Complications of Transsphenoidal Surgery
Results of a National Survey, Review of the Literature, and Personal Experience
Ivan Ciric, M.D., Ann Ragin, Ph.D., Craig Baumgartner, P.A.-C., M.B.A., Debi Pierce, B.S., Division of Neurosurgery, Evanston Hospital, Northwestern University Medical School, Evanston, Illinois
The primary objectives of this report were, first, to determine the number and incidence of complications of transsphenoidal surgery performed by a cross-section of neurosurgeons in the United States and, second, to ascertain the influence of the surgeon's experience with the procedure on the occurrence of these complications. The secondary objective was to review complications of transsphenoidal surgery from the standpoint of their causation, treatment, and prevention.
Questionnaires regarding 14 specific complications of transsphenoidal surgery were mailed to 3172 neurosurgeons. The data reported were analyzed from the 958 respondents (82%) who reported performing the operation. The neurosurgeons surveyed were asked to estimate the number of transsphenoidal operations performed, to identify any complications observed, and to estimate the percentage of operations that had resulted in any of the 14 specific complications. The 958 respondents were placed into three experience groups, based on the number of transsphenoidal operations performed. The data were analyzed by using x2 tests and Spearman correlation coefficients. The secondary objectives were met through a detailed review of the literature, in light of our experience.
Of the respondents, 87.3% reported having performed <200 operations and 9.7% reported 200 to 500 previous operations. The remaining 3% reported more than 500 previous operations. More extensive previous experience with transsphenoidal surgery was associated with a greater likelihood of having witnessed each specific complication. The mean operative mortality rate for all three groups was 0.9%. Anterior pituitary insufficiency (19.4%)and diabetes insipidus (17.8%) were complications with the highest incidence of occurrence. The overall incidence of cerebrospinal fluid fistulas was 3.9%. Other significant complications, such as carotid artery injuries, hypothalamic injuries, loss of vision,and meningitis, occurred with incidence rates between 1 and 2%. An inverse relationship was found between the experience group and the likelihood of complications, as indicated by significant negative Spearman correlation coefficients for all but 2 of the 14 complications listed in the survey (P < 0.05). Thus, increased experience with transsphenoidal surgery seems to be associated with a decreased percentage of operations resulting in complications. Some caution should be exercised in interpreting these data, because they are based on the respondents estimates.
Transsphenoidal surgery seems to be a reasonably safe procedure, with a mortality rate of less than 1%. However, a significant number of complications do occur. The incidence of these complications seems to be higher, with statistical significance, in the hands of less experienced surgeons. The learning curve seems to be relatively shallow, because a statistically significantly decreased incidence of morbidity and death could be documented after 200 and even 500 transsphenoidal operations. Better understanding of the indications for transsphenoidal surgery and improved familiarity with the regional anatomy should further lower the incidence of death and morbidity resulting from this procedure in the hands of all neurosurgeons. (Neurosurgery 40:225-237, 1 997)
The transfacial approach to the sella was first described by Giordano (27) in 1897. That report was followed by other, similar,operative techniques, all requiring extensive facial soft tissue and osseous resection (24, 39). Among these, Schloffers operation, first described in 1907 (85), is the best known and most widely cited in the literature. In 1909, Kanavel (43)suggested a transnasal approach through a subnasal and lateral rhinotomy incision. The same year, Hirsch (37) introduced the endonasal technique, via an incision through the nares, thus avoiding lateral rhinotomy. Hirsch (38) continued to use this technique over a period of 5 decades, with excellent long-term results. The sublabial transnasal approach, a modification of Hirschs operation, was first suggested by Halstead (31) in 1910. Cushing (17, 18) adopted the sublabial-transseptal technique of Halstead and used it in more than 200 of his early pituitary tumor operations, with a mortality rate of 5.2%. The confined and poorly illuminated operative field and the risk of meningitis eventually limited the usefulness of the transsphenoidal technique, until it was completely abandoned by Cushing in favor of the transfrontal operation. The availability of corticosteroids and antibiotics paved the way for reintroduction of the transsphenoidal operation by Dott and Bailey (22) and Guiot and Thibant (30). It was not until Hardy (32) and Hardy and Wigser (36) began using the operating microscope, however, that the transsphenoidal technique gained wider acceptance. The transsphenoidal operation is today the principal surgical technique used by pituitary surgeons for removal of pituitary tumors and other lesions enlarging or otherwise involving the sella. During the past 3 decades, transsphenoidal surgery has proven effective and safe, with most authors reporting mortality rates between 0 and 1% (8, 13, 25, 26,52, 61, 71, 72, 77, 98, 101).
Nevertheless, the transsphenoidal operation can be a treacherous procedure, as evident from numerous case reports on various complications of this procedure (3, 4, 10, 11, 40, 60, 62,69, 75, 76, 79, 84, 88, 91). In addition, Barrow and Tindall (6), Black et al. (8), Landolt (47), Onesti and Post (72), Post et al.(77), Tindall et al. (90), Laws (48), and Laws and Kern (54-56) have published detailed reviews on complications of transsphenoidal surgery and treatment. However, with the exception of an international survey on the results of transsphenoidal surgery performed by experienced pituitary surgeons (101), there are no literature data on the cross-sectional morbidity and mortality rates for transsphenoidal surgery as it is practiced across the United States. With this in mind, we thought that it might be of interest to survey the neurosurgeons in the United States regarding their experience with this procedure.
