Perioperative Outcomes of a Hydrocortisone Protocol after Endonasal Surgery for Pituitary Adenoma Resection

Abstract

Introduction  In pituitary adenomas (PAs), the use of postoperative steroid supplementation remains controversial, as it reduces peritumoral edema and sinonasal complaints but disrupts the detection of adrenal insufficiency (AI). It is unclear whether postoperative cortisol supplementation has a measurable effect on improving outcomes in patients with pituitary adenoma undergoing endoscopic transsphenoidal surgery (ETS). The objective of the study was to evaluate a postoperative steroid treatment protocol on various surgical outcomes in patients with PA undergoing ETS.

Methods  A retrospective cohort study was performed for patients undergoing ETS from 2005 to 2020 for PA at a single tertiary academic center. Patients were divided into two groups: those managed by a routine postoperative glucocorticoid supplementation protocol (steroid protocol) and those who received supplementation based on postoperative cortisol laboratory assessment (steroid sparing protocol). Management was otherwise the same between groups. Evaluation of length of stay (LOS), sinonasal outcomes, 30-day readmission, and perioperative complications, including AI, were performed.

Results  Among 535 patients, 21% ( n  = 111) received postoperative steroids, while the remainder ( n  = 424) did not. There were no differences in mean LOS (3 vs. 3 days, p  = 0.72), sinonasal complaints (27 vs. 19%, p  = 0.12), 30-day readmission (5% vs. 5%, p  = 0.44), and perioperative complications (5 vs. 5%, p  = 0.79) between both the groups. A multivariate model supported that both groups were comparable in predicting LOS, 30-day readmission, and complications. No reduction in readmission for AI was seen.

Conclusion  Routine administration of postoperative glucocorticoids did not significantly improve patient outcomes in patients with PA who underwent ETS.

Introduction

Pituitary adenomas (PAs) are one of the most common intracranial primary brain tumors, with an incidence of 4:100,000 per year and prevalence of 10 to 25% of the population. ^{1 2 3 4 5 6 7} Postoperative endocrinopathies occur in 10 to 30% of patients who undergo pituitary adenoma surgery but vary depending on tumor type, tumor morphology, and specific hormonal axis (e.g., gonadotrophic, corticotrophic, thyrotropic, or somatotrophic). ^{8 9 10} Moreover, recovery of function can be seen in 5 to 50% of patients also depending on tumor characteristics and the hormonal axis. ^{9 11 12} The measurement of cortisol to predict adrenal insufficiency and postoperative administration remains controversial in the treatment of pituitary adenomas.

Postoperative management of cortisol remains a critical component of pituitary adenoma surgery. Among hormonal deficits, postoperative adrenal insufficiency can be observed in 1 to 12% of patients with sellar masses and subsequent postoperative recovery can be seen in 15 to 20% of these patients. ^{9 11 13 14} In patients with normal preoperative endocrinological function, permanent and transient postoperative adrenal insufficiency can be seen in up to 9 and 18% of cases, respectively. ^{11 13 15} Measurement of early postoperative hypocortisolemia or hyponatremia to predict secondary adrenal insufficiency remains limited. ^{13 16 17 18 19 20 21 22 23} While several studies suggest an early postoperative serum cortisol cut-off level of 4 to 16 µg/dL to suggest a high risk for adrenal insufficiency, cortisol normalization can occur approximately 48 hours after surgery, even in cases when long-term steroid supplementation may still be needed. ^{13 16 17 18 19 20 22} Thus, early postoperative cortisol levels (within 48 hours of surgery) may not be able to predict late or sustained postoperative adrenal insufficiency. Similarly, insulin tolerance, short synacthen, cortrosyn stimulation, and metyrapone testing may not be accurate in the postoperative setting and no clear guidelines exist on the diagnosis and management of postoperative adrenal insufficiency. ^{13 17}

The administration of postoperative steroids remains controversial. ^{10 13 24} Patients with preoperative hypocortisolemia may develop electrolyte disturbances, increased postoperative pain, as well as hemodynamic instability. ¹⁶ However, the indiscriminate use of postoperative steroids can impact osteopenia, weight gain, cardiovascular disease, and steroid psychosis, even with short treatment courses. ^{25 26 27 28} One study also suggested an increased rate of mortality in patients with adrenal insufficiency and high hydrocortisone replacement doses. ²⁹ Studies evaluating the efficacy of postoperative steroid use after pituitary adenoma surgery have been limited by small sample sizes and a lack of clear correlation of steroid use and outcomes. ^{30 31 32 33 34} More work is needed to explore the role of a postoperative steroid regimen, specifically in patients with normal preoperative endocrine function, undergoing pituitary adenoma surgery. Furthermore, additional research specifically in this population is needed to better establish when glucocorticoid supplementation should be administered to prevent adrenal insufficiency and steroid complications.

