This prospective study enrolled 22 patients with NFPA who were referred to the endocrinologist for postoperative follow-up during the period between January 2018 and August 2019. All patients underwent surgery using the endoscopic endonasal transsphenoidal approach (EETSA) from two neurosurgeons at our institution (Yong-Kil Hong, Sin-Soo Jeon) with more than 20 years of experience each. Patients with other underlying diseases or taking any medications for therapeutic purpose were not enrolled in the study. In addition, patients who relapsed after a previous treatment or those whose pituitary stalk was removed during the operation were not enrolled. Since the purpose of our study was to identify HRV changes according to the pattern of hormone deficiency after transsphenoidal approach (TSA) in patients with NFPA, the pituitary hormone function of patients with postoperative NFPA was evaluated using a combined pituitary function test (CPFT) 3 to 6 months after TSA; none of the subjects had any restrictions on performing the CPFT. Patients were placed in the deficient group if their hormones did not respond normally in the CPFT, and the preserved group indicated patients with a normal response as observed by the CPFT. The groups varied according to the type of hormone. Informed consent was obtained from all study participants, and the study was approved by the Institutional Review Board of Seoul St. Mary’s Hospital (KC17OESI0205) and registered with Clinicaltrials.org (NCT03164148).

HRV data were recorded using an electrocardiographic sensor with a disposable patch-type electrode (T-Rex TR100A, TaeWoong Medical, Gimpo, Korea). The device has previously been used in other HRV-related studies, and the stability of data collection with this apparatus has been verified [11-13]. The device was attached to the participant’s anterior chest (Supplemental Fig. S1), and there were no restrictions on daily movement, although a supine position was maintained during sleep. All patients were requested to maintain a supine position during the CPFT. Collected row data were transformed into HRV data via Kubios HRV software [14], and the results were corrected for missing or incorrect peaks. HRV parameters were generated as recommended by the Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology [9]. Parameter adaptation was performed in the same manner recently described by Park et al. [13]. For time-domain analysis, the mean HR and the standard deviation of the normal-to-normal interval (SDNN) were measured. For the frequency domain, low frequency (LF) (0.04 to 0.15 Hz), high frequency (HF) (0.15 to 0.40 Hz), and the LF/HF ratio were analyzed. HF is an indicator reflecting parasympathetic effects, while LF is used as a reflection of sympathetic activity [10]. The LF/HF ratio and SDNN represent the subject’s overall autonomic function [10]. HF decreases in the disease state [15]. In the event of autonomic dysfunction, the LF/HF ratio increases and SDNN decreases [15].

Changes in hormonal function were evaluated via the CPFT, which was performed between three and six months after the surgery. The CPFT was performed after an overnight fast, and the results of testing were confirmed by two endocrinologists (C.J., J.H.). After intravenous injection of 0.1 U/kg regular insulin, 200 μg thyrotropin-releasing hormone, and 100 μg luteinizing hormone (LH)-releasing hormone, the responses of serum growth hormone (GH), cortisol, ACTH, prolactin (PRL), thyroid-stimulating hormone (TSH), LH, and follicle-stimulating hormone (FSH) were collected serially for 2 hours (at 0, 30, 60, 90, and 120 minutes). Hypoglycemia that occurred after administration of regular insulin was deemed valid when serum glucose measured after 30 minutes of intravenous insulin was reduced to 40 mg/dL or less or by 50% or more compared to baseline. The test results were interpreted as previously described [16,17]. The criteria for normal responses were as follows: the peak GH level increased above 3 ng/mL; the peak cortisol level increased above 18 μg/dL or increased by more than 7 μg/dL compared to the basal level; the peak TSH level increased by more than 5 μIU/mL compared to the basal level, while free thyroxine was within the normal range; the basal PRL level ranged from 2 to 15 ng/mL, and the peak PRL level was more than twice the basal PRL level. For gonadotropin (Gn), an increase in FSH by more than 2 mIU/mL and an increase in LH by more than 10 mIU/mL compared with the basal level were considered normal responses. When a particular hormone did not show a normal response, it was deemed to be deficient, and when the normal response was confirmed, the hormone was deemed to be preserved. The presence of multiple hormone deficiencies was defined as deficiencies of two or three individual anterior pituitary hormonal axes. Panhypopituitarism was defined as the coexistence of GH, ACTH, TSH, and Gn deficiencies. None of the patients were receiving growth or sex hormones prior to CPFT. If their cortisol level measured at 8:00 AM after TSA surgery was less than 3 μg/dL or if they exhibited symptoms related to cortisol deficiency, hydrocortisone was administered. However, hydrocortisone was discontinued at least 24 hours before CPFT was performed.

