Neurocognitive function in acromegaly after surgical resection of GH-secreting adenoma versus naïve acromegaly

Abstract

Patients with active untreated acromegaly show mild to moderate neurocognitive disorders that are associated to chronic exposure to growth hormone (GH) and insulin-like growth factor (IGF-I) hypersecretion. However, it is unknown whether these disorders improve after controlling GH/IGF-I hypersecretion. The aim of this study was to compare neurocognitive functions of patients who successfully underwent GH-secreting adenoma transsphenoidal surgery (cured patients) with patients with naive acromegaly. In addition, we wanted to determine the impact of different clinical and biochemical variables on neurocognitive status in patients with active disease and after long-term cure. A battery of six standardized neuropsychological tests assessed attention, memory and executive functioning. In addition, a quantitative electroencephalography with Low-Resolution Electromagnetic Tomography (LORETA) solution was performed to obtain information about the neurophysiological state of the patients. Neurocognitive data was compared to that of a healthy control group. Multiple linear regression analysis was also conducted using clinical and hormonal parameters to obtain a set of independent predictors of neurocognitive state before and after cure. Both groups of patients scored significantly poorer than the healthy controls on memory tests, especially those assessing visual and verbal recall. Patients with cured acromegaly did not obtain better cognitive measures than naïve patients. Furthermore memory deficits were associated with decreased beta activity in left medial temporal cortex in both groups of patients. Regression analysis showed longer duration of untreated acromegaly was associated with more severe neurocognitive complications, regardless of the diagnostic group, whereas GH levels at the time of assessment was related to neurocognitive outcome only in naïve patients. Longer duration of post-operative biochemical remission of acromegaly was associated with better neurocognitive state. Overall, this data suggests that the effects of chronic exposure to GH/IGF-I hypersecretion could have long-term effects on brain functions.

Figure 1. Relationship between clinical predictors obtained from multivariate regression analysis and cognitive data.

(A) longer duration of untreated acromegaly (in months) is independently associated with worse visual memory performance. Higher GH levels are independently associated with worse visual memory recall in patients with naïve acromegaly, but not in cured patients. (B) shorter biochemical remission is positively associated with worse visual memory recall. CFT, Complex Figure Test. Partial R² were calculated controlling for the other variables in the model.

Figure 2. Neurophysiological results of patients with acromegaly.

(A) Power spectra differences (frequency range: 1–40 Hz) among patient groups and healthy controls. Decreased EEG power was observed in fast bands (alpha and beta) for both patient groups. Note the decreased power in patients with acromegaly at peak alpha frequency, 10 Hz. (B) LORETA comparisons between acromegaly patients and healthy controls. Upper panel: comparison between cured patients and healthy subjects. Highest significant LORETA differences were found in left medial temporal cortex in beta 2 band (4 significant neighboring voxels, 1.37 cm³, p<0.04) and beta 3 band (15 significant neighboring voxels, 5.14 cm³, p<0.007). Cured patients showed significantly lower LORETA activity than healthy subjects in cortical areas shown in blue. Lower panel: comparison between naïve patients and healthy subjects. Highest significant LORETA differences were found in left medial temporal cortex in beta 2 band (8 significant neighboring voxels, 2.74 cm³, p<0.012) and beta 3 band (30 significant neighboring voxels, 10.29 cm³, p<0.005) and in right dorsolateral PFC in alpha (8 significant neighboring voxels, 2.74 cm³, p<0.012) and beta 1 (28 significant neighboring voxels, 9.604 cm³, p<0.006) bands. Naïve patients showed significantly lower LORETA activity than healthy subjects in cortical areas shown in blue. Scales show maximal t value and maximal significant t value (p<0.05, corrected for multiple comparisons). (C) ROI analysis comparisons among groups. ROI LORETA activity in acromegaly patients is reduced as compared to healthy subjects. *p<0.05; **p<0.01, corrected for multiple comparisons.

Figure 3. Relationship between clinical predictors obtained from multivariate regression analysis and LORETA data.

(A) lower LORETA Z-scores in left medial temporal cortex were associated with longer duration of untreated acromegaly. (B) higher GH levels were associated with decreased LORETA Z-scores in rPFC in patients with naïve acromegaly, but not in cured patients. (C) lower LORETA Z-scores in beta lMTC and rPFC was associated to shorter biochemical remission. rPFC, right prefrontal cortex; lMTC, left medial temporal cortex. Partial R² were calculated controlling for the other variables in the model.