively evaluated patients who’s serum Cr fell into 1 of 4 Cr cutoff values (1.0, 1.2, 1.5, and 1.8 mg/dL) to determine the percentage of patients who would have been at risk for contrast-induced nephropathy (CrCl b60 mL/min) after contrast injection at each Cr cutoff level. These specific Cr cutoff values were chosen by the authors as clinically useful breakpoints above and below a serum Cr of

1.5 mg/dL, which had the potential to alter the sensitivity and specificity of Cr for detecting reduced CrCl. A Cr of

1.5 mg/dL is felt to be a marker for bsafeQ administration of

IV contrast and is considered by most experts to best correspond to a CrCl of 60 mL/min [1,14].

Data analysis

Summary baseline characteristics are presented as means with standard deviation for continuous variables and as frequencies and percentages for categorical variables. The

These data are presented as percentages with 95% confi- dence intervals (CIs).

Table 1 Patient characteristic

Pain

Onset

	No.	Percent
Characteristic
Female sex	381	62
Male sex	234	38
No. of patients that had	363	59
an abdominal CT Age (y)	43.5 F 17

Sudden	342	45
Gradual	389	51
Characterization of pain Sharp	416	55
Crampy	300	39
Burning	54	7
Dull/Gnawing	143	19
type of pain Constant	547	72
Intermittent	187	24
Pain stopped	35	5
Location of pain Diffuse	191	25
Right, left lower quadrant/right,	438	57
left upper quadrant Flank	111	15
Epigastric	132	17
Periumbilical	51	7
Suprapubic	62	8
Related symptoms Nausea	481	63
Diarrhea	131	17
Constipation	59	8
Dysuria	40	5
Bloating/Gas	78	10
Reflux/Indigestion	45	6
Fever	74	10
Jaundice	10	1
Vaginal discharge	14	2
Vaginal bleeding	32	4
Comorbidities abdominal surgery	196	26
Pelvic surgery	70	9
Other surgery	57	7

primary outcome to be assessed was the incidence of preexisting renal insufficiency (as defined by a calculated CrCl of b60 mL/min) at predetermined serum Cr levels.

Results

Overall, 765 patients were enrolled. In total, 451 patients (59%) had an abdominal CT scan. The mean age of patients in this study was 43.5 F 17 years; 62% of patients were women; 59% were African American. The presenting characteristics of the patients and their final diagnoses are shown in Table 1.

Based upon a CrCl less than 60 mL/min, 108 patients were at risk for CIN. Forty-three (40%) of these patients had a normal serum Cr less than 1.5 mg/dL (95% CI, 31%-50%), placing them at increased risk of contrast administration. Conversely, 10 patients with a Cr higher than 1.5 mg/dL had a CrCl higher than 60 mL/min and may have been incorrectly prevented from having a contrast-enhanced abdominal CT. When administration of IV contrast is based solely upon a serum Cr value less than

1.5 mg/dL, the test has a sensitivity of detecting patients with a decreased CrCl of 60% (65/108) (95% CI, 50%- 69%) and a specificity of 98% (647/657) (95% CI, 97%-

99%) as demonstrated in Table 2.

Using a more conservative Cr cutoff of 1.2 mg/dL failed to identify 21 patients at risk for CIN (19% [21/108]; 95% CI, 12%-28%) and would have prevented 79 patients (48%; 95% CI, 40%-55%) with CrCl higher than 60 mL/min from receiving contrast for their CT scan. The sensitivity of this Cr cutoff is 81% (87/108), and the specificity is 88% (578/657).

Using an even more conservative Cr cutoff of 1.0 mg/dL still failed to identify 10 patients at risk for CIN (9.3% [10/ 108]; 95% CI, 5%-16%) and would have prevented 200

patients (67%; 95% CI, 61%-72%) with CrCl higher than 60 mL/min from receiving contrast for their CT scan; the sensitivity of a Cr cutoff of 1.0 mg/dL for detecting a CrCl less than 60 mL/min was 91% (98/108), and the specificity was 70% (457/657).

Using a less conservative Cr cutoff of 1.8 mg/dL would have increased the number of patients with CrCl less than 60 mL/min who were exposed to IV contrast medium to 59 (55%; 95% CI, 45%-64%). Only 1 patient with a CrCl higher than 60 mL/min had a Cr higher than 1.8 mg/dL and would have been prevented from receiving IV contrast. The

Table 2 Patients safe and not safe for CT (as a percentage of all patients with Cr b respective Cr cutoff) Serum Cr Patient is safe for CT with IV contrast (CrCl N60 mL/min) Patient not safe for CT IV contrast (CrCl b60 mL/min)
	#	%		#	% of all patients brespective cutoff
b1.8	656 (656/715)	91.8		59 (59/715)	8.3
b1.5	647 (647/690)	93.8		43 (43/690)	6.2
b1.2	578 (578/599)	96.5		21 (21/599)	3.5
b1.0	457 (457/467)	97.9		10 (10/467)	2.1
Note: Safe for CT (IV contrast) is defined as having a CrCl N60 mL/min.

sensitivity at this Cr cutoff was 45% (49/108), and the specificity was 99.9% (656/657).

