Quality and use of unlicensed vitamin D preparations in primary care in England: Retrospective review of national prescription data and laboratory analysis

To evaluate the type (licensed vs unlicensed) and cost of preparations used to fulfil vitamin D prescriptions in England over time, and to compare measured vitamin D content of selected vitamin D preparations against labelled claim.


| INTRODUCTION
Vitamin D has attracted considerable interest in recent years, not least because of increased awareness of the global prevalence of vitamin D deficiency. 1 Since 2016, the advice from the UK government is that individuals should consider taking daily vitamin D supplement of 400 international units (IU) during the winter months (from October to March), and all year round for anyone who spends limited time [Correction added on 3 September 2020, after first online publication: the middle initial of the fifth author has been inserted in this current version.] outdoors. 2 Vitamin D has also received wide media coverage arising from research linking vitamin D deficiency to an increased risk of myriad adverse health problems including diabetes, autoimmune conditions, cancers and infections, among others. 3 This heightened interest has, in turn, been accompanied by a significant rise in healthcare and consumer spending on vitamin D preparations. In England, medical prescriptions for vitamin D monotherapy preparations in primary care increased by more than 8000% between 2008 and 2013, with the latest annual spending reported at £38 million. 4,5 Likewise, consumer demand for vitamin D supplements has soared; 2018 retail data from the United Kingdom (UK) reported that vitamin D was the most popular vitamin monotherapy in this multimillion-pound industry. 6  However, vitamin D is also used clinically to treat or prevent vitamin D deficiency, and products making a medicinal claim would be subjected to the legislation governing medicinal products. [11][12][13] These legal distinctions have consequently led to the availability of two types of vitamin D preparations which are regulated by two different quality standards. Notably, the regulatory requirements for manufacturing, quality control and labelling which apply to medicinal products are far more stringent than those relating to food supplements. While medicinal products are subjected to premarket licensing assessment and hence are referred to as licensed medicinal products, food supplements are outside of the licensing process (referred to as unlicensed preparations) and are readily available to the general public without medical supervision through retail outlets including supermarkets, pharmacies, health food stores and online retailers.
Deciding whether a product is a medicinal product or a food supplement is dependent on a number of factors and the amount of an ingredient is not the only criterion. 11 In practice, the classification is sometimes ambiguous in that two apparently similar preparations are classified differently, but to the consumer can appear to be virtually indistinguishable. The extent to which these situations arise with vitamin D preparations is not known and is further complicated by concerns about product quality, with reports of poor quality food-grade vitamin D preparations on sale in New Zealand, the United States and India. [14][15][16] Questions over safety and legal issues associated with the use of food supplements to fulfil medical prescriptions of vitamin D in England has also been highlighted. 5 However, neither the quality of food-grade vitamin D preparations, nor the trends in supply of these preparations, have been investigated in England to date. Therefore, the aims of this work were to evaluate the type and costs of preparations used to fulfil vitamin D prescriptions in primary care in England over time, and to compare the measured vitamin D content of selected vitamin D preparations marketed in England against their labelled values.

| Licence status of vitamin D preparations
The National Health Service (NHS) Dictionary of Medicines and Devices (dm+d) via the NHS Technology Reference Data Update Distribution (TRUD) platform was used to identify all vitamin D preparations in use across primary care in England 16,17 ; both ergocalciferol (vitamin D 2 ) and colecalciferol (vitamin D 3 ) preparations were included. Combination preparations (eg, multivitamins, colecalciferol and calcium) were excluded in this study. For each preparation, we extracted data on its availability, dose strength, dosage form and type of vitamin D preparation, categorised as either licensed prescriptiononly medicines, licensed pharmacy medicines, licensed general sale list medicines, food supplements, "specials", "imports" or "undetermineds" (the legal definitions of these categories are provided in Supporting Information Method S1). • Little is known about the quality difference between licensed and unlicensed vitamin D preparations marketed in England.

What this study adds
• The supply of licensed vitamin D preparations against prescriptions has increased over time.
• While the number of licensed vitamin D preparations has increased, 40% of all vitamin D prescriptions dispensed in 2018 were still of unlicensed preparations.
• Vitamin D content of unlicensed vitamin D preparations varies widely from labelled claim. reimbursement data. 4 The PCA dataset complies with all aspects of the Code of Practice for Official Statistics and prescription processing activity has been internally audited to 99% accuracy. For each vitamin D presentation listed (further details on data processing of PCA dataset are provided in Supporting Information Method S2), we extracted data on the number of prescription items, unit quantity (ie, tablets or millilitres), net ingredient cost (NIC) and NIC per unit quantity (ie, cost per tablet). 4,17,18 The NIC is the basic cost of the preparation at list price, excluding value added tax. Due to commercial confidentiality, contract prices are not released outside of the NHS.

| Vitamin D content analysis
As ergocalciferol prescriptions accounted for less than 1% of the annual total vitamin D prescriptions since 2015, only colecalciferol preparations were included in the laboratory analysis. Eleven colecalciferol food supplements, covering a range of dose strengths and different dosage forms, were purchased from five different UK pharmacies, health food stores and online retailers; this represents approximately 6% of the vitamin D food supplements listed in the NHS dm +d. The preparations were selected based on a convenience sample reflecting food supplements that were available for retail purchase and used by the general public. We also included two licensed preparations that are routinely used in the NHS: one UK licensed preparation and one imported licensed preparation from Germany. The UK and Germany operate to the same manufacturing quality standards with regard to licensed medicinal products. All preparations were stored at controlled room temperature in cardboard boxes and protected from light following purchase. All products were within their expiry date at the time of analysis.

