Cerebrovascular complications and utilization of endovascular techniques following transsphenoidal resection of pituitary adenomas: a study of the Nationwide Inpatient Sample 2001–2010
- First Online:
- Cite this article as:
- Brinjikji, W., Lanzino, G. & Cloft, H.J. Pituitary (2014) 17: 430. doi:10.1007/s11102-013-0521-1
- 168 Views
Cerebrovascular complications following transsphenoidal resection of pituitary tumors are rare and often evaluated and treated with endovascular techniques. We determined the utilization rate and outcomes of endovascular procedures in transsphenoidal pituitary resection patients using an administrative database.
Using the Nationwide Inpatient Sample 2001–2010, patients receiving transsphenoidal resection of benign pituitary tumors were identified. The rate of cerebrovascular complications and utilization of endovascular repair procedures and cerebral angiography were compared between high (≥75 procedures/year) and low volume (<75 procedures/year) centers. Chi squared tests were used to compare categorical variables.
70,878 were patients included in this study. ICH/SAH occurred in 0.9 % of patients (652/70,878) and stroke occurred in 0.5 % of patients (327/70,878). Patients treated at high volume centers had significantly lower rates of stroke (0.5 % vs. 1.0 %, P = 0.04), and ICH/SAH (0.5 vs. 1.0 %, P = 0.05) when compared to patients treated at low-volume centers. Overall, 531 patients (0.7 %) received post-operative angiography and 83 patients (0.1 %) received endovascular repair procedures. High volume center patients underwent angiography in 0.4 % of cases compared to 0.9 % for low volume center patients (P = 0.02). There was no significant difference in endovascular repair procedure rates at high and low volume centers (0.1 vs. 0.2 %, P = 0.37).
Cerebrovascular surgical complications requiring cerebral angiography and endovascular repair are rare among transsphenoidal pituitary resection patients. These occur with higher frequency at low volume centers and are associated with high mortality rates.
KeywordsPituitary adenomaTranssphenoidal resectionEndovascularHospital outcomesHemorrhage
Hemorrhagic and vascular complications are rare but serious complications for patients undergoing transsphenoidal resection of pituitary tumors . Several cases of carotid cavernous fistulas and carotid pseudoaneurysm formation have been reported after transsphenoidal resection of skull base tumors [2–4]. Cerebral angiography and endovascular therapeutic techniques have emerged as the methods of choice for evaluation and treatment of carotid injury in the post-operative setting [2–6]. We examined a large administrative database to evaluate the incidence of cerebrovascular complications and the utilization of post-operative cerebral angiography and endovascular treatment techniques in patients undergoing transsphenoidal resection of pituitary adenomas. We also studied the relationship between hospital procedural volume and rate of cerebrovascular complications as well as utilization rates of endovascular procedures.
We purchased the Nationwide Inpatient Sample (NIS) hospital discharge database for the period 2001–2010 from the Healthcare Cost and Utilization Project of the Agency for Healthcare Research and Quality, Rockville, MD. The NIS is a hospital discharge database that represents 20 % of all inpatient admissions to nonfederal hospitals in the United States. Patient procedures and diagnoses were identified using both International Classification of Diseases 9th Revision (ICD-9). Patients who underwent transsphenoidal resection of benign pituitary adenomas were identified using ICD-9 procedure codes 762 (excision pituitary lesion-transsphenoidal) and 765 (total excision pituitary-transsphenoidal) and then crossmatched with patients having the ICD-9 diagnostic code 2273 (benign neoplasm of the pituitary).
Outcomes studied included hospital discharge status (in-hospital mortality, discharge to long-term facility and discharge to short-term facility), length of stay, cerebrovascular complications (stroke, ICH/SAH), other post-operative CNS complications, and other hemorrhagic complications. We also studied the utilization of cerebral angiography and/or endovascular repair techniques. Patients undergoing cerebral angiography were identified using ICD-9 procedure codes 0061 (percutaneous angiography precerebral vessels), 0062 (percutaneous angiography intracranial vessels) and 8841 (contrast cerebral arteriogram). Patients undergoing endovascular repair of damaged vessels were identified using ICD-9 procedure codes 0064 (percutaneous insertion of precerebral stent), 0065 (percutaneous insertion of intracranial stent), 3972 (endovascular repair of head vessel) and 3979 (endovascular repair of other vessel).
We compared the rates of the above outcomes in high volume and low volume centers. Hospitals were divided into high volume hospitals (those performing ≥75 transsphenoidal pituitary surgeries/year) and low volume hospitals (those performing <75 transsphenoidal pituitary surgeries/year). The cutoff of 75 cases was defined as high volume as this was the threshold at which 90 % of patients were treated at centers treating less than 75 cases per year and 10 % were treated at centers treating greater than 75 cases per year.
