TSH is not a reliable indicator of thyroid function, period. I don't mean this disrespectfully in any way but, this fact has become quite clear to me.
The invention of the TSH test in the 1970's was truly the undoing of things for thyroid patients. As thyroid hormone tests came about, doctors did not keep up with these advances and/or did not recognize the importance of evaluating thyroid function by looking at thyroid levels vs. pituitary hormone, TSH.
The National Institute of Health published an article in 2003 that confirms this (read final sentence):
http://www.ncbi.nlm.nih.gov/pubmed/12915350?ordinalpos=1&itool=EntrezSystem 2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pub med_SingleItemSupl.Pubmed_Discovery_RA&l inkpos=4&log$=relatedreviews&logdbfrom=p ubmed
For whatever it's worth, when I developed over 20 hypo symptoms and was diagnosed with hypothyroidism, these were my labs:
FreeT4 1.2 (.8-1.76)
FreeT3 3.5 (2.0-4.4)
TSH .019 (.45-4.50)
Now, if THAT TSH isn't diagnostically misleading, what is?
In case anyone is interested, my TSH was so low due to my Graves' antibodies. While many of you are familiar with Graves' antibody called TSI, the stimulating antibody that causes hyperthyroidism, there is another type of Graves' antibody called TBII that is the blocking antibody - it can either negate the effects of TSI, leaving the patient euthyroid or, it can cause hypothyroidism, suppressing TSH just like TSI.
While Hashi's is the most common cause of hypothyroidism, people can also develop hypothyroidism due to the presence of blocking TSH receptor antibodies without ever having had Graves' beforehand (or they had such a mild case of Graves' that their thyroid disease wasn't discovered until they went hypo.)
Obviously there are SO many reasons why doctors need to look at the actual thyroid hormone levels for diagnosis/treatment especially since TSH has no physiological implications whatsoever.