Abstract Introduction Varied presentations of vitamin B12 (B12) deficiency include hematologic (megaloblastic anemia) and neurologic (myelopathy, peripheral neuropathy). Delayed recognition and treatment may have serious consequences. Assessing B12 status can be challenging due to diverse laboratory assays and thresholds, resulting in a wide range of reported prevalence of B12 deficiency from 3-26%. Comorbidities may also spuriously elevate or suppress B12 levels. Methylmalonic acid (MMA) level, often used as a companion test to assess for B12 deficiency, may also be affected by differing laboratory thresholds and other medical comorbidities, such as renal dysfunction. Our institutional approach to B12 assessment, termed the pernicious anemia (PA) cascade, consists of reflexive testing to 1) document low B12 stores (B12 levels with or without MMA), and 2) diagnose pernicious anemia (intrinsic factor binding antibody [IFBA] with or without gastrin assay). In this study, our objectives were: (1) assess the usefulness of our B12 testing algorithm and associated laboratory thresholds for identifying B12 deficiency, and (2) evaluate clinical follow-up of test results as part of a quality project. Methods In this retrospective study, adult (age ≥18 years) inpatients who had B12 testing from 2007 to 2013 were identified. Medical records for those patients with B12 levels 150-400 ng/L were reviewed. Data abstracted included patient demographics, indication for B12 testing, B12 levels, results for MMA, IFBA, gastrin, serum creatinine, complete blood count, and clinical outpatient follow-up. The PA cascade algorithm is as follows: 1) for B12 <150 ng/L, IFBA is performed; 2) if IFBA is negative or indeterminate, then gastrin is performed; 3) for B12 150-400 ng/L, MMA is performed; 4) if MMA is >0.40 nmol/mL, then IFBA is performed. Reference range for B12 is 180-914 ng/L. Patients were determined to be B12 deficient if B12 level was <180 ng/L, or if B12 was 150-400 ng/L with elevated MMA and macrocytosis in the absence of significant renal dysfunction, defined as serum creatinine ≥2.0 mg/dL. Results Records of 326 patients (150 female) were reviewed. Indications for B12 testing were: hematologic, 175 patients (53.7%); neurologic, 127 patients (39.0 %), psychiatric, 18 patients (5.5%); other, 4 patients (1.2%). Fifty-nine of 326 (18.1%) patients were B12 deficient, of whom 17 of 59 (28.8%) had PA. There were 54 cases of elevated MMA, >0.40 nmol/mL, of which 30 cases had mildly elevated MMA, >0.40 nmol/mL and ≤0.60 nmol/mL, considered to be the MMA "gray zone". Thirty-two of the total 326 patients had serum creatinine ≥2.0 mg/dL, of which 23 patients also had elevated MMA. Elevated MMA in 13 of these 23 patients was more likely due to renal dysfunction than to B12 deficiency. B12 testing results were not mentioned in hospital discharge summaries for 173 patients, including 8 patients with B12 deficiency. For 22 B12-deficient patients, there was no documentation of B12 replacement. Conclusions The PA cascade detected B12 deficiency in a significant number of inpatients with low-normal B12 levels that might have otherwise been missed; 28.8% of this group had PA for whom long-term B12 replacement is indicated. As treatment for B12 deficiency is inexpensive, non-toxic, and non-invasive, it is reasonable to evaluate B12 status by cascade testing which may provide additional diagnostic data beyond B12 level. Notable clinical practice findings from this study include the lack of B12 results in 173 of 326 discharge summaries, and no record of B12 supplementation in 22 of 59 patients with B12 deficiency. These gaps in documentation and follow-up may be associated with the relatively long turnaround time of MMA testing. Efforts at improving documentation, communication, and treatment of B12 deficiency are ongoing. Disclosures No relevant conflicts of interest to declare.