BACKGROUND

Pharmaceutical compounding is the combining, mixing, or altering of ingredients to create a customized medication for an individual patient in response to a licensed practitioner's prescription. Compounding does not generally include mixing or reconstituting commercial products in accordance with the manufacturer's instructions or the product's approved labeling. The quality of a finished compounded drug product can be affected by numerous factors including the quality of the active pharmaceutical ingredient used and the compounding practices of the pharmacy in which the product is created.

Since 1990, FDA has become aware of more than 55 product quality problems associated with compounded products, many of which resulted in product recalls. In 2001, FDA's Division of Prescription Drug Compliance and Surveillance conducted a limited survey of drugs compounded by a group of community pharmacies located throughout the United States. The goal of the survey was to gather information on the quality, purity, and potency of compounded drug products in the marketplace. The compounded products surveyed were selected from a cross-section of commonly compounded dosage forms based on FDA's assessment of the potential health risks resulting from improper compounding.

METHODOLOGY

The survey was conducted from June to December 2001. Samples of the products to be analyzed were collected from 12 compounding pharmacies that allowed specific compounded products to be ordered over the Internet.

A. Products Tested

Thirty-seven products made by 12 compounding pharmacies were identified for sampling and analysis. The types of drug products sampled included hormonal products, antibiotics, steroids, anesthetics, and drugs to treat glaucoma, asthma, iron deficiency anemia, and erectile dysfunction. Five different dosage forms (i.e., sterile injectables, ophthalmic products, pellet implants, inhalation products, and oral dosage forms) were sampled.

Of the 37 products identified, samples of 29 were collected and subjected to original and repeat analytical testing during the survey. The remaining eight products identified for sampling were either not collected or did not undergo original and repeat analyses for one of the following reasons:

• Commercial versions of the products were provided in place of a compounded product (n = 4)
• The compounding pharmacy was unable to provide the requested compounded product or to provide the product in an amount sufficient for complete analysis (n = 4).

Of the 29 products collected and analyzed, 13 were sterile injectable products, 9 were ophthalmic products, 2 were pellet implants, 1 was an inhalation product, and 4 were oral products.

B. Sample Collection

FDA collected the samples via mail order, in the same manner a consumer would order the products over the Internet. The samples were sent to an FDA laboratory for analytical testing within an average of 2.5 days of collection (range = 1 - 7 days)¹.  With the exception of four products (three progesterone capsules and one progesterone injectable) for which specific storage information was not available, all samples were stored at the FDA laboratory under conditions appropriate for each product. Testing was initiated on sampled products within an average of 7.7 days of receipt (range = 0 - 46 days) at the FDA laboratory and was completed within an average of 41.3 days of receipt (range = 9 - 95 days) at the FDA laboratory. Analytical testing was completed prior to expiry for all samples that contained expiration dates on their labels. Five of the sampled products collected did not contain expiration dates on their label. Three of these 5 products passed and 2 failed analytical testing.

 ¹ It should be noted that one Iron Dextran product was collected on November 16, 2000 prior to initiation of the survey. Once the survey commenced, a decision was made to include this sample in those analyzed for the survey. The FDA laboratory conducted analytical testing on this sample within the expiration date per the product’s label. Data for this sample on time between collection and time to initiation and completion of analyses was not included in the calculations above. This sample passed all analytical tests performed.

C. Testing

Analytical tests performed on these products were specific to their dosage forms and applicable testing methodologies (either United States Pharmacopeia (USP)/ National Formulary (NF) methods or new drug application (NDA)/abbreviated new drug application (ANDA) methodology). Table 1 summarizes the tests performed for each dosage form category sampled during the survey.

Table 1: Analytical Tests Performed based upon Dosage Form of Compounded Product Sampled

Analytical Test Performed	Dosage Form Sampled
	Sterile Injectable	Pellet Implant	Solution or Ointment for Ophthalmic Administration	Inhalation Products	Oral Dosage Forms
Sterility	X	X	X
Identity	X	X	X	X	X
Assay (potency)	X	X	X	X	X
Endotoxin (LAL)	X
Particulates	X
Release rate		X
pH			X
Microbial limits				X
Contaminants				X
Dissolution					X
Content uniformity					X

For chemistry tests, original and check (repeat) analyses were performed for each sample, with the repeat analyses conducted by laboratory personnel different from those who performed the original analyses. Sample failure was based on the type of analytical tests performed in this survey (i.e., chemistry tests and microbiological tests). A sample was deemed to have failed a specific chemistry test if it failed both the original and the repeat analyses. Chemistry tests performed on the samples collected were identity, assay, content uniformity, dissolution, and pH tests. Microbiological tests performed on the samples collected were sterility, Limulus Ameobocyte Lysate (LAL), microbial limits, and particulates tests. Repeat testing was not required for microbiological analyses. A sample was deemed to have failed a specific microbiological test if it failed an original analysis.

