89

IDENTIFICATION OF PATIENTS WITH CKD USING MEDICARE DATA: INPATIENT VS. OUTPATIENT CLAIMS. David T. Gilbertson1, Tricia Roberts1, Thomas Arneson1, Stephan Dunning1, Allan Collins1, Brian Bradbury2. 1Chronic Disease Research Group, Minneapolis, MN, USA. 2Amgen, Thousand Oaks, CA, USA. Use of Medicare claims for analyses of pts with chronic kidney disease (CKD) is common. A commonly used method to identify these pts requires ≥ 1 inpatient claim (IP) or ≥ 2 outpatient (OP) claims with CKD as a diagnosis code. Although this method was developed for identification of diabetes, it, and similar approaches have been used for identifying CKD and other comorbid conditions. It is relatively unknown, however, how analyses investigating outcomes in these pts are affected by the approach used to identify patients; in particular, the source of the diagnosis (IP vs. OP) code. We used the Medicare 5% sample for 2001 to identify pts with CKD using the method described above. We analyzed overall 1-year and interval mortality rates (0-3, 4-6, and 7-12 months after start of follow-up), by source of CKD diagnosis (IP vs. OP). The 1-year mortality rate for pts identified from an IP code was 434/ 1,000 person-yrs vs. 247/1,000 person-yrs for those identified from ≥ 2 OP codes. This difference was strongest during follow-up months 0-3 but persisted throughout the year (figure). Analyses of claims-identified CKD patients should take into account the source of the CKD code, because the hazard for mortality differs greatly based on source. The source of the CKD diagnosis code may be correlated with treatment exposures of interest in a statistical model, and failure to account for this correlation may lead to spurious results.

020

040

060

080

0

767

373 371

221255

185

Months 0-3 Months 4-6 Months 7-12

IPOP

Mortality Rate per 1,000 Person-Years

Months After Start of Follow-up

By Source of CKD Diagnosis Code

90

THE IMPACT ON THE US BLOOD SUPPLY OF BLOOD TRANSFUSIONS IN DIALYSIS Gitlin M, Hollenbeak C, Mayne TJ, Amgen Inc., Thousand Oaks, CA; Pennsylvania State University College of Medicine, Hershey, PA. From 1989 to 2004 there has been a 53% reduction in the available US blood supply (surplus over demand), from 14% to 4.4% of collected blood (GAO Report, 1999). During this same time, the percentage of patients transfused has decreased 50%, coinciding with the increased use of erythropoiesis stimulating agents (ESAs). A model was developed to evaluate the blood transfusion impact as a result of changes to mean population hemoglobin (Hb) levels among US dialysis patients. Baseline model inputs include: 1) The 2004 available US blood supply of 648k units or 4.4% (NBCUS, 2005); 2) US dialysis patient census of 350k (USRDS ADR, 2006); 3) relative change in % of patients transfused associated with a mean population Hb (Mayne, 2007), and 4) the annual number of units administered per transfused dialysis patient (Crowley, 1987; Goodnough, 1994). Sensitivity analyses were run varying estimates of the baseline inputs. Figure 1 displays the results of the model of dialysis blood utilization based on a changing mean population Hb range of 9 g/dL to 12 g/dL. A mean Hb = 12 g/dL equates to ~97k unit demand (~97k dialysis patients receiving ~ 1 unit of blood per person) while a mean Hb = 9 g/dL equates to ~1.8M unit demand (~260k dialysis patients receiving ~7 units of blood per person). Figure 1.

0

400,000

800,000

1,200,000

1,600,000

2,000,000

9.0 9.5 10.0 10.5 11.0 11.5 12.0Mean Population Hb (g/dL)

Uni

ts o

f B

lood

per

Yea

r

The model suggests that, in the absence of increased supply, risks of blood shortages would increase markedly if Hb levels decreased much below 11g/dL.

Demand Exceeds Supply

Current DemandAvailable Supply

= 648,000

91

KNOCKOUT OF THE GLUCOCORTICOID RECEPTOR IN THE DISTAL NEPHRON DOES NOT AFFECT DEXAMETHASONE-INDUCED HYPERTENSION Julie Goodwin1, Junhui Zhang2, David Geller2 1Section of Nephrology, Department of Pediatrics, Yale University School of Medicine 2Section of Nephrology, Department of Medicine, Yale University School of Medicine Glucocorticoids (GCs) are used as a treatment for a variety of conditions and hypertension (HTN) is a well-recognized side effect of their use. The mechanism of GC-induced HTN is incompletely understood and has traditionally been attributed to promiscuous activation of the mineralocorticoid receptor by cortisol. Multiple lines of evidence, however, point to the glucocorticoid receptor (GR) as an important mediator as well. We have developed a mouse model of GC-induced HTN which we shave shown is dependent on GR. To determine the site(s) of GR action relevant to the development of HTN, we studied GC-induced HTN in a mouse with a tissue-specific knockout of GR in the distal nephron. Interestingly, although these knockout mice had similar body weight, nephron number and renal histology compared to littermate controls, their baseline blood pressure was mildly elevated (111.4 ± 1.6 mm Hg vs. 103.8 ± 2.2 mm Hg, p=0.032). Nevertheless, both distal nephron GR knockout mice and control mice had a similar hypertensive response to dexamethasone, both on a normal diet and on a low sodium diet. Urinary excretion of electrolytes, both before and after administration of GC, was also indistinguishable between the two groups. We conclude that GR in the distal nephron is not necessary for the development or maintenance of dexamethasone-induced HTN in our model.

92

TREATMENT OF SARCOID GRANULOMATOUS INTERSTITIAL NEPHRITIS WITH ADALIMUMAB roopali gupta, jack moore jr, lisa beaudet, tulsi k mehta Deparments of Nephrology and Pathology,Washington Hospital Center, Washington DC, USA Sarcoidosis is a systemic disease with multiorgan involvement which causes renal failure through several different mechanisms. Granulomatous interstitial nephritis is an important, albeit less frequent cause of clinically significant renal disease. Here, we present the case of a 46 year old woman with a history of sarcoidosis, steroid-induced diabetes mellitus, hypertension, and chronic kidney disease evaluated for rapidly worsening kidney function over several months. Her creatinine increased from a baseline of 1.8 mg/dL to 3.9 mg/dL with as much as 10 grams of proteinuria per day. Renal biopsy revealed granulomatous interstitial nephritis superimposed on chronic diabetic nephropathy. To avoid additional steroid toxicity, she was treated with subcutaneous adalimumab, a recombinant human TNF-alpha inhibitor. After 6 months of therapy her renal function steadily improved back to her baseline function with a reduction of proteinuria to 3.5 grams per day. Follow-up biopsy demonstrated a complete resolution of renal granulomata. To our knowledge, this is the first report of successful treatment of granulomatous interstitial nephritis with adalimumab.

NKF 2008 Spring Clinical Meetings AbstractsA50