Outcomes and utilization of endovascular procedures
When studying outcomes and utilization of endovascular procedures, patients were divided into three groups (1) uncomplicated patients (i.e. patients not receiving post-surgical endovascular treatment or evaluation), (2) patients undergoing post-surgical angiography and (3) patients undergoing post-surgical endovascular repair. For each group, data on demographic and comorbidity variables including age, gender, race (white, non-white) and Charlson Comorbidity Index were collected. For angiography and endovascular repair patients the likely indication for the procedure was obtained using ICD-9 codes. Indications studied included acute ischemic stroke, ICH/SAH, other post-operative bleeding complications and post-operative CNS complications. Post-operative angiography patients were identified using “number of days from admission to procedure” data from the NIS dataset.
For comparison of categorical variables, Chi squared testing was performed. For comparison of continuous variables, student’s t test was performed. With length of stay, the median value as well as the interquartile range (IQR) were reported. Medians were compared using the non-parametric median test. We performed a multivariate logistic regression analysis to determine independent predictors of endovascular procedure utilization. Variables included in this analysis were age, gender, race (white versus non-white), co-morbidities and hospital volume (high versus low). Discharge weights were applied except when studying utilization rates by hospital volume. All statistical analyses were performed using the SAS-based statistical software package JMP 9.0 (www.jmp.com).
Patient population, complications and discharge status
From 2001–2010, 70,878 patients underwent transsphenoidal resection of benign pituitary tumors. Mean age was 51.1 ± 16.4 years. 35,828 patients (51.3 %) were female and 33,880 patients (48.6 %) were male. 33,624 patients (62.9 %) were white. Median length of stay for patients undergoing transsphenoidal resection of benign pituitary tumors was 3 days (IQR = 2–5 days). Overall in-hospital mortality rate was 0.4 % (298/70,878). Discharge to long-term facility rate was 2.9 % (2,075/70,878) and discharge to short-term facility rate was 0.3 % (201/70,878). ICH/SAH occurred in 0.9 % of patients (652/70,878). Other bleeding complications occurred in 1.3 % of patients (943/70,878). Stroke occurred in 0.5 % of patients (327/70,878) and other post-operative CNS complications occurred in 5.9 % of patients (4,152/70,878).
Complications and discharge status by hospital volume
Outcomes between high and low volume centers
Low volume N (%)
High volume N (%)
Discharge to long-term facility
Discharge to short-term facility
Median (IQR) length of stay
Other bleeding complication
Post-operative CNS complications
Endovascular procedure utilization in pituitary patients
Of the patients included in this study, 70,304 (99.1 %) were uncomplicated, 531 patients (0.7 %) underwent post-operative angiography and 83 patients (0.1 %) underwent an endovascular repair procedure. 259 patients (0.4 %) underwent pre-operative angiography. Among post-operative angiography patients, the two most common indications were acute ischemic stroke (N = 95, 17.9 %) and ICH/SAH (N = 93, 17.5 %). Among endovascular repair patients, the most common indication was ICH/SAH (N = 34, 40.6 %) and stroke (N = 33, 39.9 %). The overall incidence of acute ischemic stroke was 1.0 %, the overall incidence of SAH/ICH was 0.9 %, and the overall incidence of post-operative CNS complications was 5.8 %.
Demographic characteristics and endovascular utilization rates
Mean age (SD)
N (%) Female
N (%) White
Mean CCI (SD)
Endovascular procedure indications
Post-operative CNS Complications
Hospital volume and endovascular technique utilization
Outcomes by group
Patients requiring endovascular repair procedures had significantly higher mortality rates than uncomplicated patients (11.5 vs. 0.4 %, P < 0.0001). Angiography patients also had significantly higher mortality rates (5.6 vs. 0.4 %, P < 0.0001). Median length of stay for uncomplicated patients was 3 days (IQR = 2–5 days). Median length of stay for patients receiving post-operative angiography was significantly higher (6 days, IQR = 4–12 days, P < 0.0001). The same was true for endovascular repair patients (13 days, IQR = 4.5–23 days, P < 0.0001).
OR (95 % CI)
This study confirms the notion that significant morbidity, mortality and vascular complications are low among patients undergoing transsphenoidal resection of benign pituitary tumors . The rate of stroke in our sample was only 0.5 % and the rate of ICH/SAH was 0.9 %. Post-operative angiography utilization in our study was only 0.7 % and the rate of endovascular repair procedures was only 0.1 %. As expected, patients requiring angiography and endovascular repair procedures had significantly longer lengths of stay and significantly higher mortality rates when compared to uncomplicated patients.
