The present data are, to our knowledge, the first to estimate the prevalence of v-betaLH in a Norwegian population. A carrier frequency of 17% amongst normal responders to COH in an IVF population is comparable to 11% in a similar Danish IVF population . In unselected populations a wide variation of v-betaLH carrier frequency was reported ranging from 42% in Finnish Lapps through 19% in Swedes to 7% in Hispanics in the USA . No association was found between POR and v-betaLH in this study. This finding is in accordance with the only reported genome wide association study concerning ovarian response to COH . Alviggi and co-workers  suggested that IVF patients with v-betaLH lack sufficient LH activity to adequately support FSH activity in multiple follicular development, leading to hypo-response to COH as defined above. The present results suggest that the decreased ovarian response in POR patients may require a different explanation, and point to the question of whether ovarian response to COH declines gradually from normal via hypo to poor, or if the three should be considered separate situations altogether . The hypothesis proposed by Alviggi and co-workers , that carriers of v-betaLH could benefit from exogenous LH in their COH, seems unlikely to apply to POR patients from the present results; however, a different study design is required to test this properly.
Differences in numbers of oocytes retrieved, embryos transferred and live births were as expected from the inclusion- and exclusion criteria. There were no differences between groups regarding the FSH receptor SNPs reported to influence ovarian response to COH (data not shown). The control group was also not matched to the POR group, leading to differences in age and BMI between groups. Albeit small, these differences are a possible confounder in the study inasmuch as BMI was reported to influence ovarian response to COH  and the risk of POR increases with age . The use of rFSH was similar between groups, reflecting that POR was unexpected by clinicians for the patients in the POR-group.
All COHs in this study, as in the studies by Alviggi and co-workers [8, 9], were by GnRH agonist mid luteal phase down regulation. Whether v-betaLH is associated with hypo-response or POR in an antagonist protocol is unknown. However, the low endogenous level of LH in agonist cycles could arguably make v-betaLH more influential in agonist than antagonist cycles.
Genotype data from the same patients regarding other signaling systems of importance in ovarian physiology and their association with ovarian response to COH were presented earlier [13–18]. This raises the question of multiple testing in genetic association studies. However, as the present study showed negative results, concerns over false positive findings were unwarranted.
The present study had two main limitations; sample size and inclusion criteria. The above sample size calculations show that the study had adequate power to find a difference between groups if v-betaLH prevalence was three times higher in the POR group than in controls. A study with more patients could have detected smaller differences between groups. However, to avoid a high rate of false negatives when applying v-betaLH as a predictor of POR, a high prevalence of v-betaLH in the POR group is required.
The inclusion criteria for POR patients did not comply with ESHRE’s Bologna criteria . The criteria for POR found in Table 1 were set to identify patients in which POR came unexpectedly, as it is in these patients that novel predictors of ovarian response could be most useful. Also the criteria in Table 1 assured that the patients included had few other known factors that could influence their ovarian response to COH apart from the putative genetic ones such as v-betaLH. The Bologna criteria on the other hand have ‘advanced maternal age’ and ‘previous POR’, as two of three criteria for POR, making them inadequate for identification of patients with an unexpected poor response, at least retrospectively as done in this study.
Other limitations in the present study were unavailability of s-LH measurements prior to and during COH, and no data on s-AMH or antral follicle count to predict ovarian response to COH.