It is well known that luteinizing hormone (LH), like many other glycop
roteins, is heterogeneous and presents several circulating isoforms. R
ecently, new sensitive immunometric assays measuring intact LH were de
veloped. These assays have been found to underestimate or to be incapa
ble of recognizing LH in some patients. This study was undertaken to d
etermine the prevalence of such cases and to define their clinical cha
racteristics. We compared three LH assays using as capture antibodies
either a monoclonal antibody that reacts exclusively with intact LH (E
S 600 Boehringer, Stratus Baxter) or a monoclonal antibody against the
beta subunit of LH (IMX Abbott). In 17% of 90 patients tested, ES 600
measured >50% lower LH concentrations when compared with the IMX. Mor
eover, in two cases LH was not detectable by ES 600 or Stratus, wherea
s it was normal with the IMX. We found another five such cases and dis
cuss here the clinical data and results of different hormone measureme
nts in these seven cases of 'invisible LH'. Although bioactive LH (mou
se Leydig cell assay) was normal, the existence of low or even undetec
table LH was clinically confusing and led to expensive complementary i
nvestigations such as gonadotrophin-releasing hormone analogue tests a
nd magnetic resonance imaging. The uses and limitations of these assay
s are illustrated by different clinical situations in which the result
s of the different assays have been misleading.