Vitamin D Toxicity: A 16-Year Retrospective Study at an Academic Medical Center

John P. Lee, DO; Michael Tansey, MD; Jennifer G. Jetton, MD; Matthew D. Krasowski, MD, PhD


Lab Med. 2018;49(2):123-129. 

In This Article


Consistent with the findings of previous studies,[3,15,16] we found that the vast majority of 25(OH)D test results at our academic medical center have shown concentrations below 80 ng per mL. Only 1.1% of patients tested in a 16-year period revealed 25(OH)D levels higher than 80 ng per mL, and only 0.1% had levels exceeding 120 ng per mL. However, 25(OH)D toxicity levels are not well defined.[12] Major clinical laboratories have varying cut-off points for excess levels.[13] A previous study[14] reported that a value of 125 ng per mL can be used as an upper limit of normal. In another study, toxicity was not observed until values exceeded 200 ng per mL. However, toxicity was mentioned in 1 patient report of a 25(OH)D serum/plasma concentration of only 80 ng per mL.[12]

We did not encounter such outcomes in patients until a 25(OH)D concentration of 194 ng per mL was achieved. Of the remaining patients experiencing symptoms, one had a level higher than 200 ng per mL, whereas the others had levels higher than 400 ng per mL. Further, similar to results reported by Dudenkov et al,[17] we discovered that potentially "toxic" levels of vitamin D did not strongly correlate with hypercalcemia. We limited our detailed medical-record review to cases in which the level of 25(OH)D exceeded 120 ng per mL. It is possible that some patients showed toxicity at lower 25(OH)D levels; however, our results align with those of other studies, such that these events appear to be rare.

The results of previous studies[3,12,15] have shown that most cases of potentially toxic 25(OH)D levels occurred due to use of high-dose vitamin D supplementation. Patients taking high-dose concentrations and liquid supplementation were at increased risk for higher 25(OH)D levels. Doses containing 50,000 IU were common in patients with high 25(OH)D levels in our study. This finding highlights the risk of someone overdosing or taking the incorrect amount with such concentrated doses.

We observed a higher rate of adverse outcomes associated with liquid vitamin D formulations. Specifically, we found 2 pediatric overdose cases in which an incorrect amount of supplement was dispensed from a dropper. Instead of receiving a drop of vitamin D, both patients received 1 entire dropperful due to confusion between the terms "drop" and "dropperful." Both misunderstandings resulted in these children receiving hundreds of thousands of IU or more of vitamin D during the course of a few months. Both patients showed symptoms and had elevated calcium levels and required hospital stays and treatment. Other case reports have also reported vitamin D toxicity resulting from liquid formulations.[8,18,19] In each case, the infant received droppersful instead of drops. In one case, a dropperful had been the correct dose on a previous formulation; however, the new supplement only required 1 drop. Our experience and the documented case reports highlight the danger associated with liquid formulations and the lack of standardized dosing in this area. Perhaps it is time for standardization of supplement formulation and simplification of administration methods, or other such measures, to enhance safety.

As a result of its retrospective nature, our study has certain limitations. First, the medical record review focused on common, recognizable symptoms of vitamin D toxicity. It is possible that more subtle toxicity was not evident on medical record review.

Second, we did not perform a detailed medical record review on patients who had vitamin D levels less than 120 ng per mL. It is possible that a patient with a 25(OH)D level lower than 120 ng per mL had symptoms. However, this possibility seems unlikely, based on the results of previous studies and the 25(OH)D levels at which we observed symptoms. Also, we did a search for patients with diagnosis codes related to vitamin D intoxication, overdose, poisoning, or toxicity. The results identified 7 patients with 1 or more of these diagnosis codes and who had 25(OH)D levels between 80 and 120 ng per mL. None of these patients had any symptoms consistent with vitamin D toxicity.

Third, a large cohort of patients did not have a calcium measurement performed at the time of vitamin D concentration. The study would have better power in terms of correlation between vitamin D and calcium levels with more concomitant testing. Also, vitamin D concentration and route were not listed in some of the medical records.

Last, a global challenge with 25(OH)D measurements has been the variation between assays and challenges with harmonization.[20–23] Although the criterion standard method of liquid chromatography–mass spectrometry (LC/MS) is increasingly being used for 25(OH)D measurement, immunoassays are still widely used. 25(OH)D immunoassays can vary from one another in cross-reactivity with 25(OH)D2 and 25(OH)D3, analytical measurement range, dilution protocols, and accuracy/precision. Variability between immunoassays may be especially large at very high 25(OH)D concentrations, thus making it challenging to compare patient results over time if different assays were used. This finding may be a factor in the present study, in which our institution changed 25(OH)D assays 3 times during the retrospective time period. Thus, we grouped 25(OH)D concentrations into broad ranges (eg, >80 ng/mL and >120 ng/mL) to minimize variations in the 25(OH) concentration interpretation. We hope that ongoing efforts to standardize and harmonize 25(OH)D assays will provide more consistency with measurement of potentially toxic 25(OD)D concentrations.[20–23]