While the prolactin response findings in ME/CFS studies are intriguing, some limitations limit the ability to make strong conclusions from the available data.
First, while five different published case-control studies have reported increased prolactin response to buspirone, these do not represent five different independent replications. Three of the five buspirone-prolactin studies shared an author/supervisor (P. O. Behan) [28, 32, 33], and all five studies were conducted in the United Kingdom. Thus, findings may relate to spurious effects specific to the few labs that have undertaken this research, and further independent replication is needed to validate the findings.
The sample size of each study was also relatively small, with no study’s ME/CFS group exceeding 30 cases. This raises the possibility that the demonstrated prolactin findings could be related to selection bias, as opposed to an abnormality generalizable to the ME/CFS population.
Studies have also detected a blunted prolactin response in ME/CFS, complicating a straightforward interpretation of over-responsivity of the prolactin system. Specifically, a study of exercise [40] and a study of insulin-induced hypoglycemia [41] both demonstrated blunted prolactin increases in ME/CFS patients. Future research should explore whether these effects represent abnormalities in prolactin regulation, or instead relate to differences in exercise parameters and/or blood sugar regulation. The relatively indirect influence of exercise and hypoglycemia on prolactin may help explain the opposite direction of effect, when compared to drugs that directly affect prolactin-regulating neurons in the brain.
Further, as stress is known to increase prolactin [42, 43], one must consider whether differences in stress-inducing factors may be responsible for the increased prolactin response seen in ME/CFS. Multiple papers have reported that ME/CFS patients demonstrate an increased magnitude of buspirone-induced side-effects, such as nausea [28, 29, 31]. A greater degree of nausea could lead to increased stress, with stress leading to increased prolactin response. However, even if factors such as nausea are responsible for the prolactin findings, the difference in side effects to buspirone would itself likely warrant further investigation.
Study-specific limitations also affect the interpretability of the evidence. Richardson’s 1995 paper includes very limited detail about the control group, in that the sex ratio and ages of the healthy controls were not reported. The paper also suggests that some of the controls may have been parents of the ME/CFS patients [29]. Thus, sex- or age-related confounding may play a role in the study’s findings. However, while sex ratio would be considered a clear confounder [35], note that studies do not appear to have demonstrated a significant correlation between age and prolactin response to buspirone in either healthy controls or individuals with depression [44–46]. Additionally, Behan’s 1996 paper of prolactin response included a significant error, describing the control group as 10 males instead of 30 individuals evenly split by sex [33]. As noted previously, an increased proportion of females in the control group would not be expected to lead to a spuriously increased prolactin response in cases [35]. Though, the error still raises concerns about the study’s data analysis procedures.
