BACKGROUND: Folate levels (routinely ordered for evaluation of macrocytosis with or without anemia) are poor predictors of true body stores, and low levels are only one of several causes of macrocytosis and anemia. Though better indicators of true body stores, erythrocyte folate levels may remain normal in early deficiency or may show false lows in chronic alcoholism, pregnancy, and cobalamin deficiency. Previous studies have led some to conclude that folate evaluations are ordered too often, and one report recommends empiric treatment with folic acid instead of folate testing for alcoholic patients with macrocytic anemia because of the low value of folate tests.
POPULATION STUDIED: The investigators evaluated the records of all patients who had a folate level drawn during hospitalization at 1 of 3 inpatient settings (an urban public teaching hospital, a private tertiary-care hospital, and a Department of Veterans Affairs [VA] teaching hospital) for a 1-year period (n=2998).
STUDY DESIGN AND VALIDITY: The study was a retrospective review of patient records that included a folate level determination. Several measures were recorded including total number of serum folate levels, the number of low serum folate levels (defined as from <1.5 ng/mL to <2.8 ng/mL among the 3 hospitals), erythrocyte folate levels (<95 ng/mL), anemia (hemoglobin <14 g/dL in men and <12g/dL in women), macrocytosis (mean corpuscular volume <99 fL), and microcytosis (mean corpuscular volume <80 fL). When the records revealed low folate levels, the medical records were reviewed to determine further patient characteristics and comorbid diagnoses. Reviewers determined if clinicians’ behavior was influenced by the low folate levels, including whether the low level was noted in a progress note or was an attempt made to determine cause and whether supplementation was given. A cost analysis was calculated using figures from 2 of the hospitals. Figures were not available from the VA hospital.Overall, the methods of this study were appropriate to answer the question. However, there are limitations with any retrospective study. For example, it is possible that clinicians took actions that were not documented in the medical record. Also, cost estimates may not be generalizable to private practice as they were taken from teaching hospitals where testing may be more prevalent.
OUTCOMES MEASURED: Reviewers determined the total number of low folate levels and clinical responses to those levels. Cost per abnormal test result was calculated and compared with the cost of empiric treatment with folic acid.
RESULTS: Only 68 of the 2998 (2.3%) folate measures were low. Low levels were recorded in 35 cases showing folate deficiency (53%); a possible cause was mentioned in 17 cases (24%); and folic acid was prescribed in only 16 cases (24%). Overall, the clinician response to folate level testing was less than 0.5%. This response rate results in a cost of $9979 spent on folate level determinations for each clinical response. In contrast, the cost of empiric treatment with folic acid would have been a maximum of $5.50 for a 90-day supply, totaling $6914 if all patients at the 2 hospitals included in the cost analysis had been treated (n=1257).
This study indicates that folic acid levels are rarely low in patients with macrocytosis, and low levels have little influence on therapy. We recommend checking only cobalamin levels in patients in whom both folate and cobalamin levels would be routinely ordered. If the cobalamin level is normal, treat empirically with folic acid. Reserve tests for folate deficiency for patients with persistent unexplained macrocytic anemia.