Critique: A Higher Dose of Vitamin D

Introduction

Broe et al. (2007) conducted a randomized study at the Hebrew Rehabilitation Center for Aged (HRCA) to determine whether higher doses of vitamin D supplements would reduce the risk of falls.  The study was conducted as a double-blind with a relatively small sample size of limited diversity.  The variables suffered some limitations, but the results were scientifically analyzed to determine that only the highest dose tested had any noticeable result on fall risks, as was supported by cited literature. While conducted with reasonable scientific reliability, the study was too limited in size and scope to return a significant result.

Protection of human participants

The basic premise of the study was to improve the health and safety of geriatric patients, presenting a clear benefit.  Since the study depended primarily on a measure of falls over a period of time, the risks were obvious, though not any different than the risk of someone outside the study. Measures were taken to exclude participants with previous medications or disorders that would significantly affect the vitamin D dose or the participants’ likelihood of falling (Broe et al., 2007, pgs. 234-235). The only evident risk not addressed by the study was the possibility of Vitamin D overdose, which can cause hypercalcemia and hypertension.  Evidence suggests that this only occurs at doses above those administered in the Broe et al. (2007) study, which was no more than 800 UI per day (pg. 234). It might be possible for the patients to have taken more outside the bounds of the experiment, however, and this risk should have been assessed.  Studies have shown that doses as low as 1,250 UI per day in geriatric subjects can cause hypercalcemia and hypertension, though generally these negative effects only occurred at doses over 10,000 IU (Hathcock, Shao, Vieth, & Heaney, 2007, pg. 11).

All proper consent and approval were acquired for this study.  The boards of both the HRCA and the Boston University School of Medicine were consulted about the protocol, both responded with approval.  The subjects of the study all gave informed consent (Broe et al., 2007, pg. 235).  There were no indicators that any of the participants were involuntary participants and since the study required the taking of pills about which they had been informed, it is reasonable to assume that the participants were voluntary.

Data collection

The variables of the experiment are clearly identified through context, but not explicitly labeled as dependent and independent.  There was a significant risk of data saturation with this collection. The only independent variable considered relevant by the study was vitamin D intake.  The dependent variable was the number of falls suffered by the subjects throughout the study period (Broe et al., 2007, pgs. 234-235). The independent variable was monitored by the HRCA pharmacy that provided the pills to ensure the double-blind status of the study. The incidences of falls were reported according to the HRCA computer database which was explicitly designed for tracking any deviation from the normal operation in the facility. The nursing staff were the ones responsible for actually reporting falls, as was the norm at HRCA regardless of the study (Broe et al., 2007, pg. 235). This method of data collection would have been highly susceptible to bias since there is no standard identified for what the nursing staff would consider a fall.

No rationale was given for the data collection methods used by this study, but the relatively simple structure of the study and the pre-existing system for tracking the dependent variable made the chosen methods appear logical. The study was conducted over the course of five months.  At the beginning of the study, the subjects’ age, gender, height, weight, and intake of multivitamins was gathered by consulting their charts. Each subject’s number of falls from the preceding year were gathered for comparison throughout the study period using the HRCA’s incident database and vitamin D serum concentrations were identified to determine the total supplement needed to achieve the various intakes of the independent variable (Broe et al., 2007, pg. 235). The only data collection conducted throughout the experiment for the purpose of the study was the number of falls.  When the study was completed, the vitamin D serum concentration was again tested (Broe et al., 2007, pg. 236).

Data management and analysis

The very limited scope of the variables made management and analysis of data simple.  The baseline characteristics of the subjects with continuous variables were compared with analysis of variance and those with categorical variables were analyzed using chi-square tests. The dependent variable was analyzed in two ways, first time fall and overall falls. For the first time fall results, Broe et al. (2007) used the Cox proportional hazards regression and the overall falls across the five month time period result was analyzed using the Poisson regression (pg. 235). No rigor of process considerations are mentioned explicitly. The only external verification of any of the gathered data mentioned would be the HRCA incident database, which is not made available.  There is no mention of multiple analyses to prevent bias in the results.

Findings / Interpretation of findings

The overall findings of the study were relatively inconclusive. Only one of the four doses of vitamin D had any noticeable effect and there was no observable trend of altered risk of falls in the lower doses.  Another source was cited, placing the recommended serum concentration between 28 and 32 ng/mL, which was only provided by the 800 IU dose of Vitamin D.  That dose resulted in an average serum concentration of 29.95 ng/mL.  It is suggested that serum levels within this range might also have a positive effect on muscle function, but that is unsupported (Broe et al., 2007, pg. 238). The findings were consistent with the previously cited recommendations about serum concentrations of vitamin D, but the range of tested doses was too limited to provide any results that could be interpreted more extensively. Previous research had recommended a dose as high as 1,000 IU (Holick, 2004, pg. 369). If this research had been considered, higher doses might have been tested in this study and yielded more meaningful results.

Limitations identified by the researchers were the uncontrolled variable of the multivitamins that some subjects were taking, but not all. Additionally, the inclusion of diets such as the Dash diet were noted as a variable. Broe et al. (2007) also acknowledged that the sample size was too small and not diverse enough to make the findings generalizable and the method for reporting falls may have been unreliable (pg. 238). The findings were, however, logically presented and were reliable enough to inspire confidence, within the bounds of the limitations.  The authors of the study did not explicitly recommend further research.

Conclusion

The authors of this study should have taken into account the previous recommendations made regarding the serum concentration of vitamin D before undertaking the study. Most of the results were insignificant because the serum levels resulting from the independent variable doses were below the recommended value.  While scientifically valid, the results were not meaningful beyond proving what was already known by experts. 

References

Broe, K. E., Chen, T. C., Weinberg, J., Bischoff-Ferrari, H. A., Holick, M. F., & Kiel, D. P. (2007). A higher dose of vitamin D reduces the risk of falls in nursing home residents: a randomized, multiple-dose study. Journal of the American Geriatrics Society, 55, 234-39. Retrieved from http://66.160.145.48/seaton/pdfs/27/23827663.pdf 

Hathcock, J. N., Shao, A., Vieth, R., & Heaney, R. (2007). Risk assessment for vitamin D. The American Journal of Clinical Nutrition, 85, 6-18. Retrieved from http://anaboliclabs.com/User/Document/Articles/Vitamin%20D/4.%20Hathcock,%20Vit%20D,%202007.pdf 

Holick, M. F. (2004). Vitamin D: Importance in the prevention of cancers, type 1 diabetes, heart disease, and osteoporosis. The American Journal of Clinical Nutrition, 79, 362-71. Retrieved from http://ajcn.nutrition.org/content/79/3/362.full.pdf