Cabergoline and prolactin: what the trials showed

Cabergoline has one of the better-established evidence bases in pituitary medicine and one of the worst-documented ones. This page is about the documentation, because it is the part a buyer on a research catalogue actually has to navigate: which numbers come from a randomised trial, which come from a cohort, and which two numbers in PubMed contradict each other by a factor of thirty.

The headline meta-analysis rests on four publications

The most-cited comparison of cabergoline against bromocriptine is a 2011 systematic review and meta-analysis in Pituitary, and its methods section tells the story. 418 references were identified, 34 articles obtained in full, 19 selected and 15 excluded: one randomised study that compared cabergoline against placebo, one that compared different amounts of cabergoline, five case series, four controlled non-randomised studies, three retrospective series and one cohort study.

Four publications ended up in the review and in the final analysis.

Against that base, the pooled results look solid on their face. Normalisation of serum prolactin favouroured cabergoline with a risk ratio of 0.67 and a 95 % confidence interval of 0.57 to 0.80. Menstruation with return of the ovulatory cycle was 0.74 (0.67 to 0.83). Adverse effects were more frequent with bromocriptine, 1.43 (1.03 to 1.98). The review concludes that the meta-analysis shows new evidence favouring cabergoline for prolactinomas and idiopathic hyperprolactinaemia.

The intervals are narrow, but narrow intervals computed from four publications describe the consistency of those four publications. They do not tell you how many independent patients were randomised, how long they were followed, or whether the same patients appear twice. That is answerable only from the full texts, and this page stays with what the abstracts print.

Where fertility was actually measured

Two studies here report anything downstream of prolactin normalisation, and both are in women.

The first is a comparison of ovarian stimulation in hyperprolactinaemic patients with polycystic ovary syndrome. Untreated patients needed fewer ampoules of recombinant FSH (P < 0.04), fewer days to reach hCG administration (P < 0.04) and reached significantly higher peak oestrogen concentrations (P < 0.03), with higher ovarian volume and more follicles at every size. Cycles cancelled for mild ovarian hyperstimulation syndrome were 4 of 65 untreated against 1 of 42 treated.

The result people cite from that study is not in it: the abstract states there were no significant differences in pregnancy rate or in multiple pregnancy. The measured benefit was control of the ovarian response, not a live birth number.

The second is a 2023 prospective study in 110 infertile women with polycystic ovaries and normal serum prolactin, described as euprolactinemic. Two groups, clomiphene citrate 100 mg given as 50 mg twice from day 2 to day 6 alone, or the same regimen plus cabergoline 0.5 mg at days 2 and 9. Biochemical pregnancy was 36.0 % (18 women) with the combination against 14.0 % (7 women) with clomiphene alone, p = 0.011, and clinical pregnancy was 32.0 % (16 women). No significant differences in abortion, ovarian hyperstimulation rate, multiple pregnancy or other adverse effects.

Two things to hold onto. The population has normal prolactin, so the mechanism being tested is not prolactin lowering. And the abstract describes dividing patients into two groups, not randomising them, so this is a prospective comparison rather than a randomised trial. It is the only pregnancy figure on this page, and it comes from a study with no reported randomisation.

The recurrence numbers, side by side

This is the part worth the reader’s time, because the two most-quoted answers are irreconcilable and both are in PubMed.

One is a 2021 systematic review and meta-analysis of prolactinoma recurrence after dopamine agonist withdrawal: 3,225 records retrieved, 13 studies with 616 patients and 19 arms included, no significant heterogeneity, and a fixed-effect pooled recurrence proportion of 2 % with a 95 % confidence interval of 1 to 3 %. The authors themselves write that much more prospective work with larger case numbers and longer follow-up is encouraged to confirm the finding.

The other is a 2005 retrospective analysis of clinic records in 89 treatment-naive subjects with microprolactinomas, mean age 32.7 years with a standard deviation of 8.4, 84 women and 5 men. They had received cabergoline (n = 67, 0.5 to 3 mg weekly) or bromocriptine (n = 22, 2.5 to 10 mg daily) for a mean of 3.1 years, after which withdrawal was attempted. Fifty-seven subjects, 64 %, developed recurrence, at a mean of 9.6 months with a range of 1 to 44. Thirty-two subjects, 36 %, remained in remission beyond one year, at a mean of 3.6 years with a range of 1 to 7.

