This article has generated considerable controversy in the medical and in the general press, because of attempts to prevent its publication by pharmaceutical industry sponsors.
Apart from this controversy, the topic of brand/generic equivalence is an important one.
The study compared four different levothyroxine formulations, two brand and two generic products:
Levoxine, manufactured by Daniels Pharmaceuticals (now Levoxyl, manufactured by Jones Medical Industries).
Two generic preparations, both manufactured by Pharmaceutical Basics under the labels Geneva Generics and Rugby Laboratories.
Patients previously stable on l-thyroxine replacement (either at a dose of 0.100 mg or 0.150 mg) were given each of the four preparations for at least 6 weeks, and biochemical measurements were obtained in order to compare the products. Furthermore, the four products were subjected to biochemical assays.
24 subjects were studied. All were hypothyroid, treated for at least 3 months, with at least two sets of normal thyroid function tests prior to the study. Subjects also underwent a history, physical examination, ECG and routine labwork and were excluded if they had unstable cardiac disease, anemia or any condition or treatment that would interfere with thyroxine absorption, metabolism or biochemical assays.
Subjects were randomized to one of four different drug sequences: A/B/C/D, B/D/A/C, C/A/D/B, D/C/B/A (the letters A, B, C and D refer to the individual drug products, to which the primary investigators and the subjects were blinded). Each patient received each preparation for 6 weeks, in one of the four sequences.
Patients, who took each of the four medications for 6 weeks, returned for assessment every 3 weeks. This assessment included thyroid function tests, clinical assessment and distribution of the study medication. Thyroid function tests were: total thyroxine (T4), total triiodothyronine (T3), resin thyroxine uptake and thyrotropin level (TSH). The free thyroxine index (FTI) was calculated from the product of T4 and resin uptake.
At the end of each 6-week period, patients were admitted overnight for a 24-hour bioavailability study. For this test, thyroid function tests were measured one hour before drug ingestion, and at 1, 2, 3, 4, 6, 8, 12 and 24 hours after ingestion. Drug bioavailability was determined by calculating the area under the 24-hour curve (AUC) for T4, T3 and FTI. The area under the curve at steady state is proportional to the amount of drug (or metabolite) that is actually absorbed into the blood during that time. The maximum serum concentration (Cmax) and the time to maximum serum concentration (tmax) were also determined for T4, T3 and FTI.
In addition, 20 tablets of each product were assayed for their levothyroxine content.
All products assayed fell within the FDA specified +/- 10% of the specified dose. The 100 mcg tablets varied between 90.4 mcg and 105.4 mcg; the 150 mcg tablets were between 140.2 mcg and 151.1 mcg.
24 patients completed the four 6-week periods of drug administration. Two were eliminated from the final analysis because of protocol violations, leaving 22 patients analyzed. For these 22 patients, all women:
The cause of hypothyroidism was Hashimoto's thyroiditis in 14, radioactive iodine therapy in 5 and thyroidectomy in 3.
12 out of the 22 patients had TSH levels > 15 mU/L prior to the initiation of thyroid replacement therapy (indicating the absence of significant endogenous thyroid function).
Of the 88 periods analyzed (4 per patient x 22 patients), 3 were eliminated (one each from 3 separate patients), due to protocol problems.
Compliance with study medication appears to have been excellent.
At the end of each 6-week period, subjects underwent 24-hours of multiple blood draws for bioavailability studies.
The authors make a number of points in their discussion, including:
The publication controversy
This study became the object of much controversy and publicity when it was revealed that the pharmaceutical industry sponsors of the work (Knoll pharmaceuticals), unhappy with the results, suppressed its publication and then published the same data elsewhere with a different analysis and interpretation.
The exact sequence of events is complicated and fascinating, and is described in detail in the same issue of JAMA in a long editorial by Dr. Drummond Rennie. I strongly recommend reading this editorial, which is available at JAMA's website.
Knoll pharmaceuticals present their side of the publication controversy in a letter to the editor published in the same issue of JAMA, to which the authors respond.
Furthermore, Knoll presents a factual critique of the methods and results of the study in another letter to the editor, to which the authors also respond.
In my opinion, this study is reassuring concerning the bioequivalence of brand and generic l-thyroxine preparations, but there are a few caveats. Not all generic formulations were tested here, so equivalence of other products cannot be strictly inferred. Also, the significance of a difference in serum TSH levels between the products remains unclear to me. The authors state that this difference did not reach statistical significance and that all products produced a circadian drop in TSH that was comparable. The analysis presented by Knoll in their paper (see their letter to the editor) came to a different conclusion, but they seem to have performed a statistical manipulation by "normalizing" the baseline TSH, a procedure that is rejected by the authors of the present study. Presumably, if their was a difference in TSH levels between the different products, it was not a major one.
The fact that only two specific generic formulations were studied here raises another issue: why don't physicians order specific "brands" of generic medicine? This would maintain much of the cost savings of generic prescribing, while greatly reducing the potential for inter-product variability. The choice of the specific generic product chosen for a patient now lies with the pharmacist, as a rule. In order to remain price-competitive, pharmacists will presumably choose the least expensive brand, which may vary from time to time. Perhaps "branding" individual generic products would be a cost-effective solution for medications with a narrow therapeutic index, such as l-thyroxine.
Finally, as health care takes on more and more of an "industrial" flavor, controversies surrounding corporate sponsorship of medical research and pressure on the authors of such research are likely to become a more pervasive problem.
May 23, 1997
Submit a comment about this article
Return to Journal Club home page