| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the Departments of *Anesthesiology and Neurosurgery, University of Virginia Health System, Charlottesville.
Address correspondence and reprint requests to Edward C. Nemergut, MD, Assistant Professor of Anesthesiology and Neurosurgery, University of Virginia Health System, P.O. Box 800710, Charlottesville, VA 22908. Address e-mail to en3x{at}virginia.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The transsphenoidal approach to lesions of the sella turcica, and especially to pituitary tumors, has been favored by neurosurgeons because of decreased perioperative patient morbidity and mortality of this approach compared with others (12,13). Indeed, at academic medical centers as many as 20% of all primary brain tumors are approached transsphenoidally (12,14). Despite the popularity of transsphenoidal surgery, there has been no systematic review of PONV, pain, or associated risk factors. We conducted a retrospective review to provide insight into the mechanisms of PONV and postoperative pain in the transsphenoidal patient.
|
METHODS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Clinical diagnosis in all cases was made using appropriate biochemical testing and confirmed after surgery with histopathology. Tumors larger than 10 mm in any dimension on magnetic resonance imaging were classified as "macroadenomas," whereas tumors smaller than 10 mm were classified as "microadenomas."
For the purpose of this review, vomiting was defined as the forceful expulsion of gastric contents in the postanesthesia care unit (PACU) or operating room (OR). Patients were considered to have vomited if vomiting was noted on the PACU record or anesthesia record. Any antiemetic given in the OR while the patient was still anesthetized was considered to be prophylactic therapy. After emergence from anesthesia, any patient given an antiemetic drug after vomiting was considered to have received rescue treatment specifically for vomiting. For the purpose of this review, nausea was defined as an unpleasant sensation associated with the urge to vomit. Accordingly, any patient given an antiemetic drug after complaining of nausea (without vomiting) was considered to have received rescue treatment specifically for nausea.
Any opioid given after the emergence from anesthesia we considered to have been given to treat postoperative pain. In this study, only morphine consumption in the PACU is considered to be indicative of opioid requirement for pain management. To facilitate review and comparison, opioid consumption was converted to morphine equivalents using standard conversion ratios.
Our institution has a well established protocol for monitoring fluid balance and the detection of diabetes insipidus (DI) during the postoperative period. For the purpose of this review, a patient was considered to have clinical evidence of DI if the diagnosis appeared on the discharge summary by the primary neuroendocrine team and the patient had documented voluminous urine output (more than 300 mL/h) for more than 3 h with a specific gravity of <1.005. Increasing serum sodium is considered as supportive evidence for the diagnosis of DI, but it is not required to make the diagnosis or initiate treatment.
Statistical analysis was performed using a z-test, t-test, or Mann-Whitney U-test as indicated. Comparisons were made using two-tailed tests and a P < 0.05 was considered significant.
|
RESULTS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
There was also a significant increase in vomiting among patients who required a lumbar drain (17.1% versus 6.6%; P < 0.001; Table 2). Notably, patients who had a lumbar drain were more likely to receive intraoperative prophylaxis (42% versus 31%; P = 0.047) and, despite that, vomited more often. Among patients who received a lumbar drain, intraoperative antiemetic prophylaxis did not reduce the incidence of vomiting in the PACU (21% versus 13%; P = 0.346).
|
Examining clinical disease, only a craniopharyngioma was significantly associated with an increased incidence of vomiting in the PACU when compared with patients with nonfunctioning adenomas (Table 3). Patients with a craniopharyngioma were more likely to receive intraoperative prophylaxis (52.6% versus 31.7%; P = 0.007) and, despite this, vomited more often. Among patients with a craniopharyngioma, those who received intraoperative antiemetic prophylaxis (15% versus 22.2%; P = 0.566) did not have a reduction in the incidence of vomiting in the PACU.
|
Patient sex or age, preoperative midazolam, and the perioperative use of nitrous oxide or neostigmine were not associated with postoperative vomiting; however, patients at risk received prophylaxis more frequently (data not shown).
Patients undergoing a reoperation did not have an increase in incidence of postoperative vomiting (6.5% versus 7.9%; P = 0.522) and were not more likely to receive intraoperative prophylaxis (35% versus 32%; P = 0.435). Among patients with pituitary tumors, tumor size did not affect the incidence of postoperative vomiting (7.4% for macroadenomas versus 5.2% for microadenomas; P = 0.276). The rate of intraoperative prophylaxis was similar between the two groups (15% versus 19%; P = 0.741). Patients who received ketorolac for postoperative analgesia tended to have a decreased incidence of vomiting; however, it was not statistically significant (4% versus 8%; P = 0.159). Patients who received ketorolac were no more likely to receive intraoperative prophylaxis (37% versus 32%; P = 0.285).