Two per cent against 64 % is not a rounding difference. The 2005 authors state that remission in 30 to 40 % of subjects is what their data support, and they note a direct relationship between pretreatment prolactin concentration and recurrence risk, with no difference between cabergoline (n = 21) and bromocriptine (n = 11). The reconciliations are known and neither abstract resolves them: the pooled analysis selects studies with normalised prolactin and radiographic involution before withdrawal, and the cohort included everyone regardless. Which rule applies is a clinical question, not a bibliographic one, and this page does not answer it. One thing the pair does support: no subject in the 89-patient cohort developed clinical features suggesting tumour expansion after stopping.

The valve question, and whose patients it comes from

Cabergoline is an ergot-derived dopamine agonist, and ergot derivatives have a cardiac-valve history. The meta-analysis of observational studies is the largest synthesis available. In Parkinson’s disease, five studies gave 634 cabergoline patients against 9,120 on non-ergot therapy or none, and valvular regurgitation of any degree at one valve or more was more frequent, adjusted odds ratio 7.25 with a 95 % confidence interval of 3.71 to 14.18, p below 0.0001.

In hyperprolactinaemia the effect is far smaller and confined to one valve. Seven studies gave 444 cabergoline patients against 954 untreated controls, with significantly higher odds of mild to moderate tricuspid regurgitation, adjusted odds ratio 1.92 (95 % CI 1.34 to 2.73), p = 0.0003. No significant difference in mitral or aortic regurgitation.

The follow-up cohort gives the natural-history half. Of 101 patients on ergot-derived agonists, 53 on pergolide and 48 on cabergoline, 64 % male and mean age 66.4 years, follow-up was 21.8 months. Valve abnormalities regressed in about a third of patients with significant multivalvular regurgitation and about half of those with single-valve regurgitation who stopped; seven new cases of grade three to four regurgitation developed in patients who continued. The authors conclude the drugs should be prescribed only when better alternatives are unavailable and the patient has access to echocardiography.

That cohort is Parkinson’s patients in their sixties, not hyperprolactinaemia. The odds ratio of 1.92 in the prolactin population and the 7.25 in Parkinson’s are not the same finding applied to different people, and this page keeps them apart on purpose.

What the mechanism means for the numbers

The reviews describe the chain: dopamine agonism at the pituitary usually produces normoprolactinaemia, eugonadism and tumour reduction, and cabergoline is superior to bromocriptine on efficacy and tolerance. Read as a whole, the record describes a drug that works, with a favourable comparative tolerability profile, and a genuinely unsettled long-term outcome literature. The interesting content is not the mechanism, which is well established, but the follow-up data, which are not.

Where this sits on the shelf

There is no dopamine agonist in this catalogue and this page carries no product link, for the reason in the note at the top. The honest neighbours on the same axis are clomiphene and the kisspeptin page.

What the data do not show

They do not show a live birth rate, because no abstract here prints one. They do not show a randomised fertility result in men, because the two studies that measured anything downstream of prolactin are both in women and one of them studied women with normal prolactin. They do not reconcile the 2 % and the 64 %, because both abstracts are printed here and the selection criteria differ. They do not show a valvular-regurgitation odds ratio for hyperprolactinaemia patients of the size printed for Parkinson’s patients, and they do not show one for mitral or aortic valves at all, because that difference was not significant. And they do not show how many individual patients sat behind the four publications in the headline meta-analysis, because the abstracts do not print it.

What we supply

Nothing on this page is a product of ours, and this page recommends nothing. It exists so that a reader chasing a prolactin or fertility claim can find the primary trial and read the interval.

Research use only. This page summarises published studies for research reference. It is not medical advice, not a protocol, and not a suggestion for human use. Nothing we supply is for human or veterinary use.