Barriers in the quest for quality drug information: salutary lessons from TGA-approved sources for thyroid-related medications

In iodine-replete countries, about 5% of the population have a thyroid disorder,1 of whom about a quarter will require long-term medication either to correct deficiency or to control thyroid hormone excess. In 2005, about 700 000 Pharmaceutical Benefits Scheme prescriptions were filled in Australia for thyroxine, with about 80 000 scripts for the antithyroid drugs carbimazole and propylthiouracil.2

Although no major new therapeutic agent has been introduced in the thyroid field for some years, the body of knowledge and evidence has advanced. It is important that these developments are reflected in the product information (PI) widely used by medical practitioners and a range of health professionals, who may not refer directly to current scientific literature on thyroid disorders. MIMS (Monthly index of medical specialties) annual,3 in its 30th edition in 2006, and the Australian prescription products guide (APPG) 2006, 35th edition,4 are the most commonly used compilations of PI (provided by the drug manufacturer and approved by the Therapeutic Goods Administration [TGA]). Both are mandatory pharmacy references specified by some state pharmacy boards.5,6

My review assesses whether PI-based information on thyroid-related medications as reproduced in MIMS annual 2006, MIMS Online and APPG 2006 is in accord with current peer-reviewed medical literature and contemporary therapeutic practice.

There are several reasons why patients taking thyroid-related medications, who number over 200 000 in Australia, need reliable information. First, they should aim for some self-sufficiency in relation to their medications, because long-term treatment and follow-up often extends beyond contact with any one medical practitioner. Second, they may need to make informed choices between therapeutic alternatives. For example, a young woman with thyrotoxicosis who has future plans for pregnancy may be offered ablative treatment, with the prospect of subsequent lifelong thyroxine replacement therapy that will need to be adjusted during pregnancy. Her alternative of antithyroid drug treatment, with the possibility of remission or recurrence, might be dismissed if she were given poorly documented advice about the safety of antithyroid drugs during pregnancy and lactation. Third, there is potential for ill-advised dose adjustment, failure to recognise side effects and, at times, misuse7 of thyroid-related drugs.

The stated aim of the 2006 MIMS annual is to serve as “the byword for accurate, reliable, comprehensive and independent medicines information”. It also states that “Prescribers, and health care professionals in general, need to be confident in today’s litigation-conscious environment that the information used as decision support, in both electronic and print formats, is reliable, accurate, from a trusted source AND reflects the current APPROVED information”.3 Further, the title page of the 2006 MIMS annual states that “Product monographs in MIMS annual represent Therapeutic Goods Administration (TGA) approved product information, which is the result of years of research and development by the sponsor company and of painstaking evaluation and review by the Drug Safety and Evaluation Branch of the TGA”.3 These statements imply an intention to offer both pharmaceutical information and reliable clinical advice. It follows that patient care can be influenced by the quality and accuracy of the information in PI supplied by pharmaceutical manufacturers or sponsors, and then endorsed by the TGA. This potential link to clinical decisions mandates accountability for PI-based drug information.

There are numerous discrepancies when information on thyroid-related medications in the PI-based June 2006 MIMS annual, MIMS Online and APPG 2006 is compared with the current medical literature and with Australasian8,9 and international reviews.10,11 Major points of contention that have a direct impact on patient care are summarised in the Box.

Where possible, evidence for the critiques of the PI has been presented according to the notation of National Health and Medical Research Council (NHMRC) guidelines (level of evidence [I–IV] and grade of recommendation [A–D]).12 In some, the NHMRC level of supporting evidence does not appear high, as randomised trials have not been performed — many thyroid-related medications have been in use for decades and their introduction pre-dates stringent documentation. However, the currently available lower-level evidence has been confirmed in endocrinology clinical practice over many years and has been incorporated as current best practice.

Some examples of contentious points, which are discussed further in the Box, include:

• A healthy euthyroid adult need not experience a period of iatrogenic hypothyroidism after near-total thyroidectomy, as would occur if PI-based guidelines for commencement of thyroxine treatment (“In all cases, Oroxine [Sigma] should be initiated at not more than 50 microgram/day . . .” [eg, MIMS annual 2006. Oroxine. Precautions: Initiation of therapy]) were followed.

• Further, it is apparent from clinical practice that continuation of antithyroid drug without dose modification until a patient is shown to be euthyroid can lead to serious over-treatment. (“After control of thyrotoxicosis, the dose of propylthiouracil should be gradually decreased to 50 mg twice daily.” [eg, MIMS annual 2006. Propyl-thiouracil. Use in pregnancy]). If a patient were then to abruptly stop the antithyroid drug either by own choice or on medical advice, the thyrotoxicosis would probably recur.