The flow of the study is depicted in Fig. 1. The CPFT examination was complex, and the procedure was delicate; therefore, only one subject was tested at a time, and a designated nurse with more than 5 years of experience in the endocrinology ward was in charge of the testing. For the CPFT, the subject was admitted to the ward by 5:00 PM the day before the examination. At 6:00 PM, the HRV device was mounted on the patient’s anterior chest, and the patient was allowed to ambulate without any limitations. At 10:00 PM, the subject was asked to go to sleep and to fast until the next day’s examination. The HRV device continuously recorded the patient’s HRV, and the data were observed closely at 3:00 AM (baseline), when the patient was sleeping and was considered to be in the most stable state. Staying at the hospital is an unusual experience for patients; hence, it may not be psychologically comfortable. This may be different when compared to a situation where the patient can stay comfortably at home. However, it is difficult to control the environment they would be exposed to at home; therefore, even if hospitalization causes anxiety, it was important for the patients to be in an environmentally controlled setting for the study. Preparation for the CPFT began at 6:00 AM on the day of examination, and hypoglycemia was induced between 6:30 AM and 7:00 AM according to the protocol. The period when hypoglycemia occurred was considered the most unstable state, and HRV data during this period were intensively observed. After the examination, the subjects were discharged at 10:00 AM, and HRV recording was terminated. After the HRV device was removed, all data collected from the patient were transferred to Kubios HRV software and analyzed.

Continuous data are expressed as either the mean±standard deviation or the median (interquartile range [IQR]). Categorical data are expressed as frequencies (proportions, %). A chi-square test or Fisher’s exact test was used for categorical data as appropriate. Continuous variables were compared between the two groups by Student’s t test or Mann-Whitney U test. Continuous data were evaluated for statistically significant differences using a paired t test or Wilcoxon signed-rank test as appropriate. Either a repeated-measures analysis of variance or a Friedman test was used to explore whether there were significant changes in HRV parameters, and a value of P<0.05 was considered statistically significant. Statistical analyses were performed using IBM SPSS Statistics software for Windows version 24.0 (IBM Corp., Armonk, NY, USA) and R software version 3.4.3 (R Foundation for Statistical Computing, Vienna, Austria).

Clinical characteristics at baseline are summarized in Table 1. The mean age of the subjects was 53.8±16.6 years. The serum glucose level after 8 hours of overnight fasting was 94.3±7.5 mg/dL, and the nadir of serum glucose during the CPFT was 37.7±4.0 mg/dL. The interval between surgery and the CPFT was 4.4 months (IQR, 3.2 to 5.9). The longest axis of the pituitary mass measured 25.9 mm (IQR, 13 to 43). The most commonly identified deficiency was Gn deficiency (FSH deficiency 59.1% and LH deficiency 54.5%), followed by ACTH deficiency (54.5%) and GH deficiency (50.0%). No single hormone deficiency was observed, and panhypopituitarism was observed in 12 patients (54.5%). The baseline LF/HF ratio was 2.9, and the SDNN was 58.3±5.2 ms.

The difference in HRV parameters for each pituitary hormone deficiency during sleep (the most stable period) is presented in Table 2. Although the difference was not significant, the HR was found to be low in the deficient group regardless of the type of hormone deficiency. In the deficient group, the HF, which decreases in the disease state, was lower, and the LF/HF ratio, which represents overall autonomic function and increases in the event of autonomic dysfunction, was higher than that of the preserved group. This pattern was consistently observed regardless of the type of hormone deficiency (Table 2).

Changes in HRV parameters for each pituitary hormone deficiency during the hypoglycemic state (the most unstable period) are summarized in Table 3. The HR was found to be low in the deficient group regardless of the type of hormone deficiency, but no significant difference was observed. HF and SDNN were lower in patients with impaired ACTH response than in those with normal ACTH response (HF: 87.5±39.9 in the deficient group vs. 193.7±55.2 in the preserved group, P=0.022; SDNN: 40.1±4.2 in the deficient group vs. 49.8±9.6 in the preserved group, P=0.030) (Table 3). The LF/HF ratio was significantly higher in the deficient group (6.9) than in the preserved group (1.9) (P=0.005) (Table 3). As shown in Fig. 2, in patients with impaired ACTH response (deficit group), HF (Fig. 2B), and SDNN (Fig. 2D) were significantly lowered, and the LF/HF ratio (Fig. 2C) was higher compared to the preserved group during both the stable state (sleep) and the unstable state (hypoglycemia), and significant differences were observed during the unstable state (hypoglycemia).

During hypoglycemia, changes in HRV parameters were analyzed according to the number of hormone deficiencies. HF and SDNN were significantly lower and the LF/HF ratio was significantly higher in patients with panhypopituitarism than in patients with multiple deficiencies (Table 4).