Materials and Methods
Questionnaires regarding 14 specific complications of transsphenoidal pituitary surgery were mailed to 3172 neurosurgeons. The neurosurgeons surveyed were asked to estimate the number of transsphenoidal operations performed. They were further asked to identify any complication observed during the surgeons entire history with the operation. Finally, for those complications that had been witnessed, the neurosurgeons were asked to estimate the percentage of operations that had resulted in any of the specific complications. There were 1162 respondents(37%). Data were analyzed from the 958 respondents (82%) who reported performing transsphenoidal surgery. The 958 respondents were classified into three experience groups, based on the number of transsphenoidal operations performed. Of the respondents, 826(86.2%) reported having performed fewer than 200 transsphenoidal operations, and 88 (9.2%) reported 200 to 500 transsphenoidal operations. The remaining 27 (3%) reported experience with more than 500 transsphenoidal pituitary operations. Data regarding the number of reported complications in each experience group were analyzed by using the x2 test. The association between complications in the survey and experience was assessed with Spearman correlation coefficients.
|Table 1. Number of Surgeons Reporting Complications of Transsphenoidal Pituitary Surgery in the National Survey|
No. of Surgeons Reporting
|Carotid artery injury||114|
|Central nervous system injury||83|
|Hemorrhage/swelling of residual tumor||186|
|Loss of vision||179|
|Cerebrospinal fluid leak||590|
|Nasal septum perforation||327|
|Anterior pituitary insufficiency||563|
Of all respondents, 939 (98%) reported having witnessed at least one or more of the 14 complications represented on thesurvey (Table 1). Diabetes insipidus (DI) (748 respondents), cerebrospinal fluid (CSF) fistula (590 respondents), anterior pituitary insufficiency (563 respondents), and nasal septum perforation (327 respondents) were the most frequently reported complications. Meningitis and postoperative visual loss were reported by 192 and 179 of the respondents, respectively. Of the respondents, 114 reported having injured the carotid artery. Operative death was reported by 129 of the respondents.
Data regarding the number of surgeons reporting complications within each experience group are shown in Table 2. As expected,surgeons with more extensive experience with transsphenoidal surgery were more likely to have witnessed each specific complication. The x2 tests indicated statistically significant group differences (P < 0.05) for each complication.
Table 3 shows the percentage of operations resulting in the 14 specific complications for all of the respondents. The mean operative mortality rate was 0.9%. Carotid artery injuries occurred in 1.1% of all transsphenoidal operations reported. The overall incidence of CSF fistulas was 3.9%, with 1.5% of all operations being complicated by meningitis as well. A loss of vision was observed after 1.8% of the operative procedures reported. Anterior pituitary insufficiency (19.4%) and DI (17.8%)were complications with the highest incidence of occurrence.
Table 4 shows the percentage of operations resulting in the 14 specific complications for each of the three experience groups. Examination of the relationship between transsphenoidal surgical experience and the percentage of operations resulting in complications demonstrated an inverse relationship. Significant negative Spearman correlation coefficients were obtained for all but 2 of the 14 specific complications (P < 0.05) (Table 5). These findings suggested that increased transsphenoidal experience was associated with a smaller percentage of operations resulting in complications. Some caution should be exercised in interpreting these data, because they are based on the respondents estimates. Table 6 shows the Evanston Hospital experience with complications of transsphenoidal surgery.
|Table 2. Number of Surgeons, in Three Experience Groups, Reporting Complications of Transsphenoidal Pituitary Surgery in the National Survey|
|Complication||No. of Surgeons Reporting Complication||P|
|Carotid artery injury||74||25||15||<0.001|
|Central nervous system injury||51||18||14||<0.001|
|Hemorrhage/swelling of residual tumor||125||40||21||<0.001|
|Loss of vision||120||38||21||<0.001|
|Cerebrospinal fluid leak||483||81||26||<0.001|
|Nasal septum perforation||236||65||26||<0.001|
|Anterior pituitary insufficiency||465||73||25||<0.001|
|*Number of previous operations|
|Table 3. Percentage of Operations Resulting in Each Complication of Transsphenoidal Pituitary Surgery in the National Survey|
|Complicaton||% of Operations Resulting in Complication|
|Carotid artery injury||1.1|
|Central nervous system injury||1.3|
|Hemorrhage/swelling of residual tumor||2.9|
|Loss of vision||1.8|
|cerebrospinal fluid leak||3.9|
|Anterior pituitary insufficiency||19.4|
|*For all respondents (estimation by participating neurosurgeons)|
|Table 4. Percentage of Operations, in Three Experience Groups, Resulting in Each Complication of Transsphenoidal Pituitary Surgery in the National Survey|
|Complication||% of Operations Resulting in Complicationa|
|Carotid artery injury||1.4||0.6||0.4|
|Central nervous system injury||1.6||0.9||0.6|
|Hemorrhage/swelling of residual tumor||2.8||4.0||0.8|
|Loss of vision||2.4||0.8||0.5|
|Cerebrospinal fluid leak||4.2||2.8||1.5|
|Nasal septum perforation||7.6||4.6||3.3|
|Anterior pituitary insufficiency||20.6||14.9||7.2|
|a Estimation by participating neurosurgeons.
b Number of previous operations.
c NA, not applicable.
|Table 5. Association between Experience and Respondents Estimation of Percentage of Operations Resulting in Specific Complications|
|Carotid artery injury||-0.56||<0.001|
|Central nervous system injury||-0.33||<0.001|
|Hemorrhage into resiual tumor bed||-0.30||<0.001|
|Loss of vision||-0.51||<0.001|
|Nasal septum perforation||-0.16||<0.001|
|Anterior pituitary insufficiency||-0.16||<0.001|
|Table 6. Complications of Transsphenoidal Pituitary Surgery at the Evanston Hospital (638 Cases)|
|Complication||No. of Complications|
|Carotid artery injury||0|
|Central nervous sytem injury||0|
|Hemorrhage into residual tumor||4|
|Loss of vision||3|
|Cerebrospinal fluid leak||7|