demonstrated that diagnostic blood loss alone does not explain declining hemoglobin levels, therefore more robust models are needed to understand the pathogenesis of anemia among adult critical care patients. Objectives: To predict the onset of iatrogenic anemia using an objective mathematical model that integrates the rates of phlebotomy, erythropoiesis and red cell senescence with patient characteristics. Results: The 70 g/L hemoglobin transfusion threshold was reached in 40–70 days using initial parameters of gender, average body weight (average total blood volume), 53.2 mL/day phlebotomy, active erythropoiesis and initial hemoglobin levels at mid-reference range. To mimic critical care patients an initial hemoglobin of 110g/L and suppressed hematopoiesis was evaluated with variable daily blood losses and total blood volumes. Combinations of the lack of erythropoiesis, low initial hemoglobin concentration, low total blood volume and increased daily phlebotomy accelerated the onset of the 70 g/L transfusion threshold within 9–14 days. Conclusions: This model objectively depicts how diagnostic phlebotomy influences the onset of anemia in subsets of adult critical care patients. There is potential that these subsets of patients could benefit from more conservative test ordering practices and subse- quent reduced phlebotomy and fewer transfusions. doi:10.1016/j.clinbiochem.2011.06.020 P510 Pediatric reference intervals for 14 serum chemistries: The caliper study D.A. Colantonio a,b , L. Kyriakopoulou a,b , M.K. Chan a , A. Venner b , M. Diamandis b , D. Brinc b , L. Fu b,c , P. Yip b,d , D. Armbruster e , K. Adeli a,b a Clinical Biochemistry Division, Department of Pediatric Laboratory Medicine, Hospital for Sick Children, Canada b Department of Laboratory Medicine and Pathobiology, University of Toronto, Canada c Clinical Biochemistry, Sunnybrook Health Sciences Centre, Toronto, Canada d Department of Clinical Biochemistry, University Health Network, Toronto, Canada e Abbott Diagnostics, Abbott Park, Chicago, IL, USA Objectives: The values of many analytes measured in the laboratory change throughout childhood development and should be reflected in pediatric reference intervals. Updated reference intervals based on a healthy pediatric population are not available for routine chemistries and emerging biomarkers, which can lead to misdiagnosis and unnecessary treatment. CALIPER is a multi-centre study among children's hospitals across Canada with the goal to establish current age- and gender-specific pediatric reference intervals. Methods: This study followed the CLSI C28-A3 guidelines. Samples were collected and analyzed from ethnically diverse healthy children ages birth to 18 years of age. A minimum of 120 samples were collected for each age partition investigated. Albumin [bromo- cresol green (BCG) or purple (BCP) method], total/direct bilirubin, calcium, carbon dioxide, creatinine (picrate Jaffe or enzymatic method), magnesium, phosphorus, total protein, iron, urea and uric acid were measured on the Abbott ARCHITECT c8000 system. Outliers and results affected by common interferences were removed. Age- and sex-specific partitioning was determined for each analyte individually. Parametric and non-parametric methods were used to establish the 2.5th and 97.5th percentiles for the reference intervals. The 90% confidence interval for the limits was calculated. Results: Albumin, bilirubin, calcium, carbon dioxide, magnesium, phosphorus, total protein and urea required partitioning by age only where as uric acid, creatinine and iron required both sex and age partitioning. Reference intervals for each analyte will be presented. Conclusions: This data demonstrates the need for updated pediatric reference intervals based on a large diverse healthy cohort that are age and sex specific. These results will contribute to improved assessment and management of children. doi:10.1016/j.clinbiochem.2011.06.021 P511 Evaluation of an automated method for urine albumin: Creatinine ratio in a pediatric and obstetric population D. De Silva a , A.C. Halstead a,b a University of British Columbia, Vancouver, BC, Canada b Children & Women's Hospital, Vancouver, BC, Canada Objective: To evaluate an automated method for urine albumin: creatinine ratio (ACR) (Vitros Fusion, Ortho-Clinical Diagnostics) compared to a point-of-care method (DCA Vantage, Siemens Healthcare Diagnostics) used in the laboratory. Methods: Routine random urine samples from adult patients and diabetic children were analyzed by both methods on the same day. Samples collected for evaluation of proteinuria in pregnancy were also analyzed by both methods. Between day precision was assessed using BioRad Liquichek and BioRad Lyphochek controls for the Vitros and DCA methods respectively. Results: Coefficients of variation for albumin on the Vitros and DCA were 3.9% and 9.0% respectively at a level of 10mg/L and 3.1% and 4.2% at a level of 55 mg/L. Routine sample comparisons showed: n Range of results Regression equation R 2 Mean difference Mean % difference Albumin 37 6.8–240 mg/L y = 1.02x -0.70 0.98 -0.21 -9.8 Creatinine 37 1.41–34.4 mmol/L y = 1.12x -1.32 0.98 0.35 0.9 Overall ACR 37 0.29–13.8 mg/mmol y = 1.08x -0.21 0.98 0.06 -10.48 Three samples had significantly lower urine albumin on the Vitros (percentage differences: 57–75%). Analysis by recovery did not suggest any interference. Crossovers using urine from pregnant women showed a positive bias compared to non-pregnant patients [ACR regression equation: y = 1.44x -0.82 (R 2 = 0.998)]. Conclusion: The Vitros method agrees well with the DCA, with better precision at low levels. Decision limits for “microalbuminuria” in non-pregnant patients are comparable. Further study is required to develop decision limits for albuminuria in pregnancy. doi:10.1016/j.clinbiochem.2011.06.022 P512 Effect of one week of proton pump inhibitor therapy on serum and plasma chromogranin A concentrations H.H. Mosli b , A.W. Dennis a , W. Kocha b , J.L. Asher b , S. Sharif a , S. van Uum b a London Laboratory Services Group, Canada b London Health Science Centre, Department of Medicine, Canada Background: Chromogranin A (CgA) is a sensitive marker of neuroendocrine neoplasms, and is used for diagnosis, prognosis and for monitoring of response to therapy. Use of Proton Pump Inhibitors (PPIs) results in increased production of CgA. Except for some case reports, there is no information on the extent to which PPIs increase CgA, or on the duration of the effect after discon- tinuation of PPIs. Methods: 14 healthy volunteers, ages 18–70, were given bedtime lansoprazole 30 mg for 7 days. Fasting serum samples for CgA were Abstracts 1169