At our institution, despite similarities in intraoperative and postoperative pituitary adenoma management, postoperative glucocorticoid supplementation continues to show surgeon variation. We compared a postoperative steroid supplementation protocol and a steroid sparing protocol on a large series of patients that underwent endoscopic transsphenoidal surgery (ETS) for pituitary adenomas. The benefits of early steroids (e.g., improvement in sinonasal outcomes, earlier discharge from improved pain control) were compared with the drawbacks (e.g., delayed diagnosis of adrenal insufficiency and perioperative complications).

Materials and Methods

Study Design

We conducted a retrospective cohort study of a prospectively maintained database including all patients who underwent ETS for a pituitary adenoma between January 2005 and December 2020. Patient demographic and clinical variables were analyzed. Primary tumors were distinguished from recurrent adenomas. Tumor size was defined by maximum length multiplied by width and by one half the height. Invasive tumors were defined as those that entered the cavernous sinus, thus making gross total resection less likely. Measured outcomes included length of stay, sinonasal complaints, 30-day readmission rate, and various complication rates, including rates of adrenal insufficiency. Sinonasal outcomes test (SNOT-22) scores, a validated test of patient quality of life following sinonasal surgery, were evaluated as well.

Patients were grouped based on whether or not they were placed on a postoperative steroid protocol with all other aspects of care remaining similar. Patients included on the steroid protocol were specifically under one neurosurgeon attending provider (C.F.) while patients on the steroid sparing protocol were under a different provider (J.J.E.) despite similarity, otherwise with patient management. Patients with Cushing's disease or with preoperative steroid use were excluded from the study. The steroid postoperative protocol involved 4 mg of dexamethasone at time of surgery followed by 0.5 mg daily for 5 days with measurement of cortisol and sodium levels at day 6. Patients not on the steroid protocol underwent measurement of a morning cortisol on postoperative days 1 and 2 if they stayed in the hospital long enough, as well as 1 week postoperatively. Only patients with an AM cortisol level below 7 µg/dL were started on maintenance steroid replacement. A follow-up cortisol level was acquired as an outpatient 1 week postoperatively after stopping the steroid replacement 5 days postoperatively, and patients with low values were referred to endocrinology.

Surgical resection was performed by an endoscopic approach which has been previously described. ³⁵ Briefly, a unilateral sphenoidotomy with contralateral partial sphenoidotomy was used along with routine intraoperative navigation. Surgical repair involved a combination of synthetic dural substitute for no leak or low-flow leaks with the addition of a nasoseptal flap for high-flow leaks. ^{35 36} Postoperative neurological monitoring was performed in an intermediate surgical step down unit and patients followed-up in 1 week in neurosurgery clinic and in otolaryngology clinic for nasal debridement. Follow-up sodium levels were obtained during hospitalization and 1 week postoperatively. Follow-up magnetic resonance images (MRIs) were performed at 3 months postoperatively.

Complete resection was based on intraoperative inspection of the sella by the surgical team after resection and by 3-month follow-up imaging. Tumor invasion was defined as involvement of bone or dura. Sinusitis was included as a postoperative complication only if the patient required intervention in the form of antibiotics/surgery. Epistaxis was defined as any nasal bleed that required intervention in the form of cauterization.

Statistical Analysis

Descriptive analysis using means or percentage of total are presented where appropriate. Following bivariable descriptive analyses, patients were analyzed in multivariable linear and logistic regressions. Data were analyzed using statistical software (RStudio v1.3.1093) and a p  < 0.05 was considered significant. The STROBE checklist was used in the preparation of this manuscript ( https://www.strobe-statement.org/index.php?id=available-checklists ).