Of the 258 patients who received a single antiemetic treatment in the OR with either droperidol or ondansetron, the overall incidence of vomiting was not reduced (Table 4). Similarly, in the 25 patients who received multiple antiemetics in the OR (dexamethasone plus droperidol; droperidol plus ondansetron; or dexamethasone, droperidol, and ondansetron), the incidence of postoperative vomiting was not significantly reduced. Indeed, intraoperative antiemetic prophylaxis of any kind did not affect the incidence of postoperative vomiting (5.65% versus 8.46%; P = 0.141; Table 5).
|
|
Nevertheless, patients who received intraoperative prophylaxis with a single drug were less likely to require rescue therapy with a second drug for nausea or vomiting (Table 6). Finally, patients who received initial prophylaxis in the OR were more likely to require rescue therapy with a second drug than patients who were initially treated in the PACU (15.0% versus 40.3%; P < 0.001).
|
Patients who later developed transient DI had increased morphine consumption in the PACU compared with patients who did not (6.3 mg versus 5.3 mg; P = 0.048). However, PACU morphine consumption was not related to the development of persistent DI requiring long-term treatment with desmopressin (6.1 mg versus 5.5 mg; P = 0.737). Patients who received ketorolac required more morphine in the PACU compared with patients who did not (6.7 mg versus 5.2 mg; P = 0.012). No bleeding complication immediately attributable to ketorolac usage was noted.
No clinical disease (Cushing's disease, acromegaly, prolactinoma, craniopharyngioma, or Rathke's cleft cyst) was associated with increased or decreased morphine consumption (data not shown). An intraoperative CSF leak and subsequent fat grafting (P = 0.361), tumor size (P = 0.114), use of neuromuscular blockade (P = 0.754), and redo operations (P = 0.540) were not associated with altered morphine requirements in the PACU. Finally, the use of a lumbar intrathecal catheter was associated with a significantly decreased need for morphine in the PACU (3.5 mg versus 5.6 mg; P = 0.008).
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Postoperative pain after craniotomies has long been thought to be less severe than other postsurgical pain. Some reviews have suggested that up to 60% of postsurgical craniotomy patients have moderate to severe pain (17). In this review, the overall small morphine requirements among patients likely reflect the less-invasive nature of transsphenoidal surgery. However, another study evaluating pain after craniotomy also noted small analgesic requirements (18). Finally, during surgical preparation, the mucosal surfaces of the nose are infiltrated with local anesthetic and epinephrine solution to reduce bleeding and facilitate dissection. This local anesthetic may contribute to analgesia in the PACU.
The overall incidence of vomiting in the immediate postoperative period among patients having transsphenoidal surgery was 7.5%. This is much less than the previously reported incidence of vomiting after craniotomy in both children (8) and adults (5–7,9,10). This could be due to the less invasive nature of transsphenoidal surgery, which is less likely to disturb the CTZ. Furthermore, the smaller incision and less disruption of surrounding structures may lead to less inflammation and pain, with subsequently less postoperative vomiting (11). Another possible explanation and potential shortcoming of this retrospective review is that patients may have vomited after leaving the PACU, and thus, we may have under-estimated the overall incidence of emesis. However, the peak incidence of vomiting after neurosurgery is within the first few postoperative hours (6,7,9,10); thus, any under-estimation is likely to be relatively small.
An intraoperative CSF leak and subsequent fat grafting were significantly correlated with a doubling in the incidence of postoperative vomiting in the PACU (Table 1). Notably, intraoperative antiemetic prophylaxis did not reduce this incidence. This increase in postoperative vomiting cannot be attributed to an increased need for morphine in the postoperative period because these patients did not have increased morphine consumption. The reason for increased postoperative vomiting in patients who receive fat grafts is unknown, but one possibility is the deliberate inflammatory reaction produced by the fat graft (12).
Another possibility may be that the loss of CSF pressure secondary to a leak may itself be responsible for the increased incidence of vomiting. Finally, a CSF leak and subsequent need for fat grafting may merely reflect a more aggressive surgical resection and increased likelihood for stimulation of the vomiting center and the CTZ. Fat grafting is associated with a significant increase in the incidence of DI, with a more aggressive resection being the most likely origin (19). Regardless, intraoperative prophylaxis with droperidol or ondansetron failed to impact the incidence of postoperative vomiting among patients in this subgroup. The lack of efficacy of prophylactic ondansetron for emesis after craniotomy has been reported (6,8). This review was not large enough to detect a difference between droperidol and ondansetron in this small subgroup of patients.
