Acute postsurgical care of the patient undergoing craniotomy has historically relied on intensive care unit (ICU) monitoring with resource-demanding
nursing attention and invasive physiological monitoring as cornerstones.2,11,20 While many factors inform this prac-tice, postoperative management in the ICU of all crani-otomy patients is conservative but fails to individualize care and utilizes substantial resources.9 Identifying those patients who can be safely managed in a less intensive set-ting is an area of academic, practical, and economic inter-est in providing high-value health care. It is also an area of cogent controversy guided by only weak evidence.1 We
hypothesized that it is safe to manage patients outside the ICU after uncomplicated supratentorial craniotomy for tu-mor resection. A protocol was instituted, and this report examines our initial experience.
MethodsCare Protocol
A consensus-driven intention-to-treat protocol was created and implemented by a multidisciplinary team in-cluding neurosurgeons, floor and recovery room nurses, anesthesiologists, and the neurosurgery physician extender
ABBREVIATIONS ICU = intensive care unit; PACU = postanesthesia care unit.SUBMITTED April 14, 2016. ACCEPTED October 13, 2016.INCLUDE WHEN CITING Published online March 3, 2017; DOI: 10.3171/2016.10.JNS16954.
A protocol for postoperative admission of elective craniotomy patients to a non-ICU or step-down settingJeffrey E. Florman, MD,1,2 Deborah Cushing, RN, MPH,1 Lynne A. Keller, RN, CNRN,1 and Anand I. Rughani, MD1,2
1Neuroscience Institute, Maine Medical Center, Portland, Maine; and 2Department of Neurosurgery, Tufts University Medical Center, Boston, Massachusetts
OBJECTIVE Selecting the appropriate patients undergoing craniotomy who can safely forgo postoperative intensive care unit (ICU) monitoring remains a source of debate. Through a multidisciplinary work group, the authors redefined their institutional care process for postoperative monitoring of patients undergoing elective craniotomy to include transfer from the postanesthesia care unit (PACU) to the neurosurgical floor. The hypothesis was that an appropriately selected group of patients undergoing craniotomy could be safely managed outside the ICU in the postoperative period.METHODS The work group developed and implemented a protocol for transfer of patients to the neurosurgical floor after 4-hour recovery in the PACU following elective craniotomy for supratentorial tumor. Criteria included hemodynami-cally stable adults without significant new postoperative neurological impairment. Data were prospectively collected including patient demographics, clinical characteristics, surgical details, postoperative complications, and events sur-rounding transfer to a higher level of care.RESULTS Of the first 200 consecutive patients admitted to the floor, 5 underwent escalation of care in the first 48 hours. Three of these escalations were for agitation, 1 for seizure, and 1 for neurological change. Ninety-eight percent of patients meeting criteria for transfer to the floor were managed without incident. No patient experienced a major compli-cation or any permanent morbidity or mortality following this care pathway.CONCLUSIONS Care of patients undergoing uneventful elective supratentorial craniotomy for tumor on a neurosurgi-cal floor after 4 hours of PACU monitoring appears to be a safe practice in this patient population. This tailored practice safely optimized hospital resources, is financially responsible, and is a strong tool for improving health care value.https://thejns.org/doi/abs/10.3171/2016.10.JNS16954KEY WORDS craniotomy; socioeconomics; postanesthesia care unit; intensive care unit
Safety of managing craniotomy patients on the neurosurgical floor
J Neurosurg Volume 127 • December 2017 1393
group. Referred to as “Four to the Floor,” it is summarized by the following: patients undergoing elective, uncompli-cated, supratentorial craniotomy for tumor resection will be observed for 4 hours in the postanesthesia care unit (PACU) before transfer to the neurosurgical floor if stable neurologically and physiologically. Inclusion and exclu-sion criteria for transfer to the floor are listed in Table 1. Those patients who were not automatically excluded (e.g., because of a ventricular catheter) were evaluated after 4 hours in the PACU by either a neurosurgeon or neurosur-gical physician assistant.