Results

Demographics and Tumor Differences

Among a total of 535 patients, 21% were placed on a steroid protocol ( n  = 111). Age and gender distribution was comparable between the two cohorts ( Table 1 ). Patients on the steroid protocol had a significantly lower mean body mass index (BMI; 26.7 vs. 31.6, p  < 0.01), as well as lower rates of hypertension (35 vs. 48%, p  = 0.02) compared with patients on the steroid sparing protocol. No other differences in patient comorbidities (smoking, myocardial infarction, congestive heart failure, atrial fibrillation, pulmonary embolism, and congestive obstructive pulmonary disorder) were observed.

Table 1. Demographics, preoperative comorbidities, and tumor characteristics.

	Steroid protocol ( n  = 111) n (%)	Steroid sparing protocol ( n  = 424) n (%)	p -Value
Age (y) Mean (SD)	55 (15)	55 (15)	0.78
Gender (female)	53 (48)	205(48)	1
Body mass index (kg/m 2 ) mean (SD)	26.7 (5.4)	31.6 (6.6)	<0.01 a
Tobacco smoking,	40 (36)	158 (37)	0.94
Hypertension	39 (35)	202 (48)	0.02 a
Myocardial infarction	2 (1.8)	11 (2.6)	0.89
Congestive heart failure	0	6 (1.4)	0.45
Atrial fibrillation	5 (4.5)	12 (3)	0.61
Pulmonary embolism	2 (1.8)	14 (3.3)	0.61
Chronic obstructive pulmonary disease	3 (3)	9 (2)	0.99
Tumor characteristics
Macroadenoma	104 (94)	394 (93)	0.94
Primary	98 (88)	371 (86)	0.95
Complete resection	87 (78)	354 (83)	0.22
Apoplexy	9 (9)	31 (7)	0.94
Intraoperative CSF leak	38 (34)	103 (24)	0.05
Size (cm 3 )  Mean (%)	2.38 (1.12)	2.51 (1.15)	0.3
Invasive tumor	39 (35)	142 (33)	0.78

Abbreviations: CSF, cerebrospinal fluid; SD, standard deviation.

^aStatistically significant.

Most tumor features were similar in both groups ( Table 1 ). A greater proportion of patients who were in the steroid protocol group had intraoperative leak, although not statistically significant (34 vs. 24%, p  = 0.05). There were otherwise no differences in the proportion of primary resections, complete resections, macroadenomas, operations due to apoplexy, tumor size, nor invasiveness. The mean size of tumors for those on steroid protocol was 2.38 cm ³ and for those on steroid sparing protocol was 2.51 cm ³ ( p  = 0.3). Eventually, 90% of the patients placed on steroid sparing protocol did not require postoperative glucocorticoid administration, and only a minority (10%) received steroids based on low postoperative cortisol levels.

Perioperative Outcomes

There were no differences in mean length of stay (3 vs. 3 days, p  = 0.72), 30-day readmission (5 vs. 5%, p  = 0.44), and perioperative complications (epistaxis, meningitis, pulmonary embolism, sellar hemorrhage, and thromboembolism; 5 vs. 5%, p  = 0.79) between both the groups ( Table 2 ). Readmission within 30-days was most commonly for hyponatremia ( n  = 6 total), adrenal insufficiency ( n  = 3 total), headache ( n  = 7 total), treatment of epistaxis ( n  = 7 total), or repair of cerebrospinal fluid (CSF) leak ( n  = 6). Although readmission for postoperative adrenal insufficiency occurred in 3 of 424 (0.7%) of the steroid sparing cohort, there was no difference between the steroid and steroid sparing protocols. Furthermore, patients readmitted for adrenal insufficiency were managed safely. Other readmission reasons included residual tumor ( n  = 1 total), infection ( n  = 3 total), as well as postoperative diabetes insipidus ( n  = 1 total). No readmissions for hyperglycemia or steroid psychosis were observed. There were no changes in postoperative sinonasal outcomes based on mean postoperative SNOT-22 scores taken 1 week postoperatively between the steroid and steroid sparing groups (22 ± 24 vs. 16 ± 16, respectively, p  = 0.12). A multivariate model showed that both groups were comparable in terms of length of stay, 30-day readmission, or complications after adjusting for other patient and tumor characteristics ( Table 3 ).

Table 2. Perioperative outcomes.