The placement of a lumbar intrathecal catheter increased the incidence of postoperative vomiting by approximately threefold (Table 2). The catheter can be used to manipulate CSF pressure by the injection of saline or removal of CSF. In addition, intrathecal air may be injected to increase CSF pressure and may push a suprasellar tumor down into the surgical field (14). The manipulation of CSF pressure may be directly responsible for the increased incidence of postoperative emesis because patients may have increased or decreased (20) CSF pressure. Because a lumbar drain is more likely to be used in patients with larger and more complex tumors and in patients with suprasellar extension, lumbar drain use may simply be reflective of a more aggressive surgical resection, more inflammation, and more simulation of the CTZ. As with patients who underwent fat grafting, intraoperative antiemetic prophylaxis did not reduce the incidence of vomiting among patients with a lumbar drain. This review was not large enough to detect a difference between droperidol and ondansetron in this small subgroup of patients.
Curiously, patients who had a lumbar drain placed required less postoperative morphine than patients without a lumbar drain. This is surprising because the most common symptom of altered intracranial pressure is headache (21,22), which would likely be treated with opioids in the PACU. We are unable to explain this interesting finding.
Patients presenting for resection of a craniopharyngioma were the only disease subgroup with an increased rate of postoperative vomiting (Table 3). Again, intraoperative antiemetic prophylaxis did not reduce the incidence of postoperative vomiting. In fact, the use of prophylaxis in this cohort was actually more frequent than the overall the rate in the study. Craniopharyngiomas are significantly larger than most other pituitary tumors and require significantly more aggressive surgical resection. The invasiveness of these tumors and their potential proximity to the CTZ and the vomiting center may be responsible for the increased incidence of postoperative emesis. In addition, these patients tend to be younger than most patients presenting for resection of a pituitary adenoma. Younger patients are at an increased risk for postoperative emesis (1,7).
Patients who later developed DI had an increased requirement for morphine analgesia in the PACU. It is unknown how DI or the relative deficiency of antidiuretic hormone (ADH) and pain relate to one another. It is likely that the development of DI is simply related to a more invasive surgical resection (19). A more extensive exploration of the gland and stalk and increased manipulation may result in damage and, thus, an increased risk for the disruption of ADH delivery and, hence, DI (19,23,24). A more aggressive resection may also be more painful and, thus, be associated with an increased requirement for postoperative analgesics. Also, certain emotional states can interrupt the delivery of ADH from the hypothalamus to the posterior pituitary or impair release of ADH from the posterior pituitary (19). The experience of pain produces an emotional state that can cause such disruption.
In all patients, intraoperative prophylaxis with a single drug, either ondansetron or droperidol, was not associated with a decreased incidence of vomiting in the PACU (Table 4). The lack of efficacy of prophylactic ondansetron for emesis after craniotomy has been reported in patients having a supra- and infratentorial craniotomy (5,6,8) but not in patients undergoing transsphenoidal surgery. In addition, droperidol has documented efficacy after supratentorial craniotomy (6) for the prevention of emesis but seems to lack efficacy among patients presenting for transsphenoidal procedures. Perhaps transsphenoidal surgery for a pituitary or sellar tumor provides more direct stimulation to the CTZ secondary to its greater proximity than more peripheral, supratentorial tumors. However, it is difficult to reconcile this hypothesis with the observation that the overall incidence of emesis is significantly less among patients having transsphenoidal surgery. It is most likely that prophylaxis was given to patients deemed to be at an increased risk for PONV. Consequently, the apparent lack of efficacy may merely reflect a reduction in a patient's increased risk to baseline.
One of the most interesting observations is that patients who were treated in the PACU were actually less likely to require rescue treatment than patients who received intraoperative prophylaxis. From a purely cost-effective standpoint, one could argue that this observation, combined with the apparent lack of efficacy of intraoperative prophylaxis in the prevention of postoperative emesis, it is better to hold antiemetic therapy until the PACU when patients vomit or complain of nausea. However, this would require all patients to first experience the undesirable sensation of nausea or to actually vomit before being treated, which might decrease overall patient satisfaction. Nevertheless, a similar strategy of withholding treatment until patients first complain of nausea was not associated with decreased patient satisfaction in all but the highest-risk patients (25).
The incidence of PONV after transsphenoidal procedures are less frequent than that reported for supra- and infratentorial craniotomy. We have defined risk factors for postoperative emesis after transsphenoidal surgery. These include intraoperative CSF leak and the need for fat grafting, use of a lumbar intrathecal catheter, and surgery for the resection of a craniopharyngioma. Finally, after transsphenoidal surgery patients require a relatively small amount of morphine in the PACU to provide effective analgesia, regardless of the type of tumor. Increased morphine requirements may be predictive of later, in-hospital transient DI.
Whereas transsphenoidal surgery is considered to carry less risk than other approaches to the sella, transsphenoidal resections are still associated with morbidity that can potentially alter outcome and patient satisfaction. Our study suggests possible risk factors for PONV and pain specific to transsphenoidal resection. These factors warrant consideration in further improving the outcome of this less invasive surgical technique.
|
Footnotes |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