Many factors were defined prospectively so as not to influence the decision regarding transfer to the floor. Pre-operative factors deemed irrelevant to the decision includ-ed medical comorbidities, preoperative seizure history, tu-mor size, advanced age, and supratentorial tumor location including side, lobe, depth, and intraventricular location. Intraoperative factors that were acknowledged but not se-lected as inclusion criteria were anesthesia type, seizure during awake mapping (without status epilepticus), op-erative time, craniotomy size, estimated blood loss, and pathology type. Postoperative factors that were discussed and deemed irrelevant included nausea or vomiting, blad-der catheterization, and use of intravenous pain medica-tions. Postoperative imaging is routinely performed at 12–24 hours in the form of MRI unless contraindications exist, and modification of our postoperative imaging prac-tice (e.g., obtaining an early CT scan) was considered but believed to be unnecessary in light of a favorable neuro-logical status.
PACU nursing at our institution involves a maximum of 2 patients assigned per nurse. Vital signs with neurologi-cal examination are performed every 15 minutes. Arterial lines, when placed intraoperatively, are maintained until the patient is hemodynamically stable and no intravenous antihypertensives have been given for more than 1 hour. Not all patients undergoing craniotomy at our institution routinely receive placement of an arterial line, but when placed it is usually maintained in the PACU. The blood pressure target is systolic blood pressure less than 150 mm Hg. All patients are observed on telemetry and with con-tinuous oxygen saturation monitoring. Upon transfer to the neurosurgical floor, nursing care is based on a maximum of 4 patients assigned per nurse. Vital signs with neuro-logical examination are performed every 2 hours for the first 8 hours and then every 4 hours. Between the formal nursing evaluations, nursing aids perform “bed checks” every hour with no vital signs or detailed neurological ex-aminations.
Clinical judgment was to determine the indications for escalation of care, in addition to defined criteria that in-cluded seizure, neurological change including any focal change, agitation or impulsiveness, and any cardiopulmo-nary instability, including hypertension sustained above 150 mm Hg systolic. An interim analysis was performed after the first 50 patients, and each escalation of care event was reviewed concurrently to ensure safety.
Data CollectionDemographic, clinical, and outcome details were col-
lected for consecutive patients at Maine Medical Center
for the first 200 patients that completed the “Four to the Floor” care protocol. We excluded patients from analysis if they were eligible for transfer to the floor but did not get transferred as a result of bed availability. Data were col-lected from the electronic medical records and managed using REDCap electronic, web-based, secure data capture tools hosted at Tufts University School of Medicine. In-stitutional review board approval was obtained, and a re-quirement for informed consent was waived.
Demographic data, comorbidities, and presurgical pa-tient characteristics were extracted. Surgical details and findings were collected including side of the lesion, pa-thology type, anesthesia type, estimated blood loss, use of invasive arterial blood pressure measurement, blood transfusions, operative time, use of anticonvulsants, and use of intra-urethral catheters. These variables were com-pared between the 2 groups of patients sent to the floor and those that went to a higher level of care using the Fisher’s exact test. PACU neurological status and length of observation before transfer to the floor was collected. All daily progress notes, nursing notes, and discharge sum-maries were studied in detail on all patients transferred from floor beds into a higher level of care. Postoperative complications were assessed for the entire hospital stay, and length of stay was collected. Escalation of care was defined as transfer from the floor to either the step-down unit or from the floor to the ICU.
ResultsThe study population consisted of 200 consecutive pa-
TABLE 1. Inclusion and exclusion criteria for transfer to the neurosurgical floor at 4 hours postoperatively
Inclusion criteria Age >17 yrs old Extubated intraoperatively or w/in 1 hr of arrival in PACU Cognitive status & language skills must allow clear communication
& cooperation Stable or improving neurological status over the 4 hrs of observa-
tion Mild focal deficit concordant w/ presurgical examination findings Stable or improving pulmonary status on nasal cannula oxygen No new cardiac dysrhythmia or symptoms Systolic blood pressure <150 mm Hg w/o intravenous antihyperten-
sive medication for >1 hr Cleared for transfer to floor by physician extender or surgeon after 4
hrs postop observationExclusion criteria Status epilepticus intraoperatively Seizure in the PACU Intracranial drains (subgaleal drains acceptable for floor care) Need for transfusion Coagulopathy Diabetes insipidus monitoring (i.e., hourly urine output assessment) Need for intravenous medication drips at the time of transfer (e.g.,
insulin)
J. E. Florman et al.
J Neurosurg Volume 127 • December 20171394
tients transferred to the floor following 4-hour observation in the PACU from 2011 through 2015. This represented 59% of the 342 craniotomies performed for tumor resec-tion during the same time period. In addition to the 200 transferred to the floor, 13 patients were eligible for floor care per the protocol but there was no floor bed available. These patients were excluded from further analysis. The ages of the patients ranged from 18 to 93 years old (me-dian 59 years old), with 45% male and 55% female. Medi-cal comorbidities are described in Table 2 with the most common being hypertension, noted in 47% of patients. Surgical details and intraoperative findings are described in Table 3; 75% of the tumors were primary brain tumors.