Measures	Steroid protocol ( n  = 111)	Steroid sparing protocol ( n  = 424)	p -Value
Length of stay (d) Mean (SD)	3 (2)	3 (2)	0.72
Postoperative day 1 cortisol Mean (SD)	39.7 (32.8)	33.3 (23.0)	0.17
30-day readmission n (%)	7 (5)	25 (6)	0.87
Headache ( n )	0	7
CSF leak ( n )	2	4
Hyponatremia ( n )	1	5
Adrenal insufficiency ( n )	0	3
Other ( n ) a	4	6
Postoperative sinonasal complaints n (%)	21 (19)	125 (27)	0.12
Septal perforation ( n )	2	9
Sinusitis ( n )	16	70
Synechiae ( n )	3	37
New pan-hypopituitarism n (%)	2 (2)	1 (<1)	0.11
Postoperative SNOT score Mean (SD)	22 (24)	16 (16)	0.12
Perioperative complications n (%)	6 (5)	20 (5)	0.79
Epistaxis ( n )	2	5
Meningitis ( n )	0	0
Pulmonary embolism ( n )	2	3
Sellar hemorrhage ( n )	2	2
Thromboembolism ( n )	0	10

Abbreviations: SD, standard deviation; SNOT, sinonasal outcomes test.

^aRelated causes included dyspnea, UTI, major depressive disorder, septicemia, bowel resection, residual tumor, infection, and postoperative diabetes insipidus. Unrelated causes of readmission included major depressive disorder, septicemia, and bowel resection.

Table 3. Multivariate regression.

	Complications		Length of stay		30-day readmission
	OR a	p -Value	Coefficient a	p -Value	OR a	p -Value
Steroid sparing protocol (ref. standard protocol)	0.99 (0.02)	0.67	0.24 (0.28)	0.38	0.99 (0.03)	0.73
Age (y)	1 (<0.01)	0.18	0 (0.01)	0.83	1 (<0.01)	0.35
BMI	1 (<0.01)	0.94	−0.03 (0.02)	0.1	1 (<0.01)	0.59
Hypertension (ref. no HTN)	0.99 (0.01)	0.71	−0.12 (0.22)	0.58	0.99 (0.03)	0.8
Intraoperative CSF leak (ref. no CSF leak)	1.01 (0.02)	0.52	0.1 (0.26)	0.69	1.12 (0.03)	<0.01 b
Size (cm 3 )	0.99 (0.01)	0.19	0.7 (0.10)	<0.01 b	1 (0.01)	0.69
Primary tumor (ref. recurrent tumor)	0.97 (0.03)	0.25	0.29 (0.33)	0.37	0.95 (0.04)	0.22

Abbreviations: BMI, body mass index; CSF, cerebrospinal fluid; HTN, hypertension; OR, odds ratio; ref., reference.

^aStandard error in parentheses.

^bStatistically significant.

Discussion

Impact of Postoperative Steroids

Our results demonstrated no improvement in patient outcome nor reduction in postoperative complications in patients placed on a postoperative steroid sparing protocol for the treatment of pituitary adenomas. This included no difference in length of stay, 30-day readmission, sinonasal complaints, or complication profile.

Perioperative Adrenal Insufficiency and Hydrocortisone Treatment

Several other studies on pituitary adenomas have aimed to evaluate the use of postoperative steroid protocols to reduce adverse events from adrenal insufficiency and predict long-term adrenal insufficiency. ^{10 13} The detection and interpretation of hypocortisolemia, as well as medication treatment and follow-up, vary depending on the institution.

A meta-analysis of 18 studies recently published in 2015 challenged the necessity of steroids for pituitary adenoma surgery. ¹³ A total of 1,224 patients were evaluated in studies from 1987 to 2013. Among 739 patients with sufficient data, there were 41 events (5.55%) of postsurgical adrenal insufficiency (range: 0.96–12.9%) during postoperative days 2 to 14. Regarding diabetes insipidus, 235 patients were evaluated with no significant difference in incidence between patients supplemented with steroids versus those not supplemented ( p  = 0.8). Considering the predictive value of postoperative cortisol levels, various cut-offs were included ranging from 4 to 16 µg/dL on postoperative day 1 or day 2 with a sensitivity ranging from 50 to 100% and specificity from 47 to 98%. One of the challenges in identifying adrenal insufficiency involves the somewhat subjective clinical definition. The same meta-analysis suggested the clinical picture of adrenal insufficiency, including “symptoms of fatigue, loss of appetite, nausea, vomiting, arthralgias, and, in some instances, hypotension and electrolyte abnormalities,” which may often resemble expected postoperative recovery in any patient. In addition, differing postoperative cortisol values and cut-offs delineated adrenal insufficiency in individual studies. The use of postoperative adrenal axis testing was also heterogeneous as was the criteria for stopping steroid supplementation in patients deemed not to have adrenal insufficiency. The impact of steroids on postoperative adrenal insufficiency is unclear partly due to the lack of clarity in understanding factors that impact insufficiency in the postoperative period. Perioperative steroids, such as hydrocortisone, can be picked up by the assay creating an artificial elevation and steroids not picked up by the assay, such as dexamethasone, can cause suppression of the AM cortisol.

