helvella – Dosing By Formula: Have just completed... - Thyroid UK

Thyroid UK

117,619 members136,417 posts

helvella – Dosing By Formula

helvella profile image

Have just completed a first draft discussing the NICE-recommended approach of dosing by weight.

This is entirely my personal opinion. Not official in any sense. Not Thyroid UK.

As such, it is certainly open to discussion and disagreement. Even spelling and grammar! :-)

helvella – Dosing By Formula


21 Replies

Excellent post. Can you send it to 'the committee'? Or publish it in bmj?

helvella profile image
helvellaAdministrator in reply to BB001

I put several of those points in my submission to NICE during the guideline development.

Feel free to take whatever you wish of it and send to NICE! They can always ignore you as well. :-)

One of the papers associated our group discusses this very thing. Two alternatives are discussed and the use of TSH as a titration for dose questioned.

Do Thyroid-Stimulating Hormone, Body Weight, or Body Mass Index Serve as Adequate Markers to Guide Levothyroxine Dose Titration?

October 2013Journal of the American College of Surgeons 217(4):752-3

DOI: 10.1016/j.jamcollsurg.2013.06.008

Johannes W. DietrichJ, Melvin Khee Shing Leow, Simon L. Goede, Rudolf Hoermann

helvella profile image
helvellaAdministrator in reply to diogenes

Thank you!

TSH110 profile image
TSH110 in reply to diogenes

Is there any way to read this it seems to be behind a paywall with no abstract? Ta

diogenes profile image
diogenes in reply to TSH110

I've extracted the body of the paper as underneath:

Ojomo and coauthors published a retrospective evaluation of patients post-thyroidectomy, concluding that maintenance levothyroxine (L-T4) doses are best predicted by body mass index (BMI) rather than by the traditional practice of measuring body weight.

Others concomitantly used an even larger cohort, showing

instead that L-T4 doses were better estimated by body

weight than by BMI.

Why do 2 trials addressing the same issue in similar

settings arrive at opposite conclusions? One of the reasons

may emanate from the statistical methodologies used.

Multivariate logistic regression may result in fallacious

rejection of explanatory variables in situations of multi-

collinearity, when measurements of body weight and

BMI were both included in one of the trials.

Both articles demonstrated that a daily L-T4 substitution dose between 1.5 and 1.6 mg/kg was required to achieve euthyroidism, which does not depart significantly

from traditional dosing recommendations. However, the

same outcome was attainable by BMI-based dosing

between 25 and 34 kg/m


The weight-based formula is inaccurate at the extremes of body weight unlike BMI; the former does not account as well for the volume of L-T4 distribution as the latter.

Notably, both trials relied purely on a thyroid-

stimulating hormone (TSH)-centric approach to estimate

adequacy of substitution dose. It was recently demonstrated that TSH alone might be an insufficient yardstick

to define the degree of euthyroidism because hormone

concentrations in the hypothalamic-pituitary (HP) unit

may equilibrate with circulating iodothyronines at

different rates compared with peripheral tissues.

Moreover, TSH secretion is influenced not only by T4 levels,

but also by structural parameters of the feedback loop,

neuromodulators/cytokines, and hysteresis.

A disconcerting weakness in both articles is that individ

uals were collectively compared using inferential statistical

techniques without regarding underlying physiologic prin

ciples. A composite of euthyroid TSH values within the

population reference range from diverse people, when

decoupled from corresponding free T4 (FT4), yields

a noisy scatter of “pseudo-random” data points concealing

an ordered structure, which the statistics fail to unravel.

Owing to wide variation, the proposed formulas, although

showing significant association at the population level, are

not very useful according to our data as instruments

for fine-tuning of dose-adequacy in an individual

It is crucial to appreciate that individuals behave

fundamentally as stochastically independent entities,

each controlled by their own characteristic HP axis. This

should be best analyzed using groups of individualized

TSH-FT4 curves. We believe that systems theory repre-

sents the appropriate approach to study such TSH

responses at the individual patients’ level to better

comprehend targeted dosing requirements of hypothyroid


For now, these conflicting observations question the

validity of presently adopted clinical practices. Future

studies on T4 substitution therapy should include addi

tional parameters of peripheral thyroid signaling such as ceruloplasmin or sex-hormone binding globulin, and

quality of life metrics for assessment of neuropsychologic

effects. The utility of bivariate reference regions for FT4

and TSH

personalized T4 substitution via an inferred homeostatic

HPT axis set-point computation from multiple paired

measurements of TSH-FT4 should also be investigated.

helvella profile image
helvellaAdministrator in reply to diogenes

Thank you - it is good to read questioning.

