High Vitamin B ₁₂ Level: A Strong Predictor of Mortality in Elderly Inpatients

To the Editor: Cobalamin levels are frequently examined in geriatric practice. The focus is usually on detecting vitamin B12 deficiency, as shown by extensive literature.1 In contrast, little is known about the meaning of high cobalamin levels.2 Previous data indicate that an elevated serum vitamin B12 level is a predictive factor for mortality in elderly patients with cancer.3 It therefore seems that elevated levels of serum cobalamin may be a sign of a serious, even life-limiting, illness. Elevated cobalamin levels often accompany diseases such as leukemia, myeloproliferative disorders, hypereosinophilic syndrome, and metastatic cancer—especially when involving the liver.4 An upregulation of transcobalamin synthesis, increased release of cellular cobalamin, or decreased clearance of cobalamin from plasma may explain mechanisms of high vitamin B12 level.4 In daily geriatric practice, (biochemical) predictors that help to estimate mean survival are of great interest in that they might possibly prevent futile hospitalizations and overdiagnosis and subsequent overtreatment. For 3 months, we assessed vitamin B12 level in 488 consecutive admissions (138 men, 350 women; average age 84.7, range 65–101) to the Geriatric University Hospital in Geneva. Laboratory screening of vitamin B12 was routinely performed the day after admission, using the chemiluminescent immunoassay system technique (IMMULITE®2000 Vitamin B12, L2KVB6, Diagnostic Product Corporation, Los Angeles, CA, http://www.dpcweb.com). Two years after B12 measurement, the state population office registry of Geneva was consulted to obtain patients' date of death. Median serum vitamin B12 level was 298 pmol/L (interquartile range=285 pmol/L), and 33.8% of the patients had high vitamin B12 levels (≥400 pmol/L). The best discriminating cobalamin threshold level (400 pmol/L) was determined using a sensitivity analysis after applying a logistic regression model. One hundred fifty-eight patients died during the 2-year period; we observed a statistically significant mortality increase in the group with high serum vitamin B12 (log rank test P<.001). A Cox regression model showed that the risk of dying was 1.74 times as high (95% confidence interval (CI)=1.25–2.41, P=.001). Adjustment for sex (hazard ratio (HR)=1.22, P=.24), age (HR=1.01, P=.56), and the presence of a cancer (HR=1.85, P=.02) did not significantly influence this finding. To confirm our hypothesis that high vitamin B12 levels may be a biomarker for mortality risk, we focused on a follow-up period of 6 months. Survival for all patients according to B12 levels started to differ significantly at Day 50 (log-rank test, P<.001). Mortality rate at Days 50, 60, and 90 were 3.42% (95% CI=1.91–6.08), 3.73% (95% CI=2.13–6.47), and 4.66% (95% CI=2.84–7.61), respectively, in the group with lower B12 levels. These values were to 9.84% (95% CI=6.15–15.56), 11.69% (95% CI=7.62–17.72), and 17.25% (95% CI=12.25–24.00) in the group with high vitamin B12 levels. Our group has already documented greater mortality rates in cancer patients with high vitamin B12 levels.3 Here, we have now also analyzed patients without cancer (solid tumors or malignant hemopathy, n=447). Figure 1 shows that survival curves according to B12 level start to differ significantly at Day 60 in this group. Mortality rate at Days 60 and 90 were 3.0% (95% CI=1.57–5.69) and 4.0% (95% CI=2.29–6.94) in the group with lower B12 levels, respectively, versus 10.4% (95% CI=6.39–16.63) and 15.9 (95%=CI 10.89–23.00) in the group with high B12 levels. Kaplan-Meyer survival curves according to prospectively analyzed cobalamin levels in 447 consecutive inpatients without cancer. The vertical line crossing the x-axis at Day 60 represents the time from which the two curves are statistically different (log-rank test, P=.003). Theses results clearly demonstrate that supranormal serum cobalamin levels (>400 pmol/L) occur in one-third of elderly hospitalized patients. Furthermore, high B12 levels are associated with a 4.48 times greater risk (95% CI=2.14–9.89) of mortality in noncancer patients within the following 90 days. Cobalamin level may therefore be an essential parameter that should be included in severity-of-illness measures in all elderly inpatients, helping in the decision-making for adequate care processes in this mainly frail elderly population even in the absence of cancer. Further research is needed to define mechanisms of high B12 levels found in this older population.