The 200 patients were discharged to the neurosurgical floor after a mean of 284 minutes (median 272 minutes, range 201–495 minutes) of observation in the PACU. In the first 48 hours, 5 (2.5%) of these patients were transferred to the neurological step-down unit, an open floor plan unit attached to the neurosurgical floor with 6 beds, a central nurses station, and a 1:3 nurse-to-patient ratio, able to perform examinations every 2 hours. No patient required transfer to the ICU. The decision to transfer and the choice of unit was based on clinical judgment independent of the “Four to the Floor” protocol. These patients are described in Table 4. Three of these patients were transferred for agitation, 1 for focal seizure, and 1 for focal neurological change. Investigations into the sources of agitation were unrevealing in all 3 patients, and no structural, metabolic, or epileptic source was identified. None of the 5 patients suffered any morbidity as a result of delayed diagnoses. No correlation was observed between the need for esca-lation of care and the patients’ demographics, preopera-tive clinical characteristics, or surgical details. All of the study patients underwent postoperative imaging within 48 hours, and no clinically significant hematoma, swelling, or hydrocephalus requiring intervention was identified in any of the 200 patients. Median length of stay was 2 days and 95% were discharged within 1 week. Ninety percent of patients were discharged home.
Of the 342 craniotomies performed during this time frame, there were 28 additional patients excluded from analysis because their disposition was determined by bed availability and not the protocol. There were 13 patients excluded from analysis because they spent an overnight in the PACU, despite otherwise meeting criteria for floor disposition. There were 15 patients excluded from analysis
because they went to either the ICU or the step-down unit, again because of lack of floor bed availability. There were 114 patients who underwent craniotomy and were not sent to the floor postoperatively, and these are further charac-terized in Table 3. The specific reasons for being sent to the ICU or step-down unit are illustrated in Table 3, with the most common reason being unfavorable postoperative cognitive status in approximately one-third of patients. In-terestingly, this was the most common reason for escala-tion of level of care from the floor to the step-down unit.
Reasons for deviation from the protocol are summa-rized in Table 5. In comparing the cohort of patients who successfully followed the pathway from PACU to floor to those who deviated, it can be seen that the demograph-ics are quite similar (Table 3). There was no significant difference with regard to age, sex, or pathology treated. One of the principal exclusions listed in the protocol is supratentorial tumors, which accounts in part for the lower proportion of patients who underwent awake craniotomy,
Safety of managing craniotomy patients on the neurosurgical floor
J Neurosurg Volume 127 • December 2017 1395
left-sided surgery, or biopsy only. There were more pa-tients with an arterial line and Foley catheter who had ICU or step-down disposition, which could be reflective of the need for increased physiological monitoring both intraoperatively and operatively. Prolonged mean surgical time (67 vs 104 minutes) and increased blood loss were also observed in the ICU/step-down cohort. As would be anticipated, the patients admitted to the floor were more likely to be discharged home and had a significantly short-er length of stay.
DiscussionThere is emerging opinion that many neurosurgical pa-
tients do not require ICU care postoperatively to be pro-vided safe and appropriate care.2,3,6,9 Studying inpatient postoperative care following craniotomy exposes many opportunities to tailor the resource allocation and consider the costs and benefits of the options.2,3,11 Few authors, how-ever, have described altering their practices or vetting new care paradigms. Same-day discharge has been described in well-selected patient populations although many might be hesitant to adopt such a practice.4,6,7 Certainly, it is desir-able to observe patients closely when risk for acute decline is high, but the benefits of intensive postoperative monitor-ing appear overstated in some neurosurgical populations, and disadvantages have been recognized including cost, delayed mobilization, suboptimal bed allocation, and po-tentially increased length of stay.2,8,11 The challenges are to identify those candidates within a given institution who can be monitored outside the ICU and implement change in care paradigms safely.