TSH110 profile image
TSH110 in reply to diogenes

Thanks diogenes 😉

I like it , i can't find any mistakes , and i think it makes a clear and compelling case that they should use their brains rather than a formula. I particularly like the emphasis on smaller titrations , and more frequent reviews of those needing to start on a low dose.

The difference between patient 1's 125 mcg and patient 3's 25mcg is truly ridiculous, and it's good to see it clearly spelled out like that.

Having said that , i haven't seen much evidence on the forum of GP's actually using this 'formula'.. if they did we wouldn't see so many middle aged patients stuck on 50mcg for a decade.

helvella profile image
helvellaAdministrator in reply to tattybogle

It's been whirling around my head since the NICE guidelines were in draft form.

Finally got to typing it out and trying to make a cogent (hopefully!) argument.

tattybogle profile image
tattybogle in reply to helvella

Showed 'im indoors the difference in persons 1-4 doses .

He used to work in A&E... his reaction was;

" deleted expletive* , looks like they're just making it up as they go along" :)

helvella profile image
helvellaAdministrator in reply to tattybogle

If GPs are using it, then my points are valid.

If GPs are not using it, what are they using instead? It is fine to be selective, intelligently, about guidelines but how can there be such a disjoint between practice by GPs and NICE?

Scrumbler profile image
Scrumbler in reply to helvella

I worry about what’s whirling around your head. Do you get many headaches 🙃 ?

TSH110 profile image
TSH110 in reply to tattybogle

They used it on me at the endocrinologists ☹️I asked them why my sister who is tiny was on 175mcg but me who towers over her was only in 125mcg Levo - no satisfactory explanation was provided. I felt pants on 25, 50, 75, 100, 125 125/100, 150 self medicating and 175 self medicating I gave up and switched to NDT. Once on 1.25 grains - normality was restored. I realised later that thyroid cancer was probably the reason for the dose differential - even that’s been abandoned now in her case 🙄

Scrumbler profile image
Scrumbler in reply to TSH110

They didn’t use any formula on me when they put me on 200 mcg after TT when I was 61 kg.

I’ve sometimes wondered how they arrived at this. I was successfully taking 60 mcg of T3 at the time. Perhaps they thought that 200 mcg Levo was some sort of appropriate ‘equivalent’.

But then that would require some sort of logical thought processes. Silly me. ☺️

Not sure what 'formula' they used on me after a sub-total thyroidectomy (9/10ths removed) to arrive at 75 mcg Levothyroxine Scrumbler. When the Levo was Cox, Norton or Goldshield Eltroxin, it worked quite well. Could not take Actavis at the time and stll could not in 2010, when only generics were available to us. NDT over the past 3 years has worked quite well, only causing a few gastric problems since only being able to hae Erfa. WP NDT worked very well for me.

Thank you!

Excellent discussion could not be agree more. Thanks for this.

Great article - thanks for producing this.The Nice guidelines suggest ‘rounding’ to the nearest 25mcg. Your example of someone weighing 70 or 71 k suggests the latter would be rounded up. Do you think Nice would always ‘consider’ rounding up or would it be rounding down if that was closer?

helvella profile image
helvellaAdministrator in reply to Bearo

In working out the examples, I used the ROUND function in Excel, which rounds to nearest - that is, 1.49 rounds down to 1, and 1.50 rounds up to 2.

As I did that, I was pretty comfortable with my interpretation. But I certainly agree someone could easily switch to round down or round up without thinking it through.

In the end, the effect of rounding will be the same, sudden jumps, but the point on the scale at which it occurs would shift.

helvella profile image

This is the relevant bit of the spreadsheet I used.

Screenshot of Excel calculations of dose by formula

You may also like...