| Data analysis
The characteristics of vitamin D preparations used across primary care in England were summarised using descriptive statistics. For prescription trend analysis, vitamin D prescriptions dispensed each year were calculated and presented as percentage of total prescription items and total NIC by type of vitamin D preparation. Using 2018 data, the NIC reported for each unlicensed preparation (food supplements, "specials", or "imports") was compared with the estimated NIC that would For RP-HPLC analyses, peak areas were used for quantification of colecalciferol concentrations. The mean percentage difference from labelled content and standard deviation (SD) were calculated for all 13 preparations. An acceptable criterion of 90-125% of the stated amount was set in line with the British Pharmacopoeia, the official UK pharmaceutical standards for medicinal products. 20 We also compared our results against the food supplement standard, which has a wider acceptance range of 80-150% of the stated label amount. 21 The relationship between preparation dose strength and the magnitude of percentage difference between measured and labelled contents was assessed with Spearman's correlation coefficient.
Data analysis was conducted using Stata 15 (Stata Corp, College Station, Texas, USA).

| Ethical approval
As the study used routinely collected data that was nonidentifiable and publicly available, ethics approval was not required.

| Characteristics of vitamin D preparations
As of 9 September 2019, there was a total of 328 oral vitamin D preparations listed as available in the NHS dm+d; 35 (10.7%) were categorised as licensed prescription only medicines, 186 (56.7%) as food supplements, 102 (31.1%) as "specials" and 5 (1.5%) as "imports". A total of nine injectable formulations were also listed with three preparations each under the three categories of "licensed", "specials" and "imports".
The oral licensed medicinal preparations were available in a range of dose strengths (Table 1); tablets and capsules were available from 400 IU to 50 000 IU, and liquid preparations from 2 400 IU/mL to 50 000 IU/mL. Food supplements, presented as tablets or capsules, spanned the same range of dose strengths (400 IU to 50 000 IU) as the licensed medicinal preparations (Table 1), but the liquid formulations were available in a narrower range from 400 IU/mL to 10 000 IU/mL. Oral preparations categorised as "specials" covered the broadest range of dose strengths of all the preparations from 20 IU/mL to 600 000 IU/mL.

| Cost comparison between licensed and unlicensed vitamin D preparations
In 2018, the total cost of vitamin D prescriptions was £38.4 million ( Figure 3). Of that, £18.6 million was for tablets/capsules containing ≤800 IU of vitamin D, £15.9 million for tablets/capsules containing >800 IU of vitamin D, £3.7 million for oral solution and £0.1 million for injections.
A licensed equivalent preparation was identified for all 236 unlicensed preparations (food supplements, "specials", and

| DISCUSSION
Using national prescribing data, this study showed that patients pre- The substantial difference between measured and declared vitamin D content in food supplements shown in our study is in keeping with published data from other countries. [14][15][16]22,23  New Zealand, our results showed a narrower range of variability. 14,15 This difference could be a reflection of the differences in legislative requirements between countries, but as eight of the 11 food supplements tested in the present study failed to comply with national food quality standards, this seems unlikely. 21 Rather, the variability in vitamin D content in food supplements probably arises from differences in manufacturing and quality control practices between manufacturers, which in turn may reflect the general lack of regulatory oversight of food supplement manufacturers in these countries. 10  lists both medicinal and food-grade vitamin D preparations, which may inadvertently be misinterpreted by clinicians to mean that the preparations can be used interchangeably. However, without evidence that a particular brand is superior to another and along with 81% of primary care prescriptions in England prescribed generically, 29 it is unlikely that clinicians are deliberately choosing food-grade prep-

| Strengths and limitations of the study
We used the complete prescription reimbursement data set for all vitamin D prescriptions dispensed in primary care in England, not just a sample. Moreover, as our data set was based on prescription reimbursement and thus data interpretation was not affected by generic prescribing, we were able to provide a comprehensive analysis of the different vitamin D preparations used, providing real-world evidence of practice. We note that our analysis was based on data in England, but we expect similar prescription patterns across the UK given the similarities in primary care practice for prescribing and dispensing prescriptions.
We evaluated the vitamin D content of only a small sample of vitamin D preparations which may not be representative of all the preparations marketed in England and further comprehensive analysis is warranted. Nevertheless, the results we obtained were consistent with published literature. Our RP-HPLC method was verified as fit for purpose against the ICH standards in terms of accuracy, precision and linearity, and was used rather than the method stated in the British Pharmacopeia due to its ability to assess the vitamin D content in both food-and medicinal-grade preparations. We did not investigate the reasons for the variability in vitamin D contents using stability indicating methods because we were unsure of the grade of vitamin D originally employed in the food-grade preparations. We also did not investigate batch-tobatch variability in colecalciferol content of the different preparations, and the number of samples per preparation tested was less than that recommended by the British Pharmacopeia. However, our study conclusion would remain valid in the case of minimum variability while the presence of batch-to-batch/sample-to-sample variability would further highlight the problem of substandard quality with these food supplement preparations.

| CONCLUSIONS
The study showed that whilst food-and medicinal-grade vitamin D preparations were being used interchangeably to fulfil primary care vitamin D prescriptions, food-grade preparations, which are man-

DATA AVAILABILITY STATEMENT
The NHS Dictionary of Medicines and Devices and the NHS Prescription Cost Analysis dataset are publicly available via NHS Digital at https://isd.digital.nhs.uk/trud3/user/guest/group/0/home and https://digital.nhs.uk/data-and-information/publications/statistical/ prescription-cost-analysis, respectively. Openprescribing.net is also a free, openly accessible tool available at https://OpenPrescribing.net.