Generic vs Name Brand L-Thyroxine Products: Interchangeable or Still Not? (2024)

Search for other works by this author on:

Article history

This issue of the JCEM provides us with new clinical data on L-thyroxine (LT4) equivalence (1, 2). The studies seem to reach opposite conclusions, but an understanding of the regulatory history shows us that they assess different aspects of the question. Before T₄ was comprehensively regulated by the Food and Drug Administration (FDA), Hansen (3) asserted that the 2 leading name brand preparations of LT4 were equivalent. Within 5 years, no less than 10 studies assessed the “bioequivalence” of these 2 products using various methods and concluded that they either were or were not equivalent (4). In 1997, due to concerns with product potency and variability, the FDA declared LT4 preparations to be “new” drugs and required that all existing and future LT4 products be approved through the new drug application (NDA) process in order to remain on the US market (5, 6). This established uniform expectations for drug performance and resulted in a significant improvement in the LT4 products available for clinical use (7). The NDA products were unique formulations and were not considered interchangeable. This remained the case until the abbreviated NDA process of measuring relative bioavailability (bioequivalence) of LT4 products was implemented to provide a mechanism for assessing the potential interchangeability of generic and referenced name brand LT4 preparations (8). Shortcomings of the traditional pharmaco*kinetic method (9) resulted in a modification of the process to correct for endogenous T₄ (10), but concerns remained that potentially clinically significant differences in a LT4 dose of 12.5% or more might not be recognized with the pharmaco*kinetic approach (9) in products designated as bioequivalent by the pharmaco*kinetic standard (11). Therapeutic equivalence codes were assigned once products met the bioequivalence specifications. The designation of AB was assigned if the standard for bioequivalence (8) was met, or products were rated as BX (not interchangeable) if this standard was not met (8). Current therapeutic equivalence ratings are summarized in Table 1 (12).

The theoretical insensitivity of pharmaco*kinetics was confirmed when the reference drug, Synthroid, was compared to a generic candidate (13). The generic was recognized as therapeutically equivalent, given the AB2 designation, and was advertised as interchangeable (C. J. Worell, Sandoz Introduces Levothyroxine Sodium Tablets, USP, A Bioequivalent Alternative Product to Synthroid and Levoxyl. Sandoz, a Novartis Company: Advertisem*nt to Physicians, 2004.). But the new generic was 12.5% more bioavailable than the referenced brand (13). Expert clinicians expressed great concern (11), and one of the studies in this issue of the JCEM directly assesses this same comparison (1). Further LT4 product improvement has occurred with the imposition of a narrower potency requirement by the FDA to minimize variability of the same LT4 product ingested from refill to refill (14, 15). To meet the new goal of 95–105% of labeled content (14), some products have been reformulated (16, 17). In both cases of reformulation, the FDA has been shown data documenting that the new formulations are therapeutically equivalent with the older versions of the same preparations by pharmaco*kinetic analysis (16, 17), but previous therapeutic equivalence ratings have not been reassessed.

So now to the new work that advances our understanding of bioequivalence, therapeutic equivalence, and the potential interchangeability of LT4 products. The reports in the current issue of the JCEM appear to reach opposite conclusions on the interchangeability question, but in fact, differences in study design and execution result in the documentation of 2 very different aspects of the issues outlined above.

In the study from Lomenick et al (2), the authors base their investigation on several very important issues and cite the highly relevant concern expressed by the American Academy of Pediatrics that “any change in source of the LT4 requires retitration of the dose” (2). That study was a retrospective chart review that assessed the quality of thyroid function control in patients treated with LT4 for congenital hypothyroidism. Subject records were reviewed and grouped based on clinician decisions to treat with Synthroid only (n = 35) or generic LT4 (n = 27). These results provide us with an important assessment of intraproduct (refill to refill) reliability because the design is consistent with a goal of maintaining the same LT4 source as recommended by the American Association of Clinical Endocrinologists (AACE), the American Thyroid Association (ATA), and The Endocrine Society (TES) (18). Verification that either generic or name brand LT4 was dispensed was made by phoning pharmacies, but a record of actual dispensing was not available for the duration of therapy. No standard for LT4 ingestion could be documented. At each clinic visit, TSH and usually a free T₄ (FT4) were drawn at unknown intervals before or after ingesting the LT4 for the day. Between visits, lab testing was done outside of the university laboratory for those who had had dose adjustments, and these data were included in the final assessment of TSH and FT4 variability.

Individuals were considered poorly compliant if the medical record documented 4 or more doses that were missed per month. There was no indication that a standard approach to compliance assessment had been followed. These limitations on treatment ingestion, data collection, and methods were appropriately acknowledged by the authors. The primary end points of TSH and FT4 were analyzed by determining the median and SD to measure variance by the traditional Wilcoxon rank sum test as well as by a linear mixed model (2).

Among those on a generic preparation, 26% were treated with exclusively Lannett LT4 (AB1, AB2), 11% ingested only Mylan (AB1, AB2), 19% ingested at least 2 different sources of LT4 not otherwise identified, and in 44% it was not possible to determine which generic was received. Characteristics of the 2 groups were balanced for the baseline degree of hypothyroidism. The assessed compliance with therapy and frequency of testing over the follow-up period was similar, although there was a trend of more outside testing in the Synthroid group. Neither TSH nor FT4 variance was different by Wilcoxon rank sum testing, but using the linear mixed model variance was lower in the group treated with generic LT4. The number of dose adjustment events was similar over the observation period. I am reassured by the observation indicating that contemporary high-quality LT4 products, both generic and name brand, had similar and fairly consistent clinical outcomes from refill to refill when used to treat congenital hypothyroidism.