Samples were assayed for sterility using the USP method (USP 24, pp. 1818-1824). All but three of the samples were assayed for potency using the appropriate USP Monograph (USP 24). Because there was no applicable USP monograph, the three other samples (two injectable products and one inhalation product) were assayed for potency in accordance with methodology contained in an ANDA or NDA for the product, respectively. An additional product (i.e., Alprostadil injection) was assayed for potency in accordance with the Pharmacopeial Forum (PF), Vol. 27(3) May-June 2001, pp. 2509, 2512-14, and another product (i.e., TriMix injection containing Papaverine, Phentolamine, and Prostaglandin E1) was assayed for potency via a method modified from that of Tu et. al. (American Journal of Hospital Pharmacy, 1987; 44:2524-2527) and validated by the FDA laboratory. This latter method was utilized because of the nature of the specific product and the fact that there was no applicable USP monograph or NDA/ANDA methodology.

The microbial limits assay was performed according to USP 24, Chapter <61>, pp. 1814-1818 and USP 24, Chapter 71, pp. 1820, 1822. The LAL test for endotoxin was conducted using the procedure described in a USP 24, Chapter 85, pp. 1829-1231. Particulate matter was determined in accordance with USP 24, Chapter 788, pp. 1971-1977, and uniformity of dosage units (UDU) analyses were performed according to USP 24, Chapter 905, 2^nd Supplement, pp. 2904-2906.

D. Results

Ten (34%) of the 29 sampled products failed one or more standard quality tests performed. Nine of the ten products with failing analytical results failed assay or potency testing. All of the sampled products that failed potency analyses had subpotent results, indicating that the products analyzed contained less of the active ingredient(s) than expected (as described in the product's label). The average percent of declared potency for these nine products was calculated from the original and repeat analyses performed for each sample, with a range of 59 percent to 89 percent of expected potency. In addition to the potency failures described, two other analytical test failures were noted during the survey (i.e., a failed LAL test for an injectable product and a content uniformity failure for an oral product that also failed potency testing).

None of the compounded products analyzed in this survey failed identity testing. In addition, none of the compounded products sampled and subjected to sterility testing (sterile injectables, pellet implants and ophthalmic products) or testing for microbial limits (the single inhalation product) failed these analytical tests. Appendix A lists the specific compounded products sampled for the survey, the expected concentration of active ingredient as listed on the product's label, a notation for the compounding pharmacy from which the product was obtained, and the type of analytical test failed.

CONCLUSIONS

Each year, FDA routinely samples drug products made by commercial manufacturers and analyzes these samples in FDA laboratories. More than 3,000 drug products from commercial manufacturers have been sampled and analyzed by FDA since fiscal year 1996. The analytical testing failure rate for commercially produced samples has been less than 2 percent. When compared to this failure rate, the percentage of sampled compounded products failing analytical testing in this survey (i.e., 34 percent) was higher than expected.

This survey had several limitations including a small sample size, the inability to collect and complete original and repeat analyses on all product samples originally identified for the survey, and the fact that the compounding pharmacies selected for the survey were limited to those permitting Internet purchase of the drug products chosen for sampling. Despite these limitations, this survey provided valuable, preliminary information on the quality of selected compounded drug products currently marketed. The findings also highlight the importance of carefully and properly compounding drug products to minimize risks to consumers.

Appendix A: Specific Compounded Products Sampled in FDA's Survey and List of Failed Analytical Test Results

Compounded Product	Concentration (per label)	Pharmacy	Type of Analytical Test Failed
Sterile Injectables
Alprostadil	20 mcg/ml	H	LAL
Betamethasone	3 mg/ml	A	Assay/Potency
Dipyridamole	5 mg/ml	E	None
Dipyridamole	5 mg/ml	F	None
Droperidol	5 mg/ml	A	None
Hyaluronidase	150 U/ml	D	None
Hyaluronidase	150 U/ml	K	Assay/Potency
Iron Dextran	50 mg/ml	F	None
Iron Dextran	50 mg/ml	F	None
Papaverine HCl	30 mg/ml	A	Assay/Potency
Progesterone	50 mg/ml	D	Assay/Potency
Progesterone	50 mg/ml	L	None
TriMix - Papaverine - Phentolamine - PGE1	17.65 mg/ml 0.59 mg/ml 5.9 mcg/ml	I	None
Pellet Implants
Estradiol	25 mg pellet	D	None
Estradiol	25 mg pellet	I	None
Ophthalmic Solutions and Ointments
Atropine SO4	1% Preservative Free (PF) solution	G	None
Atropine SO4	1% PF solution	J	None
Betaxolol	0.125% PF solution	G	None
Carbacol	1% PF solution	G	None
Ciprofloxacin	0.3% PF solution	G	Assay/Potency
Ciprofloxacin	0.3% PF solution	J	None
Dexamethasone	0.05% PF ointment	G	None
Dexamethasone	0.1% PF solution	J	Assay/Potency
Timolol Maleate	0.25% PF solution	G	Assay/Potency
Inhalation
Tobramycin	60 mg/ml	E	Assay/Potency
Oral
Progesterone	200 mg Sustained Release capsule	I	None
Estradiol	2 mg tablet	B	None
Progesterone	300 mg capsule	I	Assay/Potency Content Uniformity
Progesterone	100 mg capsule Micronized in Oil	M	None