To develop our protocol, we found value in assembling a group of clinical stakeholders at Maine Medical Center focused on the care of the inpatients after craniotomy. Ret-rospective institutional and practitioner data informed pa-tient selection criteria to construct a care paradigm. Fur-thermore, our institutional history of utilizing a step-down unit on the same floor as our neurosurgical floor provided a facility-specific environment for nursing familiarity of such patients outside the ICU. Additional confidence was provided by our historical practice of managing a majority of craniotomy patients immediately in the recovery room and not taking them directly from the operating room to the ICU. These factors facilitated the shift to caring for patients on the floor without altering our historically pre-ferred physiological targets, invasive versus noninvasive
blood pressure monitoring practices, or institutional rules around intravenous antihypertensive medication usage. Manpower resources were inventoried to ensure thought-ful consideration of opportunities to use physician extend-ers, nurses’ aids, and nurses. Lastly, oversight for imple-mentation and review of the new protocol was shared by the extended neurosurgical team including all practicing neurosurgeons, neurointensivists, nursing, and the neuro-surgical physical extender group. The structure of our ex-perience is believed to be transferrable to other institutions and can contribute favorably to quality and affordability. “Right-sizing” postoperative care by identifying those ex-pected to have uncomplicated courses has substantial im-plications for costs and bed availability.2,9
In the first day after craniotomy, the concerns for in-tracranial complications include hematoma formation, hydrocephalus, and seizures.5,16,19 While medical condi-tions such as cardiopulmonary compromise also threaten the early postoperative course, the major source for early postoperative morbidity and mortality is hemorrhage.13 Hematoma formation complicates about 1%–4% of cra-nial tumor surgeries,19 with surgeon volume,18 patient age, and biopsy versus resection identified as potential risk fac-tors.12,13 The need for reoperation for hematoma in the first 30 days after craniotomy is approximately 2%.13 Unfor-tunately, there is a paucity of data offering granular de-scription of the chronology of events expected with early postoperative hematomas. Taylor and colleagues17 provide commonly referenced historical data suggesting that clini-cal deterioration from hematomas often occurs within the first 6 hours of surgery according to their review of a large cohort of unselected craniotomy patients, including emer-gency and elective surgeries. Other published experiences describe detection of postoperative hematomas requiring surgical intervention as typically within 2–4 hours.9,15 It is our contention that in the absence of clinical decline in the first 4–6 hours following otherwise uncomplicated su-pratentorial tumor resection, acute intracranial events are unlikely and can be identified in a timely manner outside the ICU.
The services often provided in the ICU and interme-diate care units such as electrocardiographic monitoring, arterial line monitoring, frequent nursing examination, measurement of intake and output, and intravenous medi-cation management are focused on observation and not intervention.9,11 It is the ability to make timely diagnoses and initiate interventions that is at the heart of postopera-tive disposition decisions;16 however, postoperative care is often not individualized and is variable among institutions and practitioners. This current report exposes that histori-cal practices at a given institution might strongly bias the resourcing needs and even preclude management outside the ICU. For example, invasive arterial blood pressure lines are often used despite a paucity of evidence about their utility or strict blood pressure target adherence. Such invasive monitoring and intravenous antihypertensive prescription often limit options for disposition within an institution. Furthermore, while patient factors such as dia-betes, advanced age, long operating room times, large in-traoperative blood loss, and high anesthesia risk have been suggested as predictive of the need for ICU interventions
TABLE 5. Reason for deviation from protocol
Reason %
Cognitive status 34Infratentorial tumor 28Preop bed in ICU/step-down 9Ventricular drain 8Monitoring for diabetes insipidus 6Cardiac monitoring 5New or worsened neurological deficit 5New seizure in PACU 4
J. E. Florman et al.
J Neurosurg Volume 127 • December 20171396
following craniotomy,3,8,15 such data may not necessitate ICU admission after craniotomy for an individual patient.
