Predictors of Outcome and Long-Term Results

The February, l999 issue of the journal Neurosurgery contains a report by Tyrrell and colleagues from the University of California in San Francisco (UCSF) in which early outcomes and long-term results are analyzed in a large series of patients with prolactinomas following transsphenoidal microsurgery. Two groups of patients treated in different periods of time were followed, and outcomes including complications were evaluated to provide information on the benefits and risks of microsurgery in the management of prolactinomas. The author, J. Blake Tyrrell, is a Professor of Medicine and Head of the Section of Endocrinology at UCSF. All patients were operated on by Charles B. Wilson, Professor of Neurosurgery at UCSF. The statistical analysis was performed by Kathleen R. Lamborn, also at UCSF.

Data were reviewed for two groups of female patients: 121 patients treated surgically between 1976 and 1979 (Group 1) and 98 patients treated between 1988 and 1992( Group 2). Males were excluded because the numbers of males in these two periods were too few. The two time periods were chosen to determine if the later series had better outcomes based on greater experience and improved imaging, and the earlier series was chosen to obtain long-term outcomes in a series of patients treated at least fifteen years earlier.

Definitions

Normal prolactin (PRL) values were defined as 20ng/ml or less, or when a particular laboratory had higher upper limits of normal values, the laboratory’s definition of normal was accepted as normal. The postoperative PRL value was the lowest value documented during the first 7 days after surgery. Postoperative remission was defined as a PRL value of 20ng/ml or less measured within this interval. The duration of follow-up was the time in years to the date of the last available PRL value, and patients classified as experiencing continued remission exhibited no clinical symptoms of hyperprolactinemia, had undergone no additional therapy and demonstrated normal PRL values. Patients initially experiencing remission were classified as having the recurrent disease at the date of the first documented occurrence of any of the following: a PRL value of >20ng/ml or higher than the normal range for the laboratory, recurrence of amenorrhea or galactorrhea, or further treatment. Technical recurrence was defined as an elevated PRL value without symptoms after an initial remission and thereafter normal PRL values with no additional therapy. Continued clinical remission with hyperprolactinemia was defined as a slight elevation of PRL (22-33 ng/ml) with no symptoms and no requirement for additional therapy.

Results

The two groups were very similar in all respects although the earlier patients were more likely to have had amenorrhea for longer periods before treatment, and the later series of patients were more likely (61%) to have received bromocriptine before operation. Grade and stage of the adenomas was similar. Combining the two groups, 72% in the series achieved initial surgical remission: (91% for intrasellar microadenomas and 88% for intrasellar macroadenomas). However, there was a significant difference between women in Group 1 and Group 2 in the median postoperative PRL value (6 versus 2 ng/ml respectively).

Of the 132 patients who experienced initial remission, at the last follow-up evaluation, 89% remained in remission and 11% had experienced recurrence. Reflecting both a shorter period of follow-up and greater experience, 97% of patients in Group 2 remained in clinical remission at the last follow-up (median follow-up of 3.2 years) and only 2 patients ( 3%) had experienced recurrence.

Univariate analysis of the time to recurrence revealed that the postoperative PRL value was the best predictor of continued remission, although older age and lower adenoma stage were also significant.

Complications were more frequent in Group 1 ( 7 complications), 6 of seven being related to a postoperative cerebrospinal fluid leak or hematoma requiring re-operation or treatment for meningitis. In Group 2 there were only two serious complications, a postoperative intrasellar hematoma that did not require re-operation, and a cerebrospinal fluid leak that was treated successfully with temporary insertion of a spinal drain without re-operation. No deaths have occurred in more than 1,000 operations for prolactinoma spanning a period that includes the patients in Groups 1 and 2.

Discussion

The results of this study confirm the immediate and long-term efficacy of transsphenoidal surgery in the treatment of women with prolactinomas. Initial remission was achieved in 92% of those with PRL values of <100ng/ml. Patients with preoperative values >200ng/ml and large or invasive adenomas achieved less favorable outcomes. There were no differences in remission rates for Groups 1 and 2, although postoperative PRL values were significantly lower in Group 2. The likelihood of recurrence is lower in patients with the lowest postoperative PRL levels (2ng/ml or less). Prior therapy with bromocriptine had no effect on immediate surgical outcomes. This series compares favorably with 31 previously reported series with initial remission being achieved for 91% of women with intrasellar microadenomas, 82% of all women with microadenomas, and 66% of women with macroadenomas.

The most controversial aspect of surgical therapy is the wide variation in reported rates of recurrent hyperprolactinemia, the figure of 50% being quoted commonly in quantifying the risk of recurrence. The patients in our Group 1 represent those with the longest follow-up periods reported to date for transsphenoiday microsurgery, and our results do not confirm previously reported high recurrence rates. In our Group 1, with a median follow-up period of 15.6 years ( mean 13.0 years), 82% of patients exhibited normal PRL values and one patient demonstrated mild recurrent hyperprolactinemia without symptoms at their latest evaluations. Therefore, 84% of these patients with initial remission continued in clinical remission, yielding a recurrence rate of 16%. Other series reporting comparably low recurrence rates had follow-up periods of 5 years or less.

Conclusion

Both medical therapy and transsphenoidal surgery have roles in the optimal treatment of PRL-secreting adenomas. Although treatment with dopamine agonists has the advantages of non-invasive therapy, transsphenoidal microsurgery can afford selected patients immediate normalization of PRL levels, without the need for continuing dependence on medication. However, patient selection must be done with the knowledge that those patients with the highest likelihood of long-term remission are those with PRL levels of 200ng/ml or less and a microadenoma or noninvasive macroadenoma. It is equally important to assess the skill and experience of the surgeon performing the procedure. Transsphenoidal microsurgery is an effective alternative to long-term medical therapy for selected patients with prolactimomas.