Setup: A 43 year old male presents to clinic with several months of headaches, decreased libido, and erectile dysfunction. He also says that he has been bumping into door frames more frequently of late, though he is does not feel off balance. Vision is grossly normal in all quadrants bilaterally on physical exam, though formal vision testing was not performed. There is no personal or family history of pituitary, parathyroid, or pancreatic cancer. A 1.7cm a T1 and T2 isointense, contrast enhancing sellar mass compressing the optic chiasm.
A prolactin level was drawn and found to be 174 ng/mL. The remaining labs were unremarkable. A resident presented this patient on radiology rounds, and the following conversation ensued:
Chief resident: What do you think about that prolactin level?
Resident: It's pretty low, given the clinical picture.
Chief resident: So what now?
Resident: I sent the lab again with serial dilutions.
Chief resident: Nice. Medical students, why did she do that?
Medical students: [pretend to be invisible].
Chief resident: Look up the Hook Effect, and get back to me.
In actuality, this chief resident then explained the answer because they were of the teaching-on-rounds phenotype (!!!!), but transcribing the answer in the above format would have been tiresome. Any rate, the Hook Effect. first, this is NOT an an eponym, which makes it easier to remember. In fact, it describes the pattern of signal response in ELISA based assays. As the concentration of the antigen gets very high, the signal increases to a maximum and then begins to decrease. Check it out:
Prolactin levels are tested by sandwich ELISA. Briefly, antigen is added in a solution to plates that have adherent antibodies to the antigen. A second set of free antibodies in solution are are added, and then the plate is washed, removing excess antigen and unbound free antibodies. An enzyme-linked antibody to the free antibody's constant region is added, and excess is then washed away. A reagent is added to cause color change or luminescence that is proportional to the amount of enzyme-linked antibody bound in the plate (which, under normal circumstances, is proportional to the amount of antigen present).
To create the Hook Effect, an excess of antigen is added that binds all of available sites of the plate-adherent antibodies with remaining antigen in solution. If enough antigen is added, the antigen in solution will bind a large proportion of the free antibody, preventing them from binding antigen that is also bound by antibody adhered to the plate and ultimately causing it to be washed away. After the plate is washed, some of the antigen-bound adherent antibody will not have a free antibody bound. For this reason, despite having bound antigen, these adherent antibodies will not have an associated enzyme linked antibody, leading to an inappropriately low signal. Honestly, this was a very confusing explanation. Just look at the diagram.
The solution to this problem is to decrease the concentration of antigen. This can be achieved in a controlled fashion with serial dilution. This CANNOT be achieved by washing the plate before the addition of free antibody, because the amount of antigen lost in the wash would be unknown. (Or so I am told).
Pearls:
If a lab value is inappropriately low (or high), it is often reasonable to repeat to rule out lab error. In some cases, the error might be repeatable due to the nature of the test. One such example is the hook effect in ELISA-based tests. Not an eponym.
In the case of suspected hyperprolactinemia (e.g. when there is a big ol' goomba in the pituitary and clinical symptoms of hyperprolactinemia), reorder the prolactin level with serial dilutions (or whatever your institution does to deal with Hook Effect).
The sensitivity of confrontational visual field testing is not great for bitemporal hemianopsia (50%, though sample size limited in this study). Check it out
Comments