However, the authors conclude that their observations “suggest that generic LT4 and Synthroid are at least interchangeable in young children with congenital hypothyroidism” (2). Such a conclusion is not supported by the data because this retrospective, parallel treatment assessment did not address the interchange of the name brand product with the AB2 (Synthroid referenced) rated generics listed. To assess interchangeability, the best bet would be to evaluate what little interchange of generic (AB1 [Unithroid referenced]) products occurred. This would need to be analyzed separately to gain insight into the question of interchange among the named AB1 generics because no subjects were treated with both name brand and a corresponding generic product.

Carswell et al (1) studied children with severe congenital hypothyroidism and acquired hypothyroidism. These investigators enrolled subjects who were severely hypothyroid (TSH > 100 mU/L at the time of initial diagnosis), and then restudied them prospectively later in life at a mean age of 10 years (8.4 y congenital hypothyroidism, and 12.9 y acquired hypothyroidism) when euthyroid on LT4. Thirty-one children were analyzed with an intention-to-treat model. Subjects underwent a 16-week, prospective, randomized, open label, crossover study of treatment with their usual dose of outpatient LT4. Sixteen received name brand LT4 first, and 18 received the same dose of an AB2-rated generic (Sandoz). At the end of each 8-week treatment period, TSH, FT4, and total T₃ (TT3) were measured within an individualized but consistent 2-hour window for each subject. Each subject was instructed to take the LT4 with water in the morning, 30 min before eating, and all were alerted to avoid soy- and iron-containing products. All study drugs were dispensed through a research pharmacy, and fresh lots of all medications were maintained throughout the study. Missed doses were to be doubled up the next day to ensure consistent dosing. Compliance was assessed with pill counts and patient/parent recall. The primary end point was the serum TSH value after each 8-week period, with FT4 and TT3 values assessed as secondary measures. All serum samples were kept frozen and run in duplicate in the same hospital laboratory and batch to eliminate interassay and interlaboratory variability (1).

At the end of the Synthroid phase of the study, TSH values were significantly lower than they were after the 8-week generic LT4 period. This difference in TSH outcomes was most pronounced in the congenital hypothyroidism group. There were no significant differences in FT4 and TT3. Of the 16 patients who were receiving the name brand product Synthroid at study entry, follow-up TSH values at the end of their Synthroid period were not different from those observed at baseline, indicating consistent TSH outcomes.

The authors chose the Sandoz-distributed generic for comparison because it is AB2 rated to Synthroid (see Table 1; Ref. 12). This generic appears to have been 12.5% more bioavailable than Synthroid when originally submitted to the FDA (13). The observation that the generic seemed to be less bioavailable than Synthroid in this head-to-head comparison remains unexplained (1). Because no tablet content analysis of the 2 products was reported, we can only speculate to explain this observation. Perhaps a reversal of the bioavailability relationship between the Sandoz generic and Synthroid has occurred since the abbreviated NDA approval (13). Such a significant change could be associated with reformulation or formulation drift of one product or the other, which seems very unlikely given the surveillance accorded these products by the FDA. Speculating further, is this a not-so-surprising example of the inability of traditional FDA pharmaco*kinetic testing with supratherapeutic doses to predict clinical outcomes when using significantly lower clinical doses? Or is this in some way representative of the issues in pharmaco*kinetic bioequivalence assessment recently reported in conjunction with a bupropion product (19)? Clearly, further research is required to reproduce these results and appropriately assess the ability of pharmaco*kinetics to predict clinical outcome. Consideration should be given to executing the AACE/ATA/TES proposed TSH-based protocol using clinically relevant doses of LT4 in athyreotic patients to evaluate its potential as a substitute for the current pharmaco*kinetic-based process of determining therapeutic equivalence to definitively address this issue (20).

These 2 studies have similarities because numerous lots of LT4, both name brand and generic, were utilized during the course of treatment in both. Each shows us refill-to-refill consistency in TSH outcomes, as measured by variability of TSH results in subjects treated, for the most part, with the same LT4 source longitudinally (2) and similar TSH outcomes observed when patients were treated for 8 weeks with the same brand from different lots (1). This is clear testimony to the quality of the LT4 products currently available. I believe that this has been brought about by FDA regulations that oversee the quality of manufactured and distributed LT4 products.

Study design and results, however, justify the conclusion of Carswell et al (1) that “brand name L-thyroxine and an AB-rated generic formulation are not bioequivalent in patients with severe congenital hypothyroidism.” I would consider this better phrased as, “… are not clinically interchangeable …” because FDA definitions of bioequivalence and therapeutic equivalence in this case are both based on pharmaco*kinetic data that appear insensitive to expose clinically significant differences in these products.

Until better data become available, I will continue to follow the AACE/ATA/TES recommendations on LT4 treatment in my practice. Physicians should: 1) alert patients that preparations may be switched at the pharmacy; 2) encourage patients to ask to remain on the same preparation at every pharmacy refill; and 3) make sure patients understand the need to have their TSH retested and the potential for dosing readjusted every time their LT4 preparation is switched (18).

Acknowledgments

Disclosure Summary: I have received consulting honoraria from Akrimax Pharmaceuticals for scientific data review and Abbott Labs for educational consulting in regard to bioequivalence testing.

For articles see pages 610 and 653

Abbreviations

• FT4

  Free T₄

• LT4

  L-thyroxine

• NDA

  new drug application

• TT3

  total T₃.

References

Copyright © 2013 by The Endocrine Society

Issue Section:

Editorials > Editorials - Editorial

Email alerts

See also

Citing articles via

More from Oxford Academic