• The recommendation that liothyronine can be used for “myxoedema unresponsive to . . . thyroxine” and for “other conditions of the hypometabolic state” (eg, MIMS annual 2006. Tertroxin [Sigma]. Dosage and administration), is contentious and potentially harmful.7 There is no evidence-based confirmation that this ill-defined group of conditions actually exists; this recommendation provides an apparent endorsement of a common mode of thyroid hormone misuse.7

Thus, if such recommendations for the use of thyroid-related medications, as espoused in the PI and thus in MIMS and APPG, were mistakenly interpreted as firm guidelines for prescribing practice, these texts would legitimise potential misuse of medication, while implying that several established therapeutic strategies are “off label”.

The formulation, source, dosage and adverse effects of iodide are not documented in MIMS or APPG; no PI has apparently been submitted, as there is no commercial sponsor. Nevertheless, these texts make imprecise or dubious recommendations for its use within the information pertaining to other thyroid medications.

The recommendation that high-dose iodide should be used routinely in preparation for surgery in thyrotoxicosis (eg, MIMS annual 2006. Neo-Mercazole [Link]. Dosage and administration — last paragraph) is erroneous. While short-term high-dose iodide does temporarily inhibit thyroid hormone release and reduces the vascularity of the gland in Graves disease, it has no beneficial effect on thyroid blood flow in toxic multinodular goitre (quality of evidence: III-2, B).23 It should be noted that sustained iodide excess can exacerbate thyrotoxicosis and can impair the response to antithyroid drugs, with the potential for severe drug-resistant thyroid overactivity that may require emergency surgery (quality of evidence: IV, B).24

Of greater concern is the recommendation (eg, MIMS annual 2006. Neo-Mercazole. Use in pregnancy) that antithyroid drugs should be replaced by high-dose iodide in late pregnancy, a manoeuvre that is specifically contraindicated (quality of evidence: IV, B),19 except in the occasional situation of preparation for surgery in the second trimester (quality of evidence: IV, B).19 Iodide readily crosses the placenta and marked excess can severely impair fetal thyroid function, causing fetal goitre and neonatal hypothyroxinaemia, particularly in premature infants (quality of evidence: IV, B).18,19 It should be noted that the iodide doses used for transient control of thyrotoxicosis, or to diminish thyroid vascularity in Graves disease, using Lugol’s iodine, which contains 130 mg/mL of elemental iodine, are at least 100-fold greater than the 0.1–0.3 mg daily intake recommended for prevention of iodine deficiency in pregnancy (quality of evidence: III-1, A).25

The TGA has initiated a consultancy to improve the range and quality of consumer medicine information (CMI), which it also endorses.26 The preliminary discussion paper26 defines PI as the benchmark for CMI. While that document acknowledges that problems arise if PI is not kept up to date,26 there is no clear perception of any need to go beyond PI sources in preparing optimal CMI. That current PI-based sources can differ widely from contemporary therapeutic practice is a reality that should be considered in preparing CMI. Under present arrangements, CMI can be no better than the PI supplied by manufacturers or sponsors of repackaged imported products.

According to current regulations, initiatives to update or alter the PI of prescription medications require a manufacturer or sponsor to apply to the TGA and pay a fee for review of a revised submission.27 The TGA does not normally initiate changes in PI. While there is some provision for a sponsor to make “self-assessable” changes in PI as defined,27 it is unlikely that the problematic recommendations identified here could be altered by “self-assessable” revision under current regulations.27 Thus, each of the four companies that market the medications listed in the Box would need to make a separate submission to the TGA to initiate revision; there is no commercial incentive to do so. The situation regarding iodide is unusual because it has no commercial sponsor; accountability for the suggestions on its use is undefined.

The quality of information on thyroid-related medications in the PI of each drug, and thus in MIMS and APPG, is deficient to an extent that requires prompt review. The established mechanisms for updating PI are unlikely to achieve this, especially as the accountability for some information in these texts is undefined. An initiative to revise out-of-date or inaccurate sections could involve Australian clinicians who observe the effects of thyroid-related medications and are acquainted with relevant medical literature. Their advice can be sought, either directly through the Endocrine Society of Australia, or through the Therapeutics Committee of the Royal Australasian College of Physicians. A clearly defined, professional peer review process, the cornerstone of reputable medical literature, might improve the information currently available for both new and long established thyroid-related medications.

Addendum: Review of sequential MIMS annuals indicates that there has been no substantive change to the entries for carbimazole (Neo-Mercazole) or propylthiouracil since 1985. The entry for liothyronine (Tertroxin) was updated in 1988 and that for thyroxine (Oroxine) in 1990. Advice that dosage of thyroxine should be increased in pregnancy was added in 2004. Storage instructions for thyroxine were revised in 2006.