This is the verbatim reason for rejecting a paper of ours to BMJ from a socalled thyroid expert in the UK. Let me say that I am not "an enthusiast for combined T4/T3 therapy" or anything else. I simply follow where the facts lead me. But I thought you ought to see a response from a reviewer (Dr S Pearce) regarding a submission of our where we argue for a rethink in thyroid diagnosis and treatment primarily substituting the range "shoehorn" diagnostic process for individual appraisal. I believe he is a member of the NICE committee presently re-examining the guidelines. Here it is in full (I highlight some of the more beautiful statements) with my name spelled wrongly to start with:
Midgely and colleagues provide a series of arguments that current reliance on serum TSH measurements in the diagnosis and monitoring of thyroid disease leads to suboptimal treatment. The article is unstructured and many original contributions are un-cited. In addition, on several occasions the US guidelines are cited to back up a sweeping claim, with no discussion of the granularity contained within them or in original sources. One of the authors’ key arguments is largely based on extrapolating findings from a cohort of thyroid cancer patients to people with hypothyroidism, which is clearly not appropriate. Many areas of the paper lack balance, with a one sided view of the literature being taken (to use the author’s phrase-‘cherry picking’). One might be cautious that the contents of this paper might be misused as a ‘quack’s charter’, justifying use or adjustment of thyroid hormone replacement in any symptomatic patient, irrespective of serum thyroid hormone or TSH concentrations.
Given the important finding in 25,000 attendees at the Colorado Health Fair (Canaris Arch Intern Med 2000), in which it was found that 60% of biochemically euthyroid people had one or more symptoms of hypothyroidism, the approach suggested here cannot be endorsed. Similarly, even if one believes the unlikely prospect that people with true hypothyroidism may have normal serum TSH, an RCT showed that such people didn’t feel better during thyroid hormone therapy (Pollock, BMJ 2001). We wouldn't treat a patient for cancer without having histological confirmation, and we shouldn't treat patients for hypothyroidism without biochemical confirmation.
1.“In an attempt to scale back on the avalanche of purported thyroid diseases created by this strategy, the TSH threshold for treatment was raised by recent guidelines (21)”. I would say the reason for considering that the TSH threshold for treatment might be raised to 10mU in most situations is that only 2 of 9 RCTs of levothyroxine in patients with a TSH in the 5-10mU/l range showed any improvement in symptoms. Therefore the evidence to favour treatment is weak; this is not an attempt to scale back on diagnosis, but to give patients appropriate, evidence-based treatments.
2.The arguments concerning thyroid cancer are spurious and irrelevant to patients who haven’t had thyroidectomy.
3.“Using the overall preference expressed by patients at the end of double blind studies as a proxy, patients did mostly favour T3/T4 combination therapy (29).” The problem with this assertion is that patient preference was reported in under half of these studies, so although widely touted by T3/T4 enthusiasts, there is highly likely to be reporting bias in this outcome.
4. “inferior quality of life instruments” The arguments that QoL questionnaires are misleading goes against the central tenet of QoL research, which is that the patient’s QoL is what the patient says it is. The fact remains that these same ‘inferior’ QoL of life instruments were, and still are, able to demonstrate meaningful QoL changes in many other conditions, but just not in the context of combined T3/T4 treatment trials of hypothyroidism. Which should we consider more likely; that the hypothyroid patient has an invisible QoL disturbance that can’t be found on questionnaires which are sensitive in multiple other conditions? Or that these frequently very vocal patients are susceptible to the placebo effect, which is abolished in the RCT setting?