Discussion

The administration of IV contrast results in contrast- induced nephropathy in 8% to 14% of patients [1-4] with an unclear number of these patients ultimately requiring hemodialysis. With the increasingly frequent use of con- trast-enhanced CT scan in the ED for a more diverse set of medical conditions, it is important to reduce the number of patients at risk for CIN. Our data suggest that one possible method to reduce risk is to convert from use of serum Cr to use of calculated CrCl as a pretest screening tool.

Our data lend further support to prior studies that have revealed serum Cr can be an inaccurate assessment of renal function [7,13]. Duncan et al examined 2781 outpatients and found that of 2543 patients with a normal serum Cr, 387 had a calculated CrCl less than 50 mL/min. The use of serum Cr to estimate GFR is based on the assumption that Cr metabolism is in a steady state [15]. However, because of differences in age, body habitus, and the dynamic nature of renal function (particularly during illness), this assumption is often unreliable.

In turn, because the etiology of acute renal failure is multifactorial and the clinical and laboratory manifestations of ARF are delayed, ED practitioners may not recognize the subsequent deleterious effects of ED interventions, includ- ing the administration of IV contrast, nephrotoxic pharma- ceutical agents (antibiotics, analgesics), and inadequate intravascular volume resuscitation. Because there is an increased incidence of acute renal failure in patients with preexisting renal insufficiency (CrCl b60 mL/min) who are exposed to IV contrast material, better detection methods are necessary [16].

As far as outcomes are concerned, it is known that patients who develop CIN have a significantly increased hospital mortality [2-4,6]. In an article by Levy et al, 34% of patients with renal failure died, whereas only 7% of paired subjects without ARF died. Nash et al [2] similarly demonstrated very high mortality (38%) in patients that had serum Cr increases higher than 3.0 mg/dL as did McCullough et al [3], who also found that 36% of their patients who developed ARF died and, overall, had a 2-year survival rate that was 19%.

In addition, our data demonstrate that more patients are at risk for CIN than previously thought [6]. We found that in a cohort of patients presenting to a tertiary care ED for the evaluation and treatment of the chief complaint of abdominal pain, 6% of patients with a Cr less than 1.5 mg/dL had a CrCl less than 60 mL/min. Even at much lower Cr levels (1.0; 1.2), where most practitioners would not consider the possibility of an increased risk of CIN, a significant minority of patients

evaluated for the chief complaint of abdominal pain had a CrCl below 60 mL/min, including 10 (2.1%) patients with Cr less than 1.0 mg/dL and 21 patients (3.5%) with Cr less than 1.2 mg/dL.

Our study has several limitations. First, we used 60 mL/min as the minimum estimated GFR that is safe for the administration of contrast material. Other studies have used less conservative minimums (50 mL/min) [9]. Nonetheless, varying cutoffs would have resulted in similar disparities. Second, our study population consisted of patients presenting to the ED with abdominal pain; many of them had nausea and vomiting (63%) and poor oral intake, which may have resulted in an acute elevation in serum Cr due to dehydration and prerenal azotemia. However, we chose this patient population because patients with abdominal pain are the most likely to receive IV contrast in the ED. Third, our patient population was predominantly African American, and our results may not be generalizable to other patient populations. Fourth, although there is increasing evidence that serum Cr fails to identify patients with renal insufficiency, there is no universally accepted way to accurately measure GFR, especially in the acute care setting. Various investigators have, with variable results, evaluated the reliability of the Cockcroft-Gault equation in divergent populations of patients including those with known renal insufficiency and others with presumably normal renal function [9,11]. However, there are several experts who support the use of this method for calculating renal function because it is both accurate and easy to use in the clinical arena [1,7,9,15,17]. Another possible limitation is that patient-estimated weights may not accurately reflect actual lean body weight. If patients underestimate their weight, their calculated CrCl will be falsely low; if they overestimate their weight, the

calculated CrCls will be falsely high.

With regard to the accuracy of calculated CrCl using the Cockcroft-Gault equation for populations of differing body mass index, it has been demonstrated that this equation slightly underestimates GFR at very lean body weights and slightly overestimates GFR in overweight patients (up to 10 mL/min) [18-20]; however, if the more conser- vative value of 60 mL/min of calculated GFR is used (as we did in our study), this becomes a mute point. Of note, most institutions adjust radiation but not contrast dose for patients of different weight.

Finally, the purpose of this study was to identify what we feel is an extremely important discrepancy between the clinical marker that most clinicians use to assess renal function (GFR) and a method that more accurately identifies patients with renal insufficiency. We did not address outcomes in this study but plan to in the future.