Significant variation exists in the types and dosages for cortisol supplementation. For patients with adrenal insufficiency, Kristof et al recommended 100 mg of hydrocortisone on the day of the operation, 80 mg on day 1, 60 mg on day 2, 50 mg on day 3, 35 mg on day 4, 25 mg on day 5, and 15 to 25 mg on days 6 to 10 depending on clinical assessments. ³⁷ Tohti et al suggested that multiple institutions used the protocol of Inder and Hunt. ¹³ This protocol recommended hydrocortisone 50-mg every eight hours on day 0, 25-mg every eight hours on day 1, and 25-mg q8 hours on day 2 or dexamethasone 4 mg on day 0, 2 mg on day 1, and 0.5 mg on day 2, followed by reassessment of cortisol on days 3 to 5 to determine the need for further replacement or testing. Postoperative cortisol axis testing was recommended, but it was unclear what postoperative cortisol values triggered testing. A recent survey by Pecoraro et al suggested a wide variation among a group of academic skull base surgeons, using dexamethasone, hydrocortisone, or solumedrol for pituitary adenoma postoperative management. ²⁴ Furthermore, this group suggested a low dose of 50 mg of hydrocortisone intraoperatively with cortisol remeasurement on postoperative day 2 and week 6 that was sufficient to adequately treat patients and identify adrenal insufficiency. This study also showed reduced long-term steroid use in a lower dose protocol but ultimately lacked sufficient sample size and control patients.

Prior studies have not aimed to evaluate the impact of postoperative steroids on length of stay, 30-day readmission, sinonasal complaints, or complications. Our results are one of the first to suggest that routine administration of corticosteroids does not significantly improve perioperative outcomes. We did not specifically evaluate adrenal insufficiency, as patients were managed by a variety of endocrinologists and no steroid protocol for evaluation was used. However, adrenal insufficiency as a cause of 3-day readmission was not statistically reduced in patients who received steroids. The majority (90%) of patients in this study who were on the steroid sparing protocol also did not require any postoperative steroids. These outcomes suggest that routine postoperative steroids do not improve patient outcomes after ETS and are not necessary during routine perioperative care.

Limitations

Limitations of this study include its retrospective nature, as well as additional variability in decision-making for cortisol retesting, and steroid weaning. Multiple endocrinologists were involved in the outpatient decision-making for when to wean steroids in patients. While this represents a more real-world clinical situation, it makes interpretation of the hypothalamic–pituitary–adrenal axis integrity more challenging. Additionally, it is important to note that differences in surgical experience may impact endocrinological outcomes and management. There may be differences in attendings impacting overall complications, but this is also unlikely as both surgeons are experienced practitioners. One factor that could not be assessed was the patient pain scores and other postoperative symptoms (e.g., nausea and vomiting) that could either have been improved or exacerbated with steroid use. While the SNOT-22 scores provide a sinonasal quality of life assessment, they are not an ideal surrogate for mild adrenal insufficiency symptoms. Equally, the impact of steroids on other outcomes (e.g., bone density, hyperglycemia, insomnia, psychosis, and mortality) could not be fully assessed. Hence, large scale prospective studies are required to establish the guidelines. These are areas of active research and future studies can look at glucocorticoid replacement. Nevertheless, this is the first study to evaluate the impact of postoperative steroids on various outcomes including length of stay, 30-day readmission, sinonasal outcomes, and complications.

Conclusion

In patients who underwent endoscopic transsphenoidal surgery for pituitary adenoma, there were no differences in outcomes between those on the steroid postoperative steroids protocol and those on the postoperative steroid sparing protocol. This suggests that the routine administration of glucocorticoids may not be necessary following surgery for pituitary adenoma.