Although our experience serves as a large and uniquely prospective study of a change in care protocol, it has a number of limitations. It is a single-institution and single-provider experience and is not a randomized trial. Further, while this report describes a structure for protocol devel-opment and a successful paradigm at a single institution, the intent of this report is not to offer a protocol that can be widely generalized to all settings. As there were only relatively minor events in a small group of our patients with no acute hemorrhage or cardiopulmonary event, it cannot be stated with certainty that a significant delay in diagnosis or intervention could be avoided in all settings. This report suggests that such events are rare in our fo-cused study, and this limited experience describes only a limited number of postoperative scenarios. The context of this report is that it describes experience in a mature neu-rosurgical practice with 9 neurosurgeons and a full-time in-house neurosurgical extender service in a hospital with resourcing to support Level 1 trauma designation by the American College of Surgeons and stroke center certifi-cation by the American Heart Association. Many factors including facility layout, size and scope of the inpatient services, surgical team experience, nursing and extender resources, and patient characteristics are all unique to an institution; however, the opportunity to individualize care and to explore thoughtful assessment of practice ritual is not unique.
Complications occur along a lengthy timeline after surgery, and care should thoughtfully escalate and de-escalate. In our study, 98% of our patients had uneventful postoperative courses, and no patient suffered any clear morbidity or mortality even when transferred to higher acuity care. While an event that leads to escalation of care after craniotomy may reflect an unwelcome challenge, the implications of escalating care do not necessarily need to be interpreted as a care failure or complication. Other authors have reflected on a similar interpretation of un-planned care escalation as not equating with added risk or adverse outcome.3 We interpret escalation of care in the setting of our protocol as an extension of a care paradigm, as opposed to a protocol failure, because no permanent morbidity was added. All patients following our protocol were successfully managed without adverse effects on their ultimate disposition or outcome.
Acute postoperative craniotomy care is focused pri-marily on cardiopulmonary and neurological status, with limited risk for serious complications beyond 6 hours,10,14 and most required limited ICU-level interventions.8 While there may be particular concerns for reducing the risk of respiratory complications, managing labile blood pressure, and optimizing pain control, the intensity of observation, physiological targets, and role of intravenous medication are not completely driven by data. Instead, many reports simply describe the practices such as placing invasive ar-terial lines in all patients undergoing craniotomy for at least 24 hours,8 but this is certainly an overutilization of resources in some and fails to stratify based on patient needs. The findings of our study suggest that a multidisci-plinary group can assess local practice, inventory resourc-
es, vet, and implement an institution-specific protocol to safely care for a neurosurgical patient population outside the ICU after craniotomy.
ConclusionsThe evidence presented here suggests that it is feasible
and safe to manage patients undergoing craniotomy out-side the ICU after tumor surgery when they are selected thoughtfully and cared for by an experienced team. Fur-thermore, building protocols via a collaborative approach is a worthwhile effort to customize care and a compelling model for “right-sizing” access to resources. This report serves as a unique prospective study of de-escalating the intensity of early postcraniotomy care from the historical practice of ICU-based, or even step-down–based, care. Our tailored practice model safely optimizes hospital re-sources, is financially responsible, and is a strong tool for improving health care value. Care of the uneventful, elec-tive, supratentorial craniotomy for patients with tumors after 4 hours of intensive monitoring is a sound practice in our patient population and in our hospital. While this patient group may be well suited in other institutions for similar care paradigms, careful effort to integrate risk assessment systems with prospective validation appears wise.9