Our study also demonstrated that patients treated at low volume centers have significantly higher rates of discharge to facilities other than home, cerebrovascular complications and angiography utilization when compared to those treated at high volume centers. The rate of endovascular repair however was similar between the two groups. These findings suggest that patients treated at high volume centers experience significantly lower rates of vascular complications when compared to those treated at high volume centers. These findings are consistent with the large body of literature that demonstrate improved outcomes and complication rates for neurosurgical patients treated at high volume centers [8–10]. In our study, 90 % of patients were treated at centers performing <75 cases per year. Given the better outcomes seen in patients treated at higher volume centers, our findings suggest that referral of patients to high volume centers could lead to improved outcomes overall.
Rates of carotid injury among patients undergoing transsphenoidal skull base procedures range between 0.2 and 1.1 % with lower rates in series using endoscopy assistance [11, 12]. Gardner et al. demonstrated an ICA injury rate of 0.3 % in their series of 2,015 patients treated with endoscopic endonasal transsphenoidal surgery. This study also demonstrated that patients with more extensive malignant skull base tumors had higher rates of carotid injury . Of the patients who suffered carotid artery injury, three underwent endovascular repair procedures . In a single surgeon series of 300 endoscopic endonasal transsphenoidal surgeries, Mamelak et al.  reported two carotid artery injuries (0.7 %). Fatemi et al. showed 4 carotid artery injuries out of 881 transsphenoidal surgeries (0.2 %) and found lower complication rates among patients receiving endoscopy. All four patients underwent immediate angiography with one patient receiving balloon embolization and one patient receiving coil embolization for treatment of cavernous carotid pseudo aneurysms [12, 15]. In a survey of neurosurgeons before the widespread use of endoscopy, Ciric et al.  found a carotid artery injury rate of 1.1 %. Raymond et al. found 17 patients with carotid injury out of 1,800 patients (~1 %). Similar to our study, Raymond et al.  reported that angiography was used in 1.1 % of transsphenoidal surgical patients and intervention was performed in 0.3 % of patients. Carotid injury was reported in approximately 1 % of patients .
Angiographic findings after pituitary surgery include carotid occlusion, stenosis, carotid-cavernous fistulas, pseudo aneurysms, ICH/SAH and epistaxis. Such complications are generally treated with covered stent implantation, endovascular coiling, stent assisted coiling and embolization [2–6]. Patients experiencing these complications suffer high rates of neurological deficits and mortality. In a series of nine patients undergoing angiography for suspected vascular injury following transsphenoidal surgery, Cinar et al.  reported a mortality rate of 11.1 % and a significant morbidity rate of 11.1 %. In a series of 17 patients receiving endovascular treatment following carotid injury, Raymond et al.  reported five permanent neurologic deficits (29.4 %) and two deaths (18.2 %). Similar to these two prior studies, our sample of patients receiving endovascular treatment following transsphenoidal surgery demonstrated a mortality rate of approximately 11 %. Such high morbidity and mortality rates highlight the importance of understanding the relationship between the cavernous carotid artery and pituitary is important to avoid complications [17, 18].
There are a number of limitations to the use of the NIS database, largely stemming from the lack of important clinical data that could provide a more detailed explanation of outcomes. Using the NIS, we are unable to determine how successful endovascular procedures were in reducing the morbidity and mortality associated with carotid injury. We were unable to obtain any information on the use of endoscopy during transsphenoidal pituitary resection as these data are not provided in this database. Endoscopy is emerging as a safer technique that is associated with lower complication rates [13, 19]. A majority of patients undergoing angiography did not have associated diagnoses of stroke, ICH/SAH or other hemorrhagic complications. It is possible that many of these patients underwent angiography to evaluate less severe bleeding that did not originate from the carotid artery. Bleeding without carotid injury could be due to damage to smaller arteries like MMA branches or the inferior hypophyseal artery [20, 21]. Coding errors are also a potential limitation of this study as they are with any study of a large administrative database.
Our study of a large nationwide administrative database demonstrated a very low rate of cerebrovascular complications and a very low utilization rate of cerebral angiography and endovascular repair procedures among patients undergoing transsphenoidal resection of the pituitary gland. Patients treated at higher volume centers had significantly lower rates of cerebrovascular complications as well as utilization rates of endovascular procedures than those treated at low volume centers, suggesting lower rates of carotid injury at high volume centers. Patients who required such procedures suffered significantly higher mortality rates and significantly longer lengths of stay than their uncomplicated counterparts. These findings highlight the importance of careful planning of transsphenoidal surgery so as to avoid serious vascular complications.
Conflict of interest
W.B. and G.L.: None; H.J.C: Site PI at enrolling site for SAPPHIRE (Stenting and Angioplasty with Protection in Patients and High Risk for Endarterectomy) registry sponsored by Cordis Endovascular.