Future studies should address whether Screening patients undergoing contrast-enhanced CT scans using calculated CrCl compared to pretest screening based upon serum Cr level decreases the incidence of CIN. In addition, future studies should investigate the use of prophylactic measures

(ie, IV fluid, N-Acetyl cysteine, and bicarbonate adminis- tration) in this subgroup of patients.

In summary, with regard to the sensitivity of using this criterion, we believe that varying the serum Cr cutoffs (used to determine whether a patient is safe for IV contrast administration) may unnecessarily preclude patients from receiving the more appropriate study. However, we demon- strated that there is marked disparity between plasma Cr and calculated CrCl, and that the most commonly used Cr cutoff (1.5 mg/dL) for contrast administration fails to identify a significant proportion of patients at risk for CIN. It therefore may be prudent to risk stratify patients based on CrCl rather than serum Cr.

References

1. Barrett BJ, Parfrey PS. Clinical practice. Preventing nephropathy induced by contrast medium. N Engl J Med 2006;354(4):379 – 86.
2. Nash K, Hafeez A, Hou S. Hospital-acquired renal insufficiency. Am J Kidney Dis 2002;39(5):930 – 6.
3. McCullough PA, Wolyn R, Rocher LL, et al. Acute renal failure after coronary intervention: incidence, risk factors, and relationship to mortality. Am J Med 1997;103(5):368 – 75.
4. Rudnick MR, Goldfarb S, Wexler L, et al. Nephrotoxicity of ionic and nonionic contrast media in 1196 patients: a randomized trial. The Iohexol Cooperative Study. Kidney Int 1995;47(1):254 – 61.
5. Merten GJ, Burgess WP, Gray LV, et al. Prevention of contrast- induced nephropathy with sodium bicarbonate: a randomized con- trolled trial. JAMA 2004;291(19):2328 – 34 [see comment].
6. Levy EM, Viscoli CM, Horwitz RI. The effect of acute renal failure on mortality. A cohort analysis. JAMA 1996;275(19):1489 – 94 [see comment].
7. Duncan L, Heathcote J, Djurdjev O, et al. Screening for renal disease using serum creatinine: who are we missing? Nephrol Dial Transplant 2001;16(5):1042 – 6.
8. Rule AD, Larson TS, Bergstralh EJ, et al. Using serum creatinine to estimate glomerular filtration rate: accuracy in good health and in

chronic kidney disease. Ann Intern Med 2004;141(12):929 – 37 [see comment, summary for patients in Ann Intern Med. 2004 Dec 21;141(12):I40; PMID: 15611485].

1. Ashley JB, Millward SF. Contrast agent-induced nephropathy: a simple way to identify patients with preexisting renal insufficiency. AJR Am J Roentgenol 2003;181(2):451 – 4.
2. Stevens LA, Coresh J, Greene T, et al. Assessing kidney function– measured and estimated glomerular filtration rate. N Engl J Med 2006;354(23):2473 – 83.
3. Bostom AG, Kronenberg F, Ritz E. Predictive performance of renal function equations for patients with chronic kidney disease and normal serum Creatinine levels. J Am Soc Nephrol 2002;13(8):2140 – 4.
4. Swedko PJ, Clark HD, Paramsothy K, et al. Serum creatinine is an inadequate screening test for renal failure in elderly patients. Arch Intern Med 2003;163(3):356 – 60 [see comment].
5. Levey AS, Bosch JP, Lewis JB, et al. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999;130(6):461 – 70 [see comment].
6. Mehran R, Aymong ED, Nikolsky E, et al. A simple risk score for prediction of contrast-induced nephropathy after percutaneous coro- nary intervention: development and initial validation. J Am Coll Cardiol 2004;44(7):1393 – 9.
7. Perrone RD, Madias NE, Levey AS. Serum creatinine as an index of renal function: new insights into old concepts. Clin Chem 1992;38(10):1933 – 53 [see comment].
8. Freeman RV, O’Donnell M, Share D, et al. Nephropathy requiring dialysis after percutaneous coronary intervention and the critical role of an adjusted contrast dose. Am J Cardiol 2002;90(10):1068 – 73.
9. Hu KT, Matayoshi A, Stevenson FT. Calculation of the estimated creatinine clearance in avoiding drug dosing errors in the older patient. Am J Med Sci 2001;322(3):133 – 6.
10. Verhave JC, Fesler P, Ribstein J, et al. Estimation of renal function in subjects with normal serum creatinine levels: influence of age and body mass index. Am J Kidney Dis 2005;46(2):233 – 41.
11. Froissart M, Rossert J, Jacquot C, et al. Predictive performance of the modification of diet in renal disease and Cockcroft-Gault equations for estimating renal function. J Am Soc Nephrol 2005;16(3):763 – 73.
12. Cirillo M, Anastasio P, De Santo NG. Relationship of gender, age, and body mass index to errors in predicted kidney function. Nephrol Dial Transplant 2005;20(9):1791 – 8.