References 1. Awad IA: Intensive care after elective craniotomy: “all poli-
tics is local.” World Neurosurg 81:64–65, 2014 2. Beauregard CL, Friedman WA: Routine use of postopera-
tive ICU care for elective craniotomy: a cost-benefit analysis. Surg Neurol 60:483–489, 2003
3. Bui JQ, Mendis RL, van Gelder JM, Sheridan MM, Wright KM, Jaeger M: Is postoperative intensive care unit admis-sion a prerequisite for elective craniotomy? J Neurosurg 115:1236–1241, 2011
4. Carrabba G, Venkatraghavan L, Bernstein M: Day surgery awake craniotomy for removing brain tumours: technical note describing a simple protocol. Minim Invasive Neuro-surg 51:208–210, 2008
5. De Santis A, Villani R, Sinisi M, Stocchetti N, Perucca E: Add-on phenytoin fails to prevent early seizures after surgery for supratentorial brain tumors: a randomized con-trolled study. Epilepsia 43:175–182, 2002
6. Goettel N, Chui J, Venkatraghavan L, Tymianski M, Man-ninen PH: Day surgery craniotomy for unruptured cerebral aneurysms: a single center experience. J Neurosurg Anes-thesiol 26:60–64, 2014
7. Grundy PL, Weidmann C, Bernstein M: Day-case neurosur-gery for brain tumours: the early United Kingdom experi-ence. Br J Neurosurg 22:360–367, 2008
8. Hanak BW, Walcott BP, Nahed BV, Muzikansky A, Mian MK, Kimberly WT, et al: Postoperative intensive care unit requirements after elective craniotomy. World Neurosurg 81:165–172, 2014
9. Hecht N, Spies C, Vajkoczy P: Routine intensive care unit-level care after elective craniotomy: time to rethink. World Neurosurg 81:66–68, 2014
10. Khaldi A, Prabhu VC, Anderson DE, Origitano TC: The clinical significance and optimal timing of postoperative computed tomography following cranial surgery. J Neuro-surg 113:1021–1025, 2010
11. Knaus WA, Draper E, Lawrence DE, Wagner DP, Zimmer-man JE: Neurosurgical admissions to the intensive care unit:
Safety of managing craniotomy patients on the neurosurgical floor
J Neurosurg Volume 127 • December 2017 1397
intensive monitoring versus intensive therapy. Neurosurgery 8:438–442, 1981
12. Lassen B, Helseth E, Egge A, Due-Tønnessen BJ, Rønning P, Meling TR: Surgical mortality and selected complications in 273 consecutive craniotomies for intracranial tumors in pedi-atric patients. Neurosurgery 70:936–943, 2012
13. Lassen B, Helseth E, Rønning P, Scheie D, Johannesen TB, Mæhlen J, et al: Surgical mortality at 30 days and complica-tions leading to recraniotomy in 2630 consecutive cranioto-mies for intracranial tumors. Neurosurgery 68:1259–1269, 2011
14. Morasch MD, Hirko MK, Hirasa T, Burke K, Greisler HP, Littooy FN, et al: Intensive care after carotid endarterectomy: a prospective evaluation. J Am Coll Surg 183:387–392, 1996
15. Rhondali O, Genty C, Halle C, Gardellin M, Ollinet C, Odd-oux M, et al: Do patients still require admission to an inten-sive care unit after elective craniotomy for brain surgery? J Neurosurg Anesthesiol 23:118–123, 2011
16. Sughrue ME, Bonney PA, Choi L, Teo C: Early discharge after surgery for intra-axial brain tumors. World Neurosurg 84:505–510, 2015
17. Taylor WA, Thomas NW, Wellings JA, Bell BA: Timing of postoperative intracranial hematoma development and im-plications for the best use of neurosurgical intensive care. J Neurosurg 82:48–50, 1995
18. Trinh VT, Davies JM, Berger MS: Surgery for primary supra-tentorial brain tumors in the United States, 2000–2009: ef-fect of provider and hospital caseload on complication rates. J Neurosurg 122:280–296, 2015
19. Wong JM, Panchmatia JR, Ziewacz JE, Bader AM, Dunn IF, Laws ER, et al: Patterns in neurosurgical adverse events: intracranial neoplasm surgery. Neurosurg Focus 33(5):E16, 2012
20. Zimmerman JE, Junker CD, Becker RB, Draper EA, Wag-ner DP, Knaus WA: Neurological intensive care admissions: identifying candidates for intermediate care and the services they receive. Neurosurgery 42:91–102, 1998
DisclosuresThe authors report no conflict of interest concerning the materi-als or methods used in this study or the findings specified in this paper.
Author ContributionsAcquisition of data: Cushing, Keller. Analysis and interpretation of data: all authors. Drafting the article: Rughani, Florman. Criti-cally revising the article: Rughani, Florman. Reviewed submitted version of manuscript: Rughani. Approved the final version of the manuscript on behalf of all authors: Rughani. Administrative/technical/material support: Florman. Study supervision: Florman.
CorrespondenceAnand I. Rughani, Maine Medical Center Neurosurgery and Spine, 49 Spring St., Scarborough, ME 04074. email: [email protected].
