First-line surgery in prolactinomas: lessons from a long-term follow-up study in a tertiary referral center

Abstract

Context: Although consensus guidelines recommend dopamine agonists (DAs) as the first-line approach in prolactinomas, some patients may opt instead for upfront surgery, with the goal of minimizing the need for continuation of DAs over the long term. While this approach can be recommended in selected patients with a microprolactinoma, the indication for upfront surgery in macroprolactinomas remains controversial, with limited long-term data in large cohorts. We aimed at elucidating whether first-line surgery is equally safe and effective for patients with micro- or macroprolactinomas not extending beyond the median carotid line (i.e., Knosp grade ≤ 1).

Methodology: Retrospective study of patients with prolactinomas Knosp grade ≤ 1 treated with upfront surgery. The primary endpoint was patients' dependence on DAs at last follow-up. The secondary endpoint was postoperative complications. Independent risk factors for long-term dependence on DAs were analyzed.

Results: A microadenoma was noted in 45 patients (52%) and a macroadenoma in 41 (48%), with 17 (20%) harboring a Knosp grade 1 prolactinoma. Median follow-up was 80 months. First-line surgery resulted in long-term remission in 31 patients (72%) with a microprolactinoma and in 18 patients (45%) with a macroprolactinoma (p = 0.02). DA therapy was ultimately required in 11 patients (24%) with microadenomas vs. 20 (49%) with macroadenomas (p = 0.03). As for the latter, DA was required in 13 patients (76%) with Knosp grade 1 macroadenomas vs. 7 patients (29%) with Knosp grade 0 macroadenomas (p = 0.004). There was no mortality, and morbidity was minimal. Knosp grade 1 prolactinomas (OR 7.3, 95% CI 1.4-37.7, p = 0.02) but not adenoma size (i.e., macroprolactinomas) were an independent predictor of long-term dependence on DAs.

Conclusions: First-line surgery in patients with microprolactinomas or macroprolactinomas Knosp grade 0 resulted in a good chance of non-dependency on DA therapy. However, in patients with prolactinomas Knosp grade 1, first-line surgery cannot be recommended, as adjuvant DA therapy after surgery is required in the majority of them over the long term.

Keywords: Dopamine agonists; Knosp grading; Long-term outcome; Macroadenoma; Microadenoma; Primary surgical therapy; Prolactinoma.

Fig. 1

Flow chart of patient selection process. Out of 182 patients with a prolactinoma, first-line surgery was performed in 98 patients, with 86 patients included in the final analysis given the presence of long-term follow-up data

Fig. 2

Impact of first-line surgery on PRL levels as a function of adenoma size. Differences in PRL levels before and after surgery in relation to adenoma size. Both baseline and postoperative PRL levels are significantly higher in patients with macroprolactinomas than those with microprolactinomas (p = 0.01 and p = 0.04, respectively), but not at long-term follow-up (p = 0.39). PRL levels significantly decreased in both cohorts compared to baseline, independent of the initial tumor size (i.e., microadenoma or macroadenoma). There is a significant difference between postoperative and long-term PRL values (p = 0.01 for microadenomas; p = 0.03 for macroadenomas, respectively). (***p < 0.001; **p < 0.01; *p < 0.05)

Fig. 3

Long-term outcome following first-line surgery. Multimodal treatment (i.e., surgery ± DA) resulted in long-term control of hyperprolactinemia in 41 patients (95%) with a microprolactinoma vs. 35 patients (88%) with a macroprolactinoma (p = 0.25), namely in 22 macroadenomas (96%) of Knosp grade 0 vs. 13 (76%) with Knosp grade 1 (p = 0.14). Surgery alone resulted in long-term remission in 31 patients (72%) with a microprolactinoma vs. 18 patients (45%) with a macroprolactinoma (p = 0.02); namely in 15 (68%) patients with a macroadenoma Knosp grade 0 vs. 3 (18%) patients with a macroadenoma Knosp grade 1 (p = 0.004). For the long-term control of hyperprolactinemia, a significantly greater need for DA therapy was noted in patients with a macroprolactinomas (49%) than in patients with a microprolactinomas (24%, p = 0.03), and in macroprolactinomas Knosp grade 1 (76%) compared to macroprolactinomas Knosp grade 0 (29%, p = 0.004) (**p < 0.01; *p < 0.05)

Fig. 4

Kaplan–Meier estimation of recurrence-free intervals. a Recurrence-free intervals were not significantly shorter in patients with a microadenoma (354.3 ± 25.6 months) than in those with a macroadenoma (324.4 ± 33.2 months); log-rank test, p = 0.34. b However, recurrence-free intervals were significantly shorter in patients with a Knosp grade I prolactinoma (201.5 ± 25.2 months) than in those with a Knosp grade 0 prolactinoma (396.4 ± 22.5 months; log